LTV Research · SaaS growth

How Founders Actually Grow SaaS

We read 4,273 founder interviews from Starter Story, Greg Isenberg, Nathan Latka, MicroConf, Rob Walling, Lenny's Podcast and My First Million. We kept only tactics with an action, a context and proof, then ranked what founders repeat most. Every number links to the video it came from.

4,273founder interviews read in full
16,711tactics with an action, context and proof
3,199fundamentals after merging identical tactics
863used by 5+ founders across 2+ channels
78the strongest in each of 16 areas, below

Why so few: 1,372 fundamentals came from a single founder and 898 from 2 to 4, not enough repetition to call them a pattern. Of the 863 left, this page keeps the 4 to 6 most used in each area.

Key takeaways

  1. Validate by talking before you code. Ask how people solve the problem today, never whether they would use your idea. The most cited fundamental across channels: 41 founders, 7 channels.
  2. Build the audience before the product, then launch to it. Building in public, posting daily for months and repurposing what worked show up together in 5 to 7 channels.
  3. First customers come from where the audience already is: niche communities (help without selling), app marketplaces of the platform they use, and pre-sales with a landing page.
  4. Small MVP, focus on the core that already sells. Many started by delivering the service by hand and automated only what repeated.
  5. Price is a growth lever: annual prepayment funds acquisition, and raising prices in steps while measuring conversion shows up in 7 channels.
  6. Growing inside the base that already pays (add-on modules, land and expand) dominates among larger SaaS, almost all of it from Latka interviews.
  7. Paid media only after you know your CAC ceiling and payback. Those who scaled raised budget in steps while returns held.

How we built this

  • Sources: every long-form video from the 7 channels, filtered for SaaS topics. 5,206 published, 4,367 about SaaS.
  • Transcription: YouTube captions. 4,287 videos.
  • Mining: every video read in full; only tactics with an action, context and proof (a number or a real case) were kept. 4,273 videos, 16,711 tactics, 877 shallow videos cut.
  • Merging: tactics with the same mechanism, condition and effect became one fundamental, whatever their names. 3,199 fundamentals, no tactic lost.
  • 80/20 ranking: independent founders, number of channels and revenue cited, with less weight for Latka (he asks the same questions every episode). 78 fundamentals selected.
  • Authority: founders describing what they did, in interviews. Numbers are self-reported, not audited. Every number links to its video.

All 78 tactics at a glance

Each tactic shows how many different founders used it and in how many of the 7 channels it appears. The higher both numbers, the more confirmed the pattern.

01

Idea and validation

Six fundamentals of how founders checked an idea before spending months building it.

Interview your audience about today's problem, without pitching the idea

41 founders used itin 7 of 7 channels5 with revenue cited

What it is

You book short conversations with people in your audience and ask how they solve the problem today and what they have already tried. Your solution stays out of the conversation. Stop when the answers start repeating.

Why it works

Questions about the past show what the person actually did. Questions like "would you use it?" only get you politeness and empty promises. When the same problem shows up across several conversations, you have the real pain and a list of first customers at the same time.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

Before coding, when you have a problem hypothesis and access to the audience. It also works in the 0 to 10k MRR phase, while the niche is still unclear.

Step by step

  1. Write the hypothesis: "a specific type of person has a problem with a specific task". List the riskiest assumptions.
  2. Invite people with that profile by email or LinkedIn, asking for help and a 15 to 20 minute call. Don't sell anything.
  3. Ask: how do you solve this today? Walk me through the last time you did it. How much did it cost in time or money? What would happen if you did nothing?
  4. Ask for the name of someone with the same problem.
  5. Do 10 to 20 conversations (the benchmark for software sellers is 25) and note what repeats.
  6. Only build once the common pain shows up. Cap the research timeline so it doesn't turn into months of shipping nothing.

Best example

Fastlane did about 2,000 discovery calls, always asking how the person solves the problem today, and reached $69,000 a month (vídeo). In another case, Jason Cohen emailed 40 consultants, all of them agreed to talk, and he held 38 conversations (vídeo).

Numbers that came up

  • Mike Taber (Bluetick) emailed about 15 to 20 people to start the conversations (vídeo).
  • Hiten Shah (KISSinsights) did 20 phone interviews with product managers and reached more than 25 thousand signups before selling, at zero marketing cost (vídeo).
  • Greg Isenberg interviewed 7 or 8 doctors, came up with 43 problems, and picked bloating (Unbloat) based on search volume (vídeo).
  • Luna spent a year with MRR stuck at about $1,000 while doing about 5 interviews a week, and then grew about 50% a month (vídeo).

When it does NOT apply

  • In mass markets, where there are almost no early adopters, the return drops (vídeo).
  • If the question already mentions your idea, the data comes out biased (vídeo).
  • Conversations give direction, not proof of demand (vídeo). With thousands of users, A/B testing starts to measure better (vídeo).

Turn your own problem into a product and sell it to peers in the same field

37 founders used itin 6 of 7 channels23 with revenue cited

What it is

You first solve a pain in your own operation or day to day. Then you offer the same solution to people who have the same problem, usually peers in the same industry.

Why it works

Whoever lives the pain knows what's missing and already has a first test user: you. Peers in the field have the same pain, so the first customers come through referrals and the tool arrives already tested.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

When you have a business, a job, or a routine with a recurring problem and you know other people with the same pain. The highest-revenue cases came from people who already had a customer base.

Step by step

  1. Write down the repetitive problems in your day to day and your work. Check whether a solution already exists.
  2. Pick the one you suffered through firsthand and build the minimum version for your own use.
  3. Use it in your operation until it works.
  4. Offer it to friends and peers who have the same problem and see whether they close a deal with the tool.
  5. Use the requests that come in to decide what to sell.
  6. Confirm that other people pay, because solving your own problem doesn't guarantee a sale.

Best example

Alex Canter connected several delivery apps into a single device at his family's restaurant. Ordermark went from about 15 restaurants in November 2017 to more than 1,000 in 2018 (vídeo).

Numbers that came up

  • Lancer: two agency-owner friends closed, together, three five-figure clients in two weeks with the tool (vídeo).
  • Jane App: initial version built in about six weeks for about US$30K, today with 18 thousand paying practitioners and about US$4.7M in ARR (vídeo).
  • Habit Kit, built for personal use: 300 thousand downloads and 15 thousand in MRR (vídeo).
  • CardStock: about 15 thousand active subscribers and up to 75 thousand in MRR, from an app the founder couldn't find ready-made (vídeo).

When it does NOT apply

  • Without a ready customer base the path is slower. Vantica and Lancer had an operation and a network to test with (vídeo, vídeo).
  • Solving your own problem doesn't guarantee a sale. You need to see whether other people pay (vídeo).
  • It can take a long time: Wish List took about 6 years to reach about US$150K a year (vídeo).

Open pre-sale to the public, with a page and a goal, to pay for the build

35 founders used itin 6 of 7 channels24 with revenue cited

What it is

You publish a paid offer for something that doesn't exist yet, with a page, a price, and a payment button. The advance sale shows demand and funds development.

Why it works

Money coming in measures interest better than "I would pay". If nobody buys, you haven't spent on product or inventory. If people buy, you already have cash and a list of customers to listen to.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

When the product hasn't been built yet and you have a list, community, or traffic to receive the offer. It works best with easy-to-buy products at low or medium prices.

Step by step

  1. Build a short page with the problem, what you will deliver, the price, and a payment button (it can be ugly).
  2. Send a personal email to your list and contacts, or run a little advertising.
  3. Set a minimum sales goal and only start coding once you hit it.
  4. Offer an easy refund and state the delivery deadline.
  5. Call the first buyers and ask what they expected.
  6. If it sells out, open more spots and use the cash to build.

Best example

Omar, from Webinar Ninja, made an ugly page with PayPal and sold 150 beta spots at US$300 in 48 hours to a list of about 3 thousand people. He opened 100 more, which sold out in 24 hours, and raised about US$75K to develop the product (vídeo).

Numbers that came up

  • Subscribr: US$20K in 2 or 3 days, with 50 lifetime licenses, before the app existed (vídeo).
  • Mike Taber (Bluetick): talked to about 15 people, 11 paid upfront, almost US$1,500 in revenue without a line of code (vídeo).
  • The Hustle: a Google Doc with a payment link brought in about US$50K with about US$5K in ads. The pretty page barely sold (vídeo).
  • Hype4: more than 1,000 pre-sales in three weeks at US$35 each, at 70% off (vídeo).

When it does NOT apply

  • Without a list or audience, the effect is smaller. Trends had a list of about one million emails (vídeo).
  • Pre-sale doesn't guarantee retention. RightMessage had high churn afterward (vídeo).
  • Above about US$1,000 in ticket size, the person who presented the Trends case doesn't recommend this shortcut (vídeo). Someone who pays at signup may still not use the product (vídeo).

Publish a simple offer (page, waitlist, post) and measure interest

31 founders used itin 6 of 7 channels16 with revenue cited

What it is

You put up a page, a post, or a waitlist that describes the product and count how many people ask for it. Nobody pays yet and you haven't built the whole product.

Why it works

A page or a post costs little and shows whether the message attracts people. The number of signups, likes, or replies becomes a decision criterion, and those who sign up become the first conversation.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

When you can publish the offer and attract an initial audience (your network, newsletter, paid traffic). It serves to decide whether the idea deserves the investment, before building.

Step by step

  1. Put up a landing page explaining who it's for and what problem it solves, with email capture. Another option is to post the idea on X and see the replies.
  2. Bring traffic: your network, newsletter, ads, Reddit, or Product Hunt.
  3. Set the cutoff number beforehand (for example 500 signups, 15% conversion or more, or more than 20% retweets relative to likes).
  4. Talk in person to the people who left their email.
  5. Compare interest with payment: see whether those who signed up buy.
  6. If it passes the cutoff, build. If not, set the idea aside.

Best example

Buffer was a 3-page site with a headline, a pricing page, and a "we're not ready yet" page with email capture. It took seven weeks from idea to launch, and the first paying customer came in four days, with US$5 (vídeo).

Numbers that came up

  • Vinyl Me Please: 100 emails from the personal network turned into 12 paying subscribers at launch, about 12% (vídeo).
  • Atlassian (Jira Product Discovery): more than 3,000 waitlist signups in about two weeks, without a single line of code (vídeo).
  • Drip: cold traffic usually gives 5% to 10% signup, one page reached 30%, and the pre-launch list grew to about 3,400 interested people (vídeo).
  • Story ARB: the first tweet brought 25 DMs, and today there are 12 customers at US$7K a month (vídeo).

When it does NOT apply

  • A signup is not revenue. Harpoon gathered about 3,000 signups and the users left right afterward (vídeo).
  • Without an audience, a landing page alone has little effect (vídeo).
  • The cutoffs (500 signups, 20% retweets, 15% conversion) are each founder's personal rules and vary with the market (vídeo, vídeo).

Deliver the service by hand and automate only what repeats

38 founders used itin 6 of 7 channels30 with revenue cited

What it is

You do the work manually for the first customers, charging for it. Then you turn only the repeating parts into software.

Why it works

Doing it by hand shows the real process, the repeated pain, and how much the customer pays, and it brings in revenue before a product exists. Service money funds development, and the software is born with customers and a clear process.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

When you can deliver the service alone or with a small team and there is a repeated pain. Many cases already had an agency, consultancy, or service team in operation.

Step by step

  1. Find 3 to 5 customers with an urgent, expensive problem.
  2. Deliver the solution by hand (spreadsheet, no-code tools, a team of analysts), charging for it.
  3. Ask customers what they would use every day and note what repeats.
  4. Raise the price gradually and see whether the queue holds.
  5. Use the service revenue to pay for development.
  6. Automate what repeats and launch the sellable version without waiting for the complete product.

Best example

List Kit launched a SaaS-looking version on its site, but delivered by hand by the team within 24 hours. The service earned about 100 to 110 thousand dollars a month and the SaaS reached 1,000 paying customers in the first 6 months (vídeo).

Numbers that came up

  • Mobile.io: about 1 million dollars in revenue from the service, with about 15 people, before becoming SaaS, and today close to 1.5 million in ARR (vídeo).
  • LeadFuze: about 250 thousand dollars in cash in 2015 from the service, and the software launched with about 6,500 dollars in MRR (vídeo).
  • Folderly: a US$2K audit and then about US$10K with the first customer, which funded the engineering team (vídeo).
  • Drivex: a pilot of 100 inspections for about US$500 with an insurer, sold before the full technology existed (vídeo).

When it does NOT apply

  • The manual mode doesn't scale, and on its own it is service revenue, not recurring (vídeo, vídeo).
  • It depends on having a base: Folderly came out of an agency with a ready team, and Aptum had about US$1.5 million of its own to invest (vídeo, vídeo).
  • Shutting down the service to focus on the software can cause a period of negative revenue, as happened at LeadFuze (vídeo).
02

Product

Launch a lean MVP with minimal scope and a short deadline

32 founders used itin 6 of 7 channels14 with revenue cited

What it is

You build only what tests the product's core hypothesis and put it live fast. Everything else waits until customers are asking for it. The goal is real feedback early, before you spend months on something nobody needs.

Why it works

The more you build before launching, the more bets you make without validating any of them. Cutting scope shortens the deadline, and a short deadline brings the first real usage sooner. That usage shows what is worth building next and what could have been left out.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

Before your first launch, when you feel tempted to think one more feature is missing. It works best if you already know which problem you want to test.

Step by step

  1. Write down the core hypothesis and list only what tests it. Coffee Meets Bagel kept four items: Facebook login, a profile with like or pass, an admin tool to create the matches, and anonymous chat.
  2. Set an hours budget and cut whatever does not deliver the core value (multi-CRM, queues, automatic data loading).
  3. Ask which feature is mandatory for the first customer. If it is not, cut it and launch.
  4. Launch with the main function, without a pretty landing page. The polished page comes once you have users.
  5. If you want to reduce risk, use the product yourself or hand it to a small beta group, then launch right after.
  6. With feedback in hand, decide the next step. If the smaller version solves about 80% of the problem, stop; if customers ask for more, go back and expand.

Best example

SiteGPT went from idea to live in two weeks and reached 10K in MRR in its first month (How I Grew My App to $13K/month).

Numbers that came up

  • EUform: MVP in 2 weeks and 11K in MRR. The pretty landing page only came with about 200 users (video).
  • SEG Metrics: launched in 50 hours, with an 8-hour MySQL optimization cut from the budget (video).
  • Cursor (Anysphere): launched about 3 months after the first line of code and hit US$ 100 million in ARR in about a year and a half (video).
  • CodeGuide: 100 users on day one, after two weeks to build the responsive app (video).

When it does NOT apply

  • If the MVP ends up too ugly: Dylan Field, from Figma, says it needs to have something "a little bit amazing" (video).
  • If the support and sales team needs to use the feature right away: at Revolut, they reduce features but keep UX, speed and polish, which takes longer (video).
  • Without an audience that already pays for the problem: SEG Metrics had a consultancy feeding the product (video).

Focus everything on the core that already sells and cut the other fronts

33 founders used itin 6 of 7 channels14 with revenue cited

What it is

You pick the product or market that already has traction, put the whole team on it and cut the rest, even if another line still brings in revenue. Once the core is strong and predictable, its slack funds the next bet.

Why it works

A team spread across several fronts delivers little on each one. Focusing creates efficiency in what already works, improves retention and frees up cash. That cash funds a new bet without pulling engineers off the core.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

When the core product already has traction and you can measure every detail of it. Also when the company finds itself with three, four or more products or markets and none of them is moving properly.

Step by step

  1. List everything the company does: products, markets, platforms, priorities.
  2. See which line sells the most or grows on its own. At AppSumo, software sold 10 times more than the other lines.
  3. Choose one bet and stop the rest, even if another line still pays. Intercom did this with customer support.
  4. Ask what the one thing is that makes the product indispensable and get that feature solved before adding the next.
  5. Measure the core in detail and invest in it more than your competitors do.
  6. Only then use the core's profit for a new bet. If an old product distracts the team, consider selling it, as Onboard Meetings did with Ensemble.

Best example

Mixpanel had 50 engineers spread across three domains and revenue churn got close to 40% in 2018. After going back to the core, retention rose from about 60 to 90 and NPS from 16 to 50 (An inside look at Mixpanel’s product journey).

Numbers that came up

  • Limelight Health: it had 405 customers (400 brokers and 5 insurers). After cutting markets, it was sold about eight months later for about US$ 93 million (video).
  • Intercom: it had US$ 80 million in ARR on the other line and still focused on customer support. Fin is growing above 300%, according to the CEO (video).
  • Basecamp: with 40 to 45 people, they kept 12 codebases before going back to focusing only on Basecamp (video).
  • Sendlane: average ticket went from about US$ 25 to 30 to about US$ 255 per month after focusing on e-commerce (video).

When it does NOT apply

  • If the cut can cost you customers: Sendlane lost about 75% of its base, from about 6,500 to 1,650 customers (video).
  • If your core is not the best in its category: Mixpanel says that may be the time to reassess (video).
  • In a single-core product that is still immature: the idea of using the core's slack for new bets applies to large, mature companies, like Uber (video).

Sell an adjacent product to the customer base that already trusts you

29 founders used itin 6 of 7 channels21 with revenue cited

What it is

After an entry product sells, you launch another one for the same buyer, using the base you have already won. Each customer starts to bring in more, and the market you reach gets bigger.

Why it works

A satisfied customer already knows your brand and has already agreed to pay. Offering something adjacent to them costs far less than winning a new customer. Each extra product increases revenue per account and widens the addressable market.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

When you already have a satisfied customer base and a validated entry product. It works best if customers are already asking for more.

Step by step

  1. Look at the integration and feature requests from current customers. ConnectBooks saw customers asking for Walmart, eBay and Shopify, even with 90% of volume on Amazon.
  2. Choose a product for the same buyer, around an adjacent pain they already have.
  3. Offer it first to the base that already trusts you, before going after new acquisition.
  4. Offer a discount to those who buy both, as FindCRA did.
  5. With no team to build it, close a white label partnership with revenue share and, if it sells well within 6 to 18 months, buy it or swap equity (the Britek case).
  6. Track sales volume from day one and fix the launch problems.

Best example

OSS Ventures started with a specific operational pain at factories and today has 80 customers, 180 factories and 8 software lines, with NRR of about 130% in the prior year's cohort (OSS Ventures CEO Renan Devillieres).

Numbers that came up

  • SafetyWing: 3x growth in five months after launching covid coverage, in 2020 (video).
  • Paddle (data from about 40 thousand SaaS companies): companies with more than one product grow 30 to 50% faster than single-product ones (video).
  • Reveleer: TAM went from about US$ 2 billion to about US$ 20 billion with new suites for the same buyer (video).
  • Batch Leads: a product launched in December passed US$ 100 thousand per month within a few months (video).

When it does NOT apply

  • If the base does not respond: ConvertKit Commerce had 250 thousand users to push to, but reached about 500 thousand in GMV, well below the 1 billion goal (video).
  • If the entry product is not yet validated: Xero itself admits that the ARPU goal of going from 23 to 30 or 35 dollars is a hypothesis (video).
  • With a white label partner: the revenue split drops as the partner grows, and culture and support weigh in (video).

Start by selling a service and migrate to software as the work repeats

35 founders used itin 5 of 7 channels29 with revenue cited

What it is

You sell the service done by hand, with no finished product. You note what repeats across clients and gradually turn that into software, charged by subscription. The service pays the bills while the product is born.

Why it works

The service funds the business and teaches you which pain the customer pays to solve. What repeats across several clients becomes a higher-margin feature. In the end you are left with recurring revenue and a validated product.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

When you can sell a service without a finished product and have the time and margin to run both. It works best with 10 to 15 recurring clients or more.

Step by step

  1. Do the service for whoever pays and note the problem that shows up in each project.
  2. List the manual tasks and automate the ones that repeat. SurefireLocal connected a small tool to the Google, Facebook and Bing APIs so the client could manage it alone.
  3. Test the software with the agency's clients, charging a monthly fee (Birdseye charged from US$ 500 to US$ 1,250).
  4. Track adoption month by month, because it is the signal to keep going.
  5. Migrate clients to a subscription to the product. Be prepared for it to take years and to have to convince the sales and support team.
  6. Keep the service as a second revenue line, if it makes sense, as Unilog did.

Best example

SurefireLocal went from an agency to about 2,000 locations paying US$ 700 per month, almost US$ 1 million per month in recurring revenue (SurefireLocal CEO Chris Marentis).

Numbers that came up

  • Click.ai: revenue of about US$ 200 thousand in 2018 (80% service), US$ 500 thousand in 2019 and MRR of about US$ 100 thousand afterward (video).
  • Bazilla: service margin near 40%, versus about 80% for SaaS (video).
  • Trustmary: 208 paying customers, about 208 thousand euros in ARR, after a year with the software inside the service package (video).
  • Birdseye: about 80 paying agency clients, with a run rate of about US$ 1.2 million (video).

When it does NOT apply

  • If the client only wants the service: SurefireLocal lost two large enterprise clients who wanted agency service (video).
  • If the client does not pay for the pain before the software exists: consulting has a low margin and demands a lot of the team's time (video).
  • If you need quick results: Click.ai's transition took years (video).

Turn your internal tool or operation into a product you sell

29 founders used itin 6 of 7 channels23 with revenue cited

What it is

You build a tool to solve a problem in your own company or agency. When it works well and other people ask for it, you package it and sell it. Your own use already proves the pain exists.

Why it works

You know the problem is real because it was getting in your own way. Daily use shows what to improve before any outside customer shows up. And you start with at least one customer: your own company.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

When the problem is yours and the tool already solves it well. Even better if other customers or companies in the same industry have the same pain.

Step by step

  1. Use the tool in your own operation and measure how much it earns. Blue Stream Analytics measured almost 4 million in closed deals with it.
  2. See whether others ask for access or have the same problem. At Sweet CX, clients saw the result and asked for a license.
  3. Translate the tool for third-party use: remove personal shortcuts, build onboarding and test with outside teams (Up Content).
  4. Start by selling to people who already know you, through word of mouth, like Maximus.
  5. With about 20 customers using it, consider spinning it off into its own company, with its own team and resources (Invoke, Folderly).
  6. Keep the parent company as the first customer or partner, to license it back.

Best example

Instantly was born from a tool the agency already used. The agency had about 22 clients; the SaaS went from US$ 15 thousand in MRR in May to US$ 200 thousand in October, with about 2,900 customers (Young CEO of Instantly goes from $0 to $2.4m in 9 months).

Numbers that came up

  • Folderly: an agency with about US$ 6 million in revenue in 2020 and a SaaS with a US$ 1.5 million run rate nine months after launch (video).
  • Maximus: about 25 paying customers in under 12 months, from the agency's media buying team tool (video).
  • Visme: a tool that cost about 10 to 12 thousand dollars to develop, which reached 8 to 10 thousand per month and generated almost 1 million dollars in 8 to 9 years (video).
  • PodPlay: about US$ 3 million in contracted ARR, more than 200 venues, from PingPod's stack (video).

When it does NOT apply

  • If cash is tight: Up Content took about 8 months to transform the tool and another 6 to get it ready for teams, with zero revenue in the first year (video).
  • If the tool is the company's main product: Beehive's idea only works if the internal tool is the secret of growth (video).
  • If your own use proves nothing: Blue Stream spent about 100 thousand on iterations before it worked and still only they use it (video).

Differentiate a common product through look, format and packaging

22 founders used itin 5 of 7 channels13 with revenue cited

What it is

In a category full of identical products, you change the look, the format or the personality of yours. The formula can be the same, but the color, the packaging or the way it is used are different. The product becomes easy to remember and seems to be worth more.

Why it works

When everything on the shelf looks the same, the customer notices what stands out. A color nobody uses or a new format grabs attention without needing an explanation. The product becomes memorable and the price is easier to sustain.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

When the category is crowded and products look the same or dull. Most cases are physical products; in software, the equivalent is landing page, visuals and positioning.

Step by step

  1. Look at the shelf or at Amazon search and list what everyone does the same: color, format, name, ingredient.
  2. Choose one or two of those things and do the opposite. IM8 used a color no competitor had, red.
  3. Change the format if it makes sense, like Grüns, which sold gummies in a packet instead of a bottle.
  4. Shift the positioning to the category the customer already buys, like Poppi, which went from health drink to soda.
  5. In software, give the product personality (mascot, its own visual style). Figma ran a one-day design sprint and came out with 20 ideas for FigJam.
  6. Invest early in packaging and website, and make sure the product itself is good.

Best example

Grüns sold packets of gummies (8 per dose) instead of bottles of 60. It made US$ 30 thousand in the first month and US$ 230 thousand in the second, and reached a billion-dollar exit in 32 months, burning about US$ 8 million before profit (This guy sold his company to Unilever for $1.2B after just 3 years).

Numbers that came up

  • IM8: run rate of US$ 200 million in 18 months (video).
  • Poppi: about US$ 3 million in the first year and later more than US$ 500 million, according to the founder (video).
  • Cat on Chair: 127 thousand downloads and about US$ 18 thousand in 30 days, with illustrations as the base of the look (video).
  • Relay HQ: customers pay US$ 59 per month for a look that, according to the founder, seems to have cost thousands of dollars (video).

When it does NOT apply

  • If the product itself is weak: form alone is not enough, according to Eric Ryan on the Olly case (video).
  • If the market is already saturated: a new format guarantees nothing, says Grüns' founder (video).
  • In serious work tools: FigJam's playful tone does not fit Figma Design (video).
03

Building with AI

Write the product plan before telling the AI to build

19 founders used itin 3 of 7 channels4 with revenue cited

What it is

Before you open the coding agent or the app generator, you write a document (spec, PRD or plan) describing what will be built. The AI gets that document and executes on top of it instead of guessing.

Why it works

A clear document of what to build reduces agent errors and rework. Without it, the generator can't guess what you want and hands back something generic. With it, the first build already comes out aligned with what you had in mind.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

Before opening the agent to build an app or MVP that will be used for real. For a throwaway prototype or a learning project, you can go straight in.

Step by step

  1. Write the idea with screens, usage flows, who the user is and the main functions. Ask a chat (ChatGPT, for example) for a lean PRD, without choosing the stack.
  2. Describe the data: a JSON with the shape of each item and what each field means.
  3. Save the plan in the repository (project.md, plan.md) and turn on the agent's plan mode (Shift+Tab in Claude Code or Cursor), with the instruction "don't write code yet".
  4. Review and iterate on the plan until it is good. Cut whatever is not the main function.
  5. Ask for one task at a time and read the result. When you get stuck, take a screenshot and ask the AI what it is seeing.
  6. Accept imperfect versions and keep fixing.

Best example

Payout, a mobile app that reached US$ 20 mil per month. The founder wrote a document explaining the app and a JSON with the shape of the data, pasted screenshots of the screens into Claude or Cursor and built the latest app in less than 2 weeks, 100% with AI (I vibe coded a $20K/month mobile app in 14 days).

Numbers that came up

When it does NOT apply

  • Quick exploration prototype: you can improvise, and documentation comes in when the exploration ends (The rise of the professional vibe coder).
  • People who don't like writing specs: the OpenAI interviewee himself says the practice is for those who enjoy it (Inside OpenAI).
  • You need to be present on the scope, you can't send it off and forget it (Inside Devin).

Prototype with AI and test with users before spending engineering

14 founders used itin 5 of 7 channels4 with revenue cited

What it is

You generate the first version of the product with an AI tool, show it to real users and only bring in engineering (or hire a dev) after you see that someone wants it.

Why it works

A tested V1 becomes an argument for the roadmap. Engineering then improves what is already validated and you avoid building what nobody wants. The prototype costs days or weeks, not months.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

At the start, before investing engineering time or money in a freelancer, especially if the app is simple (CRUD, one main function).

Step by step

  1. Write a prompt describing the app: users, functions, admin panel, look. The more descriptive, the better the result.
  2. Ask the agent (Replit, Bolt, Lovable or Cursor) for the complete prototype, with database and login.
  3. Use Claude or ChatGPT to set up what you don't master, such as subscriptions, email and database.
  4. Publish and track signups in the first 24 hours. Test with a few users.
  5. If the prototype shows value, take it to engineering, or hire a dev only after you have revenue.
  6. If you are publishing on Apple, request the developer account and the review early, because the wait is what delays things most.

Best example

Scanemon, which the founder built with the Cursor agent in one day. Launched about 1 year ago, it has 3.400 active subscribers and about US$ 17 mil of MRR (I Built Two Apps That Make $120K/Month).

Numbers that came up

When it does NOT apply

Use a cheap model for repeated tasks and the strong one only to decide

16 founders used itin 4 of 7 channels1 with revenue cited

What it is

Instead of running everything on the most expensive AI model, you split the work into steps and pick a model for each one. The strong model handles planning and hard decisions, and the cheap or fast one does the repeated work.

Why it works

Repetitive tasks, or tasks with a capped upside, don't need the most expensive model. Cutting spend on those steps frees up budget and plan limits for where quality matters, and the product margin improves.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

When token spend or the plan limit starts to weigh, whether in your daily use or in a product you sell.

Step by step

  1. Break the task into steps (idea, outline, draft, edit, image, or plan, code, patch).
  2. For each step, compare the price per million tokens (OpenRouter helps) and test open and closed models.
  3. Pick the cheapest one that solves the step. Leave the frontier model for what has big upside.
  4. Use the strong model to plan and the cheap one to execute (for example, the most expensive one describes what to change and the cheap one edits the code).
  5. Use caching to hold down cost and retest from time to time, because model rankings and prices change.

Best example

On Hermes Agent, token spend dropped from about US$ 130 every 5 days to about US$ 10 every 5 days, a reduction of more than 90% while keeping the tasks. A Qwen model costs about 1/10 of Sonnet on input (Hermes Agent: The New OpenClaw?).

Numbers that came up

When it does NOT apply

Give the AI a reference image to control the visual style

13 founders used itin 3 of 7 channels0 with revenue cited

What it is

Instead of describing the look in words, you send a screenshot or a reference image along with the prompt. The AI copies the style and you adjust in short rounds.

Why it works

An image communicates style better than text. Someone who can only describe the look by example gets closer to what they want in a few exchanges, and without the reference the first version comes out generic.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

When you want an interface with a look you can show but can't name. It works for front-end and single-function apps, with no backend.

Step by step

  1. Find a site or app with the look you want and take a screenshot (or draw the screen in Figma, on paper or on an iPad).
  2. Upload the screenshot to the tool (v0, Lovable, Claude Code, AI Studio) along with a prompt saying what the app does.
  3. If you are copying from another page, say: "don't use the content, only the style". Without that, the model copies the text too.
  4. Ask for one change at a time, in plain language, as you would with a designer (font, color, density).
  5. If you don't know the name of the style, research it first (he found "glassmorphism" on Perplexity) and write the term in the prompt.
  6. Check the text before accepting and take the approved design to Claude Code.

Best example

Snag (10X Studio): the founder starts from the value proposition of the page, draws the screen in Figma and passes the design to Claude Code. With no backend, it takes 4 to 5 hours to get a good app, and he says he built more than 40 apps in the last year with this process (I Built A $30K/Month App: Here's My Exact Process).

Numbers that came up

When it does NOT apply

04

Pricing

Charge for the year (or several years) upfront to pay for acquisition

81 founders used itin 4 of 7 channels66 with revenue cited

What it is

Instead of a monthly fee, you offer an annual or multi-year plan and collect everything at signup. It usually comes with a discount of about two months. The cash arrives before the service is delivered.

Why it works

The whole year's money lands on the day of the sale and covers the cost of acquiring the customer. With CAC paid on day one, you reinvest right away and grow without raising capital. It also locks the customer in for longer.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

The customer accepts an annual contract and CAC is high (B2B, ticket in the thousands of dollars). Works for anyone who wants to grow without investors.

Step by step

  1. Create an annual plan with about 2 months free (16% to 23% discount).
  2. Charge the full year at signup, with a short payment term (e.g. 30 days).
  3. Calculate payback: CAC against the amount received on day one.
  4. Use the upfront cash to pay for acquisition or a hire.
  5. In campaigns (Black Friday, affiliate coupon), push only the annual plan.
  6. If the customer doesn't trust you yet, accept a pilot at the start and close upfront once you have known customers.

Best example

Park Bench sells an annual plan of 4,500 to 5,000 dollars paid upfront, with a CAC of about 650 dollars. In one month they sold to about 150 real estate agents (vídeo). Predictive Index recovered 12 thousand of its 16 thousand CAC on the spot (vídeo).

Numbers that came up

  • WP Engine: 1 in every 4 signups picks the annual plan, and monthly cash went up more than 3 times (vídeo).
  • Blue Rhythm: four-digit MRR, but the cash coming in went past five digits (vídeo).
  • Homemaker: CAC of about 5 thousand dollars, ticket of 1,250 per month, payback of 4 to 5 months with annual billing upfront (vídeo).
  • Pixlee: more than 80% of deals paid upfront, with payback accepted between 6 and 12 months (vídeo).

When it does NOT apply

  • A niche where the customer won't pay annually, or the early stage, when they demand a pilot because of low credibility.
  • Weak renewal: at Park Bench it sits at 60% to 70% per year, so the return depends on the second year.
  • Black Friday concentrates renewal and churn in the same month; Nathan Latka (ConvertKit) said he doesn't know if it was a good idea.

Charge per unit of usage consumed

93 founders used itin 4 of 7 channels80 with revenue cited

What it is

The price follows a usage measure the customer understands: task, API call, order, click, session. Whoever uses more pays more, usually with volume tiers.

Why it works

The bill rises along with the value the customer gets and with your cost. When the customer grows, their spend grows without you renegotiating or sending a salesperson. Expansion becomes automatic.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

Usage is easy to measure and grows when the customer succeeds. Works well in API, logistics and e-commerce.

Step by step

  1. Choose the unit that shows value to the customer (delivery, order, query, session).
  2. Test the price range before fixing it: Intercom went through US$ 10, US$ 5 and US$ 2.50 until it landed on US$ 0.99.
  3. Build volume tiers, with a lower price per unit for those who use more.
  4. Compare with what competitors charge.
  5. Bill after usage, without requiring prepaid credit, if the customer uses the API all the time.
  6. Use the same unit to sell new use cases to existing customers.

Best example

Onfleet charges per delivery task, on average about 20 cents, with volume tiers and without taking a percentage of the order. That is about 900 customers paying an average of 900 dollars per month and more than 3 million tasks in the last month (vídeo).

Numbers that came up

  • Billbee: swapped modules for a single price based on number of orders, and ARPU rose from about 17 to about 32 dollars in seven months (vídeo).
  • Inbenta: 25% expansion on the old base, gross churn of 5% and net retention of 120% (vídeo).
  • Seon: 4 to 8 cents per API call, about 90 million calls in the month and an average ticket near 55 thousand dollars per year (vídeo).
  • PixelMe: net revenue churn of about 0% since the price change, 4 months ago (vídeo).

When it does NOT apply

  • Low-volume customers pay much more per unit, which can push small customers away (Onfleet).
  • Part of Billbee's ARPU gain came from a jump in orders during Covid and Christmas, not just from price.
  • Price has to come from value, not from cost; margin swings as cost falls or rises (Intercom).

Charge per unit of the customer's business (store, location, asset)

83 founders used itin 2 of 7 channels73 with revenue cited

What it is

You charge per vehicle, point of sale, connection, account or space of the customer, usually monthly. Each new unit they add increases the invoice.

Why it works

When the customer's operation grows, your revenue grows with it, with no new sale. The unit also makes pricing predictable, which helps when usage volume is unpredictable.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

The customer has several physical or digital units and keeps adding more over time (fleets, store chains, hospitals).

Step by step

  1. List what grows when the customer grows (vehicle, store, connection, court).
  2. Choose the unit and set a monthly value for it (e.g. 1 dollar per active vehicle, 25 to 30 per point of sale).
  3. Give a volume discount for large fleets (Coaster gives 25% to 30%).
  4. If each customer needs its own installation, charge a one-time setup fee per location.
  5. If you want to offer tiers, split them by delivery type: software only, software with hardware, full package.
  6. Watch concentration: don't let one customer become too large a share of revenue.

Best example

Zeliot charges 1 dollar per active vehicle per month, with about 100 thousand vehicles. Recurring revenue went from 48 thousand to about 100 thousand dollars per month in one year (vídeo).

Numbers that came up

  • Redox: about 1,000 dollars per healthcare system connection, customers paying 2 to 3 thousand per month (vídeo).
  • Raw Mango: 25 to 30 dollars per point of sale per month, plus 1,000 to 2,500 for setup, across more than 500 locations (vídeo).
  • Sendzilla: charges per connected LinkedIn account; the biggest customer has 400 accounts and pays about 11 thousand per month (vídeo).
  • Productsup: charges by number of products (SKUs), with a ticket of about 3,700 per month and net retention of 120% (vídeo).

When it does NOT apply

  • A giant customer becomes a risk: if they leave, revenue drops a lot (Sendzilla).
  • An institutional market with a high minimum doesn't suit a low-ticket product (Market Factory).
  • Recurring revenue swings when the unit count varies, like one-off events (Topia) or a variable fleet (Coaster).

Charge a percentage on each transaction (take rate)

57 founders used itin 4 of 7 channels44 with revenue cited

What it is

You keep a percentage of the money that flows through the platform: payments, tickets, sales, ads. Often there is no monthly fee.

Why it works

The customer only pays when they get paid, so the barrier to entry is low. The more they sell, the more you earn, and revenue tracks volume.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

The platform processes payments or handles the customer's funds, and you want to attract people who wouldn't accept a monthly fee (occasional events, small sellers).

Step by step

  1. Define the fee on volume (the examples range from 1.9% to 5% in software and up to 15% in other cases).
  2. Leave the card cost to the acquirer and charge only your part.
  3. For large accounts, negotiate a fee discount (ticketinghub goes down from 3% to about 1.5%).
  4. If there are occasional customers, charge per ticket or transaction, with a cap (Eventzilla: 1.9% plus 0.99 dollar, capped at 9.99).
  5. If you want predictability, combine a fixed retainer with a percentage (Agency Sparks, Marlin).
  6. Track volume: if transactions fall, revenue falls with them.

Best example

Paddle charges a flat fee of about 5% on volume, with no monthly fee and no separate service. With about 900 customers processing more than 200 million dollars per year, it reaches about 10 million in ARR (vídeo).

Numbers that came up

  • LeagueApps: about 4% of GMV, more than 3 thousand leagues and 1 billion dollars in GMV over nine years (vídeo).
  • ticketinghub: 3% on ticket sales, 250 to 300 customers, about 50 thousand per month (vídeo).
  • Flexi: 2.9% per salary advance; about 70% of revenue comes from this fee (vídeo).
  • Resonant: royalty of 7 to 15% of the average device price, versus 1 to 2% from large licensors, according to him (vídeo).

When it does NOT apply

  • At very high scale, 5% may stop making sense for the customer (Paddle).
  • The model is exposed when transactions fall, as the pandemic showed at LeagueApps in 2020.
  • Seasonal: 30 to 35% of Eventzilla's organizers don't come back the following year, and that is still not truly recurring revenue.

Raise the price in steps, measuring conversion, while demand holds

32 founders used itin 7 of 7 channels18 with revenue cited

What it is

You start cheap or test a price, and raise it in stages while watching conversion and churn. You stop when resistance shows up. Several founders report that almost nobody complained.

Why it works

If almost nobody says no, the price is too low, so doubling it measures the ceiling. A low price also attracts low-quality customers and signals that the product is worth little. A higher price attracts people who really have the problem.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

Price is too low for the value delivered, in software, consulting or advertising. Works best when the customer is a company that needs the result.

Step by step

  1. Choose a step (e.g. 29 to 39, then 79, then 99).
  2. Change the price and watch conversion and churn at that step.
  3. If nothing gets worse, raise it again.
  4. If someone says no, go back a little (the search works like binary search, according to Ben Orenstein, from Tuple).
  5. Stop when resistance appears or when the customer profile stabilizes.
  6. If the cheap plan has to exist, create a Pro plan for those who pay more.

Best example

At Gather, the lowest plan went from 29 to 39, 79 and 99 dollars with no increase in churn and no drop in conversion. Today it has plans at 165 and 290 dollars, at its fastest growth pace (vídeo).

Numbers that came up

  • Autoklose: from 9.99 to 49.99 dollars in about 7 months, with about 3 thousand customers and MRR of about 81 thousand (vídeo).
  • Lately: ticket rose from 99 to about 300 per month, and the higher price brought back about 20% of the customers who had left (vídeo).
  • Lynx: the increase from 4 to 5 dollars generated about 5 thousand more in MRR (vídeo).
  • YNAB: launched at 9.95 dollars without selling; at 20 dollars, with the same page, the first sale came the same day (vídeo).

When it does NOT apply

  • Hobbyist product: the increase makes the customer leave (Tweet Hunter).
  • Letting the customer pay what they wanted at the start attracted low-quality people (Autoklose).
  • There are cases with only a few months of data or no churn by tier (Getsorale, Prentus), so confirm with your own numbers.

Pricing tiers by customer size

51 founders used itin 3 of 7 channels47 with revenue cited

What it is

The price depends on the size of the customer: revenue, number of employees, members, sites or visitors. Everyone inside the company can use it, with no per-seat license.

Why it works

The customer's size grows with the business, so revenue rises without a new sale. With no per-user charge, the product spreads inside the company without friction.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

There are many users per customer and the value is tied to the customer's size (companies, communities, agencies, store chains).

Step by step

  1. Choose the size measure the customer knows how to answer (revenue, employees, software spend, visitors).
  2. Create tiers with rising prices (e.g. 500 per month up to 100 employees, 2,500 per month at 500).
  3. Let all employees use it, with no per-seat license.
  4. Lower the price per person for larger customers.
  5. For a large customer, offer a 3-month pilot at a lower price.
  6. Review the price every year against the customer's current size.

Best example

Forecaster.co measures the customer's monthly expenses (3-month average) and charges about 3% of that on the annual contract. Someone spending 500 thousand per month pays about 15 thousand per year. That is 550 customers (vídeo).

Numbers that came up

  • Saleswise: 500 dollars per month for up to 100 employees, 2,500 per month for 500 (vídeo).
  • Wigzo: stores with 2 to 20 million pay about 1,500 per month; those with 20 to 100 million, 100 thousand to 200 thousand per year (vídeo).
  • Fable: annual license of about 60 thousand dollars per group, with about 50 paying accounts (vídeo).
  • SmartCSM: ticket of about a thousand dollars per month, annual churn of 5 to 6% and NRR of about 114% (figures from memory) (vídeo).

When it does NOT apply

  • Small samples with no churn data (Saleswise, ClickIt, Wigzo with a new price still being tested).
  • In a weak season the customer pays the same for less usage (Data Duopoly).
  • Agencies and media don't accept an annual license; they pay per campaign with a guaranteed minimum (Fable).
05

Communities

Help without selling in other people's communities to find your first customers

45 founders used itin 6 of 7 channels29 with revenue cited

What it is

You join Reddit, Facebook, Discord or Slack groups where your audience already talks about the pain. You help for real, without pushing a product. The offer comes later, once trust exists.

Why it works

People with the pain are already talking about it in public. If you answer with real help, you become a go-to person in that group. When the product shows up, it arrives as a solution from someone they know, not as an ad. Result: first customers without paying for media.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

From zero to $10k MRR, when your audience already gathers in open communities and the community accepts founders taking part. Works best in a small, engaged niche.

Step by step

  1. Find the servers and groups in your niche (discboard.org or Discord search; on Reddit, the list of suggested communities in Reddit Ads targeting, even if you never pay for a campaign).
  2. Join and observe. On a new Reddit account, only comment and vote for 7 to 14 days before posting.
  3. Look for questions that show demand and answer with real help, using your tool if it fits, without mentioning the product.
  4. Send a private message to people who described exactly the problem you solve.
  5. Post real value (a case study, full content in the post itself) and leave a small product link at the end.
  6. Repeat 2 to 3 times a week at most, and avoid groups that ban promotion.

Best example

Algrow stayed around US$ 14 thousand per month, with more than 10 thousand users. The first 400 users came from Discord: the founder joined the voice channel, shared his screen using the tool, helped for about a week and only then posted the waitlist (vídeo).

Numbers that came up

  • Alchemi app: US$ 17 thousand per month, more than 1,000 paying customers and about 20 thousand signups from Reddit, with zero marketing spend (vídeo).
  • Gojiberry AI: 11 million impressions on Reddit and about 100 first customers (vídeo).
  • Batch Leads: first customers came from Facebook groups, one of them with about 120 thousand members; today they have about 18 thousand customers (vídeo).
  • Relay HQ: CAC of US$ 35 to 40 per customer paying about US$ 80 per month, with payback in the first month (vídeo).

When it does NOT apply

  • Many servers and subreddits ban self-promotion. Without a relationship built beforehand, the product post looks like spam (vídeo).
  • Mass market: the effect of the small, passionate group disappears (vídeo).
  • Groups have a ceiling. Airfocus now buys traffic through ads and partnerships (vídeo).

Answer public questions about the problem, using mention alerts

11 founders used itin 6 of 7 channels7 with revenue cited

What it is

You set up keyword alerts to get notified when someone asks about or complains about the problem your product solves. Someone on the team answers with real help and mentions the product only afterward. You can start alone and for free.

Why it works

People who ask in public are already looking for a solution. A useful answer, given right away, builds trust and stays visible to anyone who arrives later. This produces steady signups and authority, without ads.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

At any stage, as soon as there are public conversations about the problem and someone on the team can answer. Deel did this from day one.

Step by step

  1. List 5 to 15 keywords for your audience: product name, company name, your own name, competitors and niche terms.
  2. Set up alerts: F5Bot (free) for Reddit and Hacker News, and a saved search on Twitter. Google Alerts lags by 12 to 24 hours.
  3. Jump into the thread and give the real answer first (yes, it can; no, it can't; but consider X and Y).
  4. Only then say you are the founder and that the person can talk to you, or mention the tool if they ask for one.
  5. Answers come from people on the team, with no automation.
  6. Also look for conversations that rank well on Google (forums, Quora, Stack Exchange) and answer there; the answer climbs in search over time.

Best example

Kickoff Labs: 10 hours answering questions on Quora generated more than US$ 35,000 in sales in a year and a half (vídeo).

Numbers that came up

  • Brand24: more than a thousand signups per month from this channel, with one dedicated person as the main cost (vídeo).
  • Deel: the channel started at about 15% of the funnel and fell to under 5% as the rest of the funnel grew (vídeo).
  • Save Wise: went from thousands to tens of thousands of users and reached about US$ 25 thousand per month (vídeo).
  • B2B SaaS analysis: Reddit outranked all vendors on 50% to 66% of shared keywords (vídeo).

When it does NOT apply

  • There is an audience ceiling: a subreddit of about a thousand people will not bring thousands of customers (vídeo).
  • A post that is only advertising ("check out our site") gets the person kicked out of the community (vídeo).
  • The founder of Missive doubts that Twitter search is the main engine; he thinks word of mouth generates more (vídeo).

Your own group of active customers for feedback, advice and support

11 founders used itin 6 of 7 channels2 with revenue cited

What it is

You gather your most engaged customers or users in a space of your own (Facebook group, WhatsApp, Slack, weekly webinar, advisory board). The team listens and answers directly there.

Why it works

Your most active customers tell you what is missing and what breaks, around the product, with no survey. With fast feedback, the team adjusts before competitors do and customers feel like owners of the product. This produces a better product and more loyal customers.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

When you already have a user base and the team can respond often. In products where the customer is also the audience (creators, coaches), it pays off even more.

Step by step

  1. Pick the space: Facebook group, WhatsApp, Slack, weekly call or webinar.
  2. Invite the most active people (and, if you want, people who did a trial and did not convert, or people who like the product without paying).
  3. Answer personally, every day at first; later the team takes over.
  4. Paste requests and comments into a shared document and count the most repeated ones, so you have numbers and real phrases from users.
  5. Take the idea to the team, show a mockup at the next meeting and repeat the cycle every few weeks.
  6. When the group grows past thousands of people, move from WhatsApp to Discord or Reddit.

Best example

WebinarNinja: weekly webinar with users, mockup in Balsamiq and a feature launched within a few weeks, with a team of 30 people against 500 to a thousand at competitors. They reached 30 thousand paid users, according to the speaker (vídeo).

Numbers that came up

  • Fellow: a group of leaders with about 2,000 members, plus an academy with classes and webinars (vídeo).
  • Goodreads: a feedback group with almost 20 thousand people; the founder wrote more than 4 thousand comments in it (vídeo).
  • Zest: a user board with 64 people and 42 paying among the first contacts (vídeo).
  • Base44: a WhatsApp group with thousands of people, where the founder saw bugs and slowness before any other channel (vídeo).

When it does NOT apply

  • It eats the founder's time; it only scales when the team takes over (vídeo).
  • It does not scale: Zest's own founder calls it "doing things that don't scale" (vídeo).
  • WhatsApp can't handle a very large group, which is why Base44 moved (vídeo).

Small group of peers at the same stage with fixed meetings (mastermind)

9 founders used itin 4 of 7 channels3 with revenue cited

What it is

You bring together 4 to 10 founders with similar revenue and team size and schedule recurring meetings in a fixed format. Each person brings a real problem and the others give direct advice.

Why it works

People at the same size face similar problems and have tested tactics to swap. The fixed format (agenda, time per person) gets everyone talking about real things. Over time trust builds, and with it learning, network and groups that last for years.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

When there are people of the same size nearby and someone takes on the agenda and facilitation. Works at any stage, as long as nobody in the group is far ahead or far behind.

Step by step

  1. Gather 4 to 8 founders in your revenue and team range (one step ahead or behind works). If possible, from different niches that do not compete with each other.
  2. Set the rhythm (biweekly or monthly) and the format: hot seat (one person per meeting, 45 minutes, one problem) or round table (equal time for everyone).
  3. Ask the presenter to bring numbers and the question of the moment (for example CAC, acquisition channels, what is going wrong).
  4. Open a chat channel for the gaps between meetings.
  5. Hold a 3 or 4 day retreat one or more times a year, with a quarterly review agenda and plans for the year.
  6. Agree on a time frame (something like a year) and leave respectfully if someone has moved far ahead and the group cannot keep up.

Best example

Hampton: groups of 6 to 10 founders with a facilitator, where one member per meeting presents numbers and gets questions. They have 30 to 40 facilitators and run about 200 events per year; the site cites an average of US$ 23 million in annual revenue among members (vídeo, vídeo 2).

Numbers that came up

  • Rob Walling's group: one of his groups lasted 15 years and still has monthly conversations (vídeo).
  • Nik Sharma: a monthly group of about 10 e-commerce founders produced an idea that earned about US$ 100 thousand the following weekend (vídeo).
  • SaaStock: in 2022 the event gathered about 3,700 people, including 1,500 founders; today there are mastermind groups tied to it, with annual retreats of about 40 founders (vídeo).
  • Fresh Founders (Ottawa): founders met every Friday and companies like Shopify, Fluid Surveys and Fellow came out of it (vídeo).

When it does NOT apply

  • A founder far ahead (for example US$ 2 million ARR) needs mentorship, not a mastermind (vídeo).
  • Scaling is hard: recruiting and training facilitators is the biggest cost, according to the Hampton host (vídeo).
  • A group mixing 6 to 7 figures of revenue with people at their first customer did not work (vídeo).
06

Content and social

Build in public by showing the process, the journey, and the milestones

34 founders used itin 7 of 7 channels20 with revenue cited

What it is

You post what you are building (drafts, mistakes, revenue numbers, milestones) from before the product is ready. The people who follow along become your first customers, helpers, or useful contacts when launch arrives. The channel is usually Twitter/X, a blog, LinkedIn, or Instagram.

Why it works

People who watch the process up close come to trust the builder and feel part of the story. That trust turns into sales at launch, with no need for ads. The audience also brings help and invitations (podcast, accelerator, volunteer devs) that a one-off launch post never would.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

Solo founder, small product or no product yet (0-10k MRR), willing to post every day for weeks. It works best when your target audience is on the network where you post.

Step by step

  1. Pick a network where your customer really is, and open the account already themed around your product.
  2. Post every day what you built, with a real number: revenue, subscribers, mistakes, dashboard screenshots.
  3. Announce a measurable goal in the very first post (e.g., recurring revenue in X months) and update it live.
  4. At launch, talk to the people who engaged and offer them the product. If you post the link, put it in the second tweet, not the first.
  5. When you need help or need to hire, tell your audience in the same space.
  6. Give it at least about 50 days of consistency before judging the channel.

Best example

Potion showed the product's process on Twitter every day and closed 79 paying subscribers in about 4 months, with 85% of customers coming from Twitter (https://youtu.be/F5F5iEmrzvY).

Numbers that came up

  • Shipper: after about 50 days of building in public, X brought about $20K of MRR in 1 to 2 weeks (https://youtu.be/Nnpz1wsTjBI).
  • Maxed AI: about a third of revenue comes from people who follow the journey on X, within 3 months (https://youtu.be/aLidq3_FwW0).
  • Breakcold: 5 paying customers the week after the MVP, going from zero to about 7 thousand followers (https://youtu.be/VelbzA51ILI).
  • ConvertKit: with a few thousand emails on the list, 25 Ruby developers offered to help after blog posts (https://youtu.be/hCRMzMnAb40).

When it does NOT apply

  • If your target audience is not on the network where you post: Breakcold only got curious people and free trials on Twitter, and today acquisition comes from email and LinkedIn (https://youtu.be/VelbzA51ILI).
  • If you expect fast results: Ness Labs says it took time, and Nathan Barry missed his goals and took much longer than expected (https://youtu.be/clD41jX5lNA, https://youtu.be/sZGbQvX9AXY).
  • When the audience already came from years of publishing, as with Juggle, you cannot attribute everything to one post (https://youtu.be/n40ZXWBbOy8).

Build your own audience before the product and launch to it

32 founders used itin 6 of 7 channels11 with revenue cited

What it is

You publish useful, free content (Twitter, blog, podcast, newsletter, YouTube) for your future product's audience for months or years. When the product ships, the first sale happens inside that audience. You can also test the idea with them before building.

Why it works

The audience already knows and trusts you, so the cost of finding the first customer practically disappears. They also respond fast on whether the idea interests them, which validates it before you spend time coding. The launch starts with people in the room, and the cold start gets easier.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

When you have time to publish for months or years with no return and can define a clear topic for the audience. It works for a subscription product, a tool, an app, or even a paid guide.

Step by step

  1. Define the audience and a clear thesis about their problem (e.g., "boring marketing").
  2. Publish useful things for free on a recurring basis, on the channel where that audience is.
  3. Collect emails or subscribers (list, newsletter, community) so you have a direct channel.
  4. Before building, offer the idea or a simple version (even a spreadsheet) and see if people ask for it.
  5. Launch to this audience first, on day one, before any ads.
  6. Only open billing to companies or the paid subscription when the audience is already large.

Best example

Ben Orenstein, from Tuple, spoke for free for years in talks, podcast, Twitter, and newsletter (about 10 thousand followers, a few thousand on the list) and says almost all of the first customers came from MicroConf, the podcast, and Twitter (https://youtu.be/C4WOfEyogo8).

Numbers that came up

  • Magi: 10 years of blogging and over 100 thousand subscribers; with no ads, went from 3 thousand in the first month to 30 thousand per month in 1 year (https://youtu.be/DV7bBCAABg4).
  • CoSchedule: about 10 thousand people on the list on launch day, before writing any code; today more than 8 thousand customers (https://youtu.be/eNJ9jSVRSQY).
  • Marketing for Developers: 75% of the book's buyers were listeners of the author's podcast (https://youtu.be/TnWo6ex1C4E).
  • Lenny's Newsletter: 9 months free before charging; with 1,000 paid subscribers, revenue came to close to 100k (https://youtu.be/zKP2HrMc23s).

When it does NOT apply

  • People who do not have time for months of publishing with no return. One of the interviewees says that, at the start, it is shouting into the dark (https://youtu.be/FOdkglCm_oM).
  • Slow results: the Boring Marketing community earns from zero to US$ 100 per month at the start, and the compounding effect comes later (https://youtu.be/ndqX4vbR7Rc).
  • An audience does not guarantee purchase. Noah Kagan himself fears the audience will not buy (https://youtu.be/OmS6UcxM8rQ).

Post every day for months to pile up attempts until something hits

31 founders used itin 5 of 7 channels12 with revenue cited

What it is

You pick a niche and a short format, and publish every day (or almost) for weeks and months, without waiting for the perfect post. On a new TikTok or Instagram account, you watch and interact for a few days beforehand. The daily volume discovers which post hits.

Why it works

Nobody can predict which post will go viral, so more attempts give more chances to find the one that hits. Peaks usually come after dozens of posts. On a new account, watching and commenting before posting helps keep the platform from limiting your reach (shadowban).

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

New account or small audience, with a fixed routine of a few hours a day. Worth it when you can hold on for at least 30 to 90 days before evaluating.

Step by step

  1. Pick a specific niche, take care with the name and the bio, and follow only accounts in the niche.
  2. On a new account, warm it up: watch and comment for 15 to 30 minutes a day for about 3 days (or a week without posting).
  3. Post one video or text per day. If you post more than one, space them 3 to 4 hours apart and avoid a scheduler on new accounts.
  4. Set aside 45 minutes to 1 hour a day to reply to the audience.
  5. Look at the data for each post and repeat what produced results.
  6. Set a minimum deadline (30 days, 50 posts, or 90 days) before switching strategy.

Best example

Nuseir Yassin, from Nas Daily, made a thousand videos in a thousand days. It took 270 videos for one to blow up, and today he has 47 million followers in 13 languages (https://youtu.be/YIyipW04NSs).

Numbers that came up

  • One thread a day for 50 days took Twitter from 500 to 40 thousand followers and the newsletter from 2 thousand to 8 or 9 thousand subscribers (https://youtu.be/5MaPEK-yVVY).
  • One video with about 48 million views generated 12.5 thousand in sales in one day, with 60 thousand downloads, on the Glow Up app (https://youtu.be/LYdJ3dfNCVs).
  • Natural: a video a day on their own accounts, with revenue of 41 thousand, 32 thousand, and 27 thousand over the three months, and the first sale coming from the first video (https://youtu.be/1nfPvJKrYYQ).
  • Clip.co: one billion combined views in a year of daily shorts and more than 1 million subscribers on YouTube (https://youtu.be/e6COsh5wfNA).

When it does NOT apply

  • If you cannot handle the emotional cost: Nuseir says it was high and that he considered therapy (https://youtu.be/YIyipW04NSs).
  • Some posts never take off. The owner of Mylands says you cannot predict viral (https://youtu.be/gc3mX99WnaU).
  • Many cases are just the interviewee's word, with no published data, and Andy Raskin reminds us that most people try the single post that blows up and almost always fail (https://youtu.be/dkVJnaxDlXE).

Repurpose a recording or content that worked into several formats

33 founders used itin 5 of 7 channels5 with revenue cited

What it is

You record once (interview, livestream, call, podcast) or take a post that worked, and turn it into several pieces: article, thread, LinkedIn post, shorts, newsletter, book chapter. AI and an editor help with the transformation, and you review before publishing.

Why it works

The expensive work is having the idea and the conversation, and you pay for that once. Each new format costs only the conversion, so you get more pieces per hour of effort. Starting from something that already got a response also reduces the risk that the new piece will not perform.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

When you already have long recorded content or tweets and posts with good response, and want presence on more channels without adding many weekly hours.

Step by step

  1. Pick one main channel and record on it with a fixed routine (e.g., one interview per week, at the same time).
  2. Generate the transcript of the audio or video.
  3. Use a saved prompt for each format (article, thread, newsletter, description with timestamps) and let the AI do the draft.
  4. Review everything by hand before publishing, because the AI repeats phrases and gets the tone wrong.
  5. Publish on the other channels, cutting the long video into shorts and pulling the best tweets over to LinkedIn.
  6. See what got the most response and use it as the base for the next round.

Best example

At Hubstaff, Andy Baldacci interviews one expert per week (about 70 episodes), the 60-minute interview yields about 10 thousand words, and his work comes to about one hour per week, with outsourcing of US$ 1 to 2 thousand per month (https://youtu.be/vo2k4dRf61U).

Numbers that came up

  • LinkedIn went from 10 thousand to 82 thousand followers in one year, repurposing tweets that got a good response (https://youtu.be/mWXOSNq6b8E).
  • An account with 2,700 followers got 400 thousand views in 7 days on LinkedIn with content ported from Instagram, TikTok, and Shorts (https://youtu.be/tiKnAnJuswo).
  • Clip.co: illustrated newsletter built from the shorts, launched with about 10 thousand subscribers in the first month and about 20 thousand in 3 weeks (https://youtu.be/e6COsh5wfNA).
  • A 3-hour task (listening, writing, and marking timestamps) went down to a few minutes with the transcript in Claude (https://youtu.be/FqHH_rXdsMI).

When it does NOT apply

  • Without human review the result comes out generic: the AI becomes a draft, and the angle and delivery have to be yours (https://youtu.be/HTjvpTR3VgQ).
  • It only works if the starting idea is good (https://youtu.be/HhspudqFSvU).
  • If you do not yet have material that performed, the road is much longer: for a book coming from a newsletter, starting from zero would take about two and a half years (https://youtu.be/XXtSaM_AZ3Q).

Study and copy a format, hook, or topic already validated in the niche

24 founders used itin 4 of 7 channels11 with revenue cited

What it is

You look for videos or posts that already got lots of views in your niche, note the hook, structure, and length, and redo the format with your product or content. You can copy the structure and swap the subject and the face. Not the text.

Why it works

A format that already earned many views for someone else has a good chance of earning them again, because the platform and the audience have already shown they like it. You skip the guessing phase and spend your time on what changes: your topic, your product.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

When there are already viral videos in your niche to study, and you have a low creation cost (phone, 30 minutes per video). Worth it for a new account and for an impulse-use app.

Step by step

  1. Spend a day scrolling the feed and save the most-viewed videos in the niche (on TikTok, search the niche keyword and filter by most liked).
  2. Build a spreadsheet with the hook, story, length, and call to action of each one.
  3. Redo the format with your product: visual or audio hook in the first 3 seconds, entertainment first, product in the middle, 2-second CTA.
  4. Test one copied format per day and repeat until you get it right. One of the interviewees talks about 30 to 40 attempts, sometimes 100.
  5. For YouTube titles and thumbnails, look at channels a bit bigger than yours and use their "most popular" as a reference.
  6. When a topic works, repeat the topic (double down).

Best example

Puff Count built a spreadsheet with the most viral videos in the niche, did entertainment first, showed the product in the middle, and closed with a 2-second CTA. One video with 8.3 million views brought tens of thousands of downloads, and the app reached about US$ 40 thousand in MRR (https://youtu.be/loXc0Tyi4R4).

Numbers that came up

  • A copy of a viral video got 1.5 million views against the original's 150 thousand, on the Natural app (https://youtu.be/1nfPvJKrYYQ).
  • Ali Abdaal took a title from a video with 2 million views on a channel with 20 thousand subscribers, and his video got 6 million (https://youtu.be/EBAqxaJ1RMw).
  • A new UGC account reached 10.8 thousand followers in a few months, and the app reached about $11 thousand in MRR with 400 payers (https://youtu.be/QXxESN94Y6s).
  • A format validated in another niche produced a video with about 2 million views in 30 posts, on a food app (https://youtu.be/9tOmppsiO2w).

When it does NOT apply

  • High reach does not guarantee sales. A video with 11.6 million views brought almost zero installs because it had no CTA or clear pain point (https://youtu.be/hKqLceTPLWM).
  • A video that turns into a commercial does not work on TikTok, according to the founder of Puff Count (https://youtu.be/loXc0Tyi4R4).
  • A copied title has to be true to the content, otherwise retention drops (https://youtu.be/V6LnE7cXCJI).

Pay UGC creators and micro-creators to recreate a validated format showing the product

22 founders used itin 4 of 7 channels14 with revenue cited

What it is

You hire small creators (UGC) to record videos in the format that already worked, showing the app or product in use, and post on multiple accounts. Payment is usually a fixed monthly fee, sometimes with a performance part. You track which format performs and repeat it.

Why it works

Each creator and each account reaches a different bubble of the audience, so you do not depend on a single account. Video that looks like normal content gets watched, and video that looks like an ad gets ignored. With many cheap videos, volume shows which hook converts.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

When the format already works (you tested it with your own content) and there is a monthly budget to pay creators. It works best for a phone-use product, with impulse purchase, and a young audience.

Step by step

  1. Find small creators who already post in your niche (search and DM) and talk to each one to see if they have the capacity to go viral.
  2. Agree on a fixed monthly amount per package of videos (e.g., 3 or 10 videos instead of 1), provide screens and videos of the product, and ask for one video per day with a new hook.
  3. Hand over the format that already went viral and ask the creator to recreate it, showing the product in use before any download request.
  4. Track each format and sub-format. Keep the creators who perform and drop the weak ones.
  5. Make up the variable part by conversion (coupon or UTM), with a bonus if the video passes a certain number of views.
  6. Open new accounts and repeat the format with other faces. Give creative freedom: directing too much kills reach.

Best example

The founder of Jenny AI paid 4,000 dollars for 20 videos to a niche creator. In the first month the account had 7 million impressions, reached 70 thousand followers and, adding other accounts, about half a billion views (https://youtu.be/SJFs-fZ3HjI).

Numbers that came up

  • Graph: 900 dollars buys about 500 thousand impressions from a creator (CPM about 2 dollars), against about 150 thousand on Facebook (CPM about 6) (https://youtu.be/Px_X-qBQ18M).
  • Symmetry: almost 80 thousand videos, more than 500 accounts, and almost 1 billion views in the last year, with part of the revenue shared with creators (https://youtu.be/OV09yPLf8BI).
  • Margin of 50 to 80% on in-house UGC, against 25 to 70% with influencers and 0 to 30% with paid ads (https://youtu.be/5u9u8yzPEpA).
  • One creator received about 120 dollars for a video that got 2 million views and brought in tens of thousands of dollars (https://youtu.be/9WWvLj-NqEE).

When it does NOT apply

  • High views do not guarantee sales: a video with 20 million views barely moved the day's conversion (https://youtu.be/SJFs-fZ3HjI).
  • If the product requires a long learning curve or a high-consideration purchase, it does not work, and managing creators is work because many are unreliable (https://youtu.be/Px_X-qBQ18M).
  • Seeding without a sharing loop does not sustain growth (https://youtu.be/hyosynoNbSU), and Cali AI stalled after the peak with only influencers (https://youtu.be/wgTOoJ_UkmQ).
08

Product-led growth

Publish on the marketplace of the platform where your customer already works

55 founders used itin 5 of 7 channels41 with revenue cited

What it is

You put your app, theme, or plugin in the store of a big platform (Shopify, Slack, Salesforce, GitHub, Google Workspace, monday.com). People who already use the platform search for solutions there and find you. You start with a ready-made customer base, without selling from scratch.

Why it works

Marketplace traffic comes from people who already have the problem and are already inside the tool. Buying intent is high. Whoever ranks first in the category gets signups every month, without paying for ads.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

The platform has lots of customers and an app store where competition is still manageable. It works best early, when the category is empty, and for the first specific pain of people who already use the tool.

Step by step

  1. Pick a platform with lots of users and a marketplace where your specific pain still has little supply.
  2. Build a small app that solves only that pain, with a native look and feel for the platform.
  3. Submit it for review, prepare the listing page, and put the words people search for in the title and description.
  4. Make the first experience deliver value fast (guided onboarding, first result within 60 seconds).
  5. Work on reviews and rating until you reach first place in the category.
  6. Track the signups that come through the store and follow up with the people who sign up (demo, trial). Grow beyond the platform before the store gets crowded.

Best example

Privy published the free version on the marketplaces of the online store platforms and became the number 1 marketing app on Shopify, with about 15 thousand reviews. ARR went from about $1.3 million to $3.1 million in one year, with most of the growth coming from the marketplace (https://youtu.be/5twP3_CuBrQ).

Numbers that came up

  • Abi Noda, Pull Reminders: joined the GitHub Marketplace on April 1, reached about 350 paying accounts, and momentum took off after that (https://youtu.be/t_uSYaInXxk).
  • Supermetrics: 757 thousand downloads through the Google Workspace marketplace and hundreds of signups per day coming from it (https://youtu.be/jCPwAd_g2W4).
  • Rob Walling's interviewee: launched 4 apps on the monday.com marketplace and reached $30K in MRR in about a year (https://youtu.be/6WXHWKDEbIs).
  • Reamaze: 20 to 30 trials per day from the Shopify app store (https://youtu.be/vSdm1YIjeVs).

When it does NOT apply

  • The platform takes a cut of each sale: 25% on GitHub and Shopify (BillBee), 30% revenue share on Shopify (Reamaze), about 15% on AppExchange licenses (Blackthorne).
  • The platform can copy you or launch the feature natively and make your app redundant (Data Fetcher, Blackthorne).
  • The store can get crowded and you have to leave it to keep growing (Karmabot). It also takes time: the competitor mentioned in the Shed Wool video took about a decade to reach first place on the Intuit App Exchange.

"Powered by" branding on widgets, badges, or content your customer publishes

28 founders used itin 6 of 7 channels14 with revenue cited

What it is

Everything your customer publishes (form, widget, newsletter, video, page) carries a link or badge for your product. People who see the content click and land on you. The branding usually stays on the free plan and disappears if the person pays for a higher plan.

Why it works

The embedded element is shown to every visitor of the customer's site. Each use of the product becomes a free ad. The result is referral traffic, searches for your brand name, and backlinks.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

The product has a visible part on someone else's site or material (form, widget, public page). It also requires a free or cheap entry plan, where the branding stays.

Step by step

  1. Put a visible "powered by" link in the footer of the form, widget, or content the customer publishes.
  2. Keep the branding on the free or entry plan and let the customer remove it by paying for a higher plan.
  3. Add a UTM to every link so you know which site each signup came from.
  4. Compare signups, conversion, and revenue by site and double down on the sites that convert best.
  5. If you want the badge to stay installed, give a reason: an affiliate commission for whoever refers you.
  6. Make the free plan genuinely useful, so more people use it and carry the link along.

Best example

ConvertKit put "powered by" on every page and form and later improved the free plan. Visits from those links went from 2 to 3 thousand per week to 14 to 15 thousand per week (https://youtu.be/jersh1-1jxw).

Numbers that came up

  • Shorthand: 30% to 40% of readers arrive through the footer logo, with zero marketing spend (https://youtu.be/Nmh8mIVf3-o).
  • YouCanBookMe: half or more of new accounts come from the "powered by" button on the free plan (https://youtu.be/zK_YWQPg37I).
  • Arium: people who click the "powered by" link convert at close to 7 to 8%, according to the founder (https://youtu.be/YiNfXj0w3Ps).
  • JivoChat: about 230 thousand sites with the widget and 31 thousand paying customers, with branding that can't be removed (https://youtu.be/4dUBeIG-JGI).

When it does NOT apply

  • It only works if the product shows up inside other people's sites or content. It doesn't work as a generic channel (Hiten Shah, KISSinsights).
  • The traffic may convert poorly: 1 to 2% in ConvertKit's case, against 7 to 8% from other channels.
  • A big free plan attracts abuse (phishing and spam on Tally) and, in markets where free is strong (India and Indonesia for JivoChat), almost nobody pays.

Free tool as bait that asks for an email and leads to the paid step

25 founders used itin 6 of 7 channels10 with revenue cited

What it is

You build a small tool that answers a question your audience already has: a site diagnostic, a calculator, a report, a quiz. The person gets the result right away and leaves their email. Then you lead them to the paid product that solves what the tool revealed.

Why it works

The instant result is worth an email. People who use it already arrive with the problem your product solves, so the lead comes qualified. Since the tool is simple, you can generate lots of leads at scale.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

There is a question your audience wants answered and you can deliver the answer in seconds. It helps if you have data or analysis that nobody else shows.

Step by step

  1. Find the question in your niche (what's my site's score, how much is my asset worth, what's in my water).
  2. Build the simplest tool possible, one you can make in a day or two, even without code (n8n and Lovable came up as examples).
  3. Ask for the email before delivering the full result. Two of the cases cite not capturing email as a serious mistake.
  4. Show the result in an easy to share way (percentile instead of an absolute score, as in Website Grader).
  5. On the next screen, offer the paid product that solves the problem that showed up.
  6. Do the math: the cost of building the tool against the value of each email that becomes a customer.

Best example

SecurityScorecard built a widget where you type a company's URL and get a security score. The report was downloaded by more than 880 thousand companies, brought in 85 thousand leads, and they became customers with an average ticket of $23K to $25K per year (https://youtu.be/CynGGu3WaBA).

Numbers that came up

  • HubSpot, Website Grader: tens of thousands of leads at the start, grading about 800 thousand sites (https://youtu.be/5WOzdCZRDnU).
  • Flippa: free valuation tool with more than 10 million searches per month (https://youtu.be/rwz2i_7SObQ).
  • UX Exec: free presentation with more than 200 thousand downloads and a paid plan of about $4 per month (https://youtu.be/mwbIDovpOs4).
  • Claap: free tool connected to Google Calendar helped the waitlist reach about 3 thousand people (https://youtu.be/VBQcR8c9hNY).

When it does NOT apply

  • It depends on having useful data or analysis in your field, and on the team working the lead afterward (SecurityScorecard, Jolie).
  • ProfitWell: about 30 competitors with Stripe integration showed up, and the founder says he should have raised money earlier.
  • Some cases have no conversion numbers. At Estalty and Website Grader, not capturing email at the start cost leads.

Extension or add-on inside the tool the user already uses every day

19 founders used itin 5 of 7 channels14 with revenue cited

What it is

Instead of building a new app, you build a Chrome extension, a Google Sheets add-on, an Outlook plugin, or an app inside Slack. The product lives where the person already spends the day. They use it without having to remember to open something else.

Why it works

The action already happens every day (buying, applying, opening a new tab, writing an email). If the product automates that action, usage comes on its own. The extension store also brings in users without a partnerships team, at low cost.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

There is a frequent action on the web or in a tool that can be automated, preferably without requiring an account up front. It also works for a solo dev testing a cheap idea.

Step by step

  1. List what people already do every day in the browser or the tool (finding coupons, fixing text, filling out job applications, importing data).
  2. Build a small extension that solves one pain, working right at install, with no signup or credit card.
  3. Create the developer account on the Chrome Web Store (one-time fee of $5), prepare 16, 48, and 128 pixel icons and the description, and submit.
  4. Ask about 5 friends to leave a review on day one, to generate the first visits.
  5. Keep the free version and add a discreet button for the paid plan, which removes the branding and the limits ($39 or $79 per year in Digital Inspiration's case).
  6. Ask for the account only later, when you have a bigger benefit to deliver.

Best example

Simplify, a Chrome extension that fills out job application forms on any job site, passed 5 million applications sent in about a year and a half, with more than 150 thousand users and zero dollars in paid marketing, according to the team (https://youtu.be/GcZtzqvL-S0).

Numbers that came up

  • Super Power ChatGPT: about 150 thousand weekly users and $20K to $30K in MRR, built by a solo dev (https://youtu.be/ty9ZcimL6VE).
  • SalesBolt: about 80% of customers came from the extension that runs on LinkedIn Recruiter, with 30 to 40 inbound leads per week (https://youtu.be/6vzdv4kVfZQ).
  • Nimble: free Outlook plugin converted 10% to trial, against 5% from the website (https://youtu.be/RQ28ne1EvSY).
  • Honey: about 17 million users when it was sold to PayPal for about $4 billion (https://youtu.be/OialI0ZNU3g).

When it does NOT apply

  • The platform can take the extension down or launch the feature natively (OpenAI in the Super Power ChatGPT case, Google in the Grammarly case).
  • The store has approval rules and sites with strong security, like YouTube, block injected scripts.
  • Lots of signups don't prove retention: Rocket Note had about 3 thousand signups and low activity. Depending on a single channel is also a risk (SalesBolt).
09

Partners and affiliates

## Partnerships and affiliates

Embed the product in the partner's software or workflow, with revenue share or white label

69 founders used itin 2 of 7 channels53 with revenue cited

What it is

A partner that already has the customers sells your product inside their software, under their brand or yours. You bill the partner and almost never serve the end customer. It works as a shortcut to lots of customers without building a sales team.

Why it works

The partner already has the trust and the customer base. They sell your product as part of what they already deliver, so you don't pay to win attention. Each partner you close brings dozens or thousands of customers at once.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

Your product is configurable and doesn't require heavy support. You have access to partners with their own base of small or mid-sized customers.

Step by step

  1. List who already serves your ideal customer: marketplaces, payment gateways, ERPs, store networks, publishers, broadcasters, apps with a large base.
  2. Build a version of the product the partner can sell without you: white label, API, or checkout integration.
  3. Close the revenue share (in the cases, between 30% and 50% to the partner) and decide who keeps support.
  4. If the partner has a sales team, train that team and let them run the webinars and onboarding.
  5. Forbid the partner from selling the product standalone below your direct price, only inside a bundle of theirs.
  6. Use the first wave of customers as an entry point to the full product.

Best example

Rupify (credit for small merchants in India) put credit into the checkout of large B2B marketplaces. It went from about US$ 5 thousand per month in June to about US$ 60 thousand in September, with about US$ 7.5 million lent in September alone (https://youtu.be/R3W87-2nxEQ).

Numbers that came up

  • GreenRope: about 1,500 of the 3 thousand customers came from 25 resellers (https://youtu.be/5d2Xdcz1iZQ).
  • Lucky Orange: about 40% of revenue comes from partnerships, and more than half of that doesn't depend on a single partner (https://youtu.be/LRlBx19Emok).
  • Seon: 4 to 5 payment gateways bring about 4 to 5 thousand indirect stores, on top of 200 direct customers (https://youtu.be/exuWBK5noVw).
  • WPBuffs: white label partners manage 60 to 70% of the company's roughly 1,000 sites (https://youtu.be/iFXtC4Ucl5g).

When it does NOT apply

  • A few large partners concentrate the risk: at Webboss, partnership revenue fell when the partner laid people off during covid (https://youtu.be/KBiWli0OzvI), and Vivid Learning saw renewal drop to about 85% (https://youtu.be/gZUDdo8be0w).
  • Average ticket drops with small customers: at Upsales it went from about US$ 15 thousand to US$ 6 to 7 thousand per year (https://youtu.be/LcWZrNpEZJE).
  • In models like Froda's, the partner requires a banking license and its own balance sheet, so it doesn't fit pure software (https://youtu.be/RYqbTW4ApMA).

Sell through agencies, accountants, consultancies, and integrators who bring the product to their own base

38 founders used itin 5 of 7 channels30 with revenue cited

What it is

You offer the product to professionals who already serve your audience, such as agencies, accountants, consultants, and integrators. They refer or resell to their client portfolio. One sale to a professional can put dozens of customers on the platform.

Why it works

The professional already has the client's trust, so the referral arrives ready-made. They also gain from the product: it saves them time, lets them deliver more, or lets them charge the end client. You trade direct sales for a few large sales and CAC drops.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

Your product solves something the professional uses with several clients, and that professional serves the same profile you want to reach. It works best when the incentive is simple (margin, discount, or time saved).

Step by step

  1. Find the type of professional that already serves your ideal customer (accountant, agency, MSP, notary, consultant, integrator).
  2. Choose the incentive: discount or margin on the subscription, recurring commission (20% to 30% in the cases), or just the time saved.
  3. Make a direct sale first, to have happy customers and proof. Only then open the channel.
  4. Train the partner once and give them a dashboard or a simple way to refer.
  5. Calculate the cost of bringing in each professional (demo, team, events) and ask for referrals from those already participating.
  6. Track concentration: see how many customers come from each partner so you don't depend on just one.

Best example

HubDoc sold to accountants and bookkeepers, who add all their clients to the tool. About 6,500 to 7,000 of the 7,500 small businesses came from more than a thousand paying accountants, with revenue of about US$ 132 thousand per month (https://youtu.be/k97GdASFPPQ).

Numbers that came up

  • Coro: the MSP channel went from about 70% of revenue in 2023 to about 85% projected in 2024, with about 300 partners (https://youtu.be/WzPc9fwqxjA).
  • Keybot: 16 agencies bring about 1,100 of the 1,500 customers, and each agency costs about US$ 100 to 120 to win (https://youtu.be/1guKbnflfC8).
  • WhatConverts: about 400 partner agencies serve more than 10 thousand sites (https://youtu.be/JrUyYOHwGAY).
  • HubSpot: after 4 to 5 years, selling through agencies became about 50% of revenue, according to Jason Cohen (https://youtu.be/8xLquwfx6p0).

When it does NOT apply

  • When a large agency leaves, revenue drops all at once: Keybot took about a year to recover (https://youtu.be/1guKbnflfC8).
  • The channel is slow: at Coro the partner generates revenue around the second month and only becomes profitable for the company in the fifth (https://youtu.be/WzPc9fwqxjA).
  • With a small end customer (US$ 30 per month), CAC gets close to US$ 300 and payback is slow (https://youtu.be/JrUyYOHwGAY).

Monetize your own audience as an affiliate for products you use

23 founders used itin 3 of 7 channels16 with revenue cited

What it is

You recommend third-party products to your audience, with an affiliate link, and earn a commission on each purchase or signup. It can be a newsletter, a video, a comparison site, or a list of tools. You can earn revenue without building your own product.

Why it works

The audience already trusts you and already asks what to use. The link sits right where you'd make the recommendation anyway, so no ads are needed. When the commission is recurring, each referral keeps paying for months.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

You already have a direct audience (newsletter, email, channel) or search traffic, and you use products that have affiliate programs. The cheaper the content is to produce, the better, because the return only comes with volume.

Step by step

  1. List the tools and products you actually use and see which ones have affiliate programs.
  2. Compare the payment type: per thousand visits (CPM), per lead (CPL), or per sale (CPA). Prefer recurring commission when the customer pays every month.
  3. Choose the format: a newsletter about one tool, a best-of list, a comparison page, or a short video.
  4. Put the link where the recommendation happens (video description, bio, email).
  5. Test the product before recommending it and show how you use it, without looking like an ad.
  6. Track the conversion of each page and cut what doesn't pay off.

Best example

CompareTech, a site that compares VPNs, created in 2015, gets about 1 million visits per month and, according to public UK filings, had US$ 15 million in revenue and US$ 13.2 million in profit in 2021 (https://youtu.be/hbUZcRWaBHA). The host did not confirm the commission per customer.

Numbers that came up

  • An acquaintance of Greg Isenberg with a large newsletter earns US$ 40 to 60 thousand per month in commission from a supplement brand (https://youtu.be/pyYk5WPcPm4).
  • Cody earns 40% of the lifetime revenue of each US$ 99 per month subscription to a community platform (https://youtu.be/uc0y94vNqlo).
  • The Penny Hoarder earned US$ 2,000 per referred rideshare driver and reached about US$ 70 million in revenue for the year, with no investors (https://youtu.be/F2tljx5A98Q).
  • At Divi, affiliates accounted for about 50% of monetization, and the rest came from brand deals (https://youtu.be/Q1_m8ScJgYM).

When it does NOT apply

  • Without a direct audience, the channel loses strength: those who depend on SEO or the algorithm suffer more (https://youtu.be/pyYk5WPcPm4).
  • The window can close fast: on TikTok Shop, competition arrives and conversion drops (https://youtu.be/zpObHDZq5uE).
  • Commission changes depending on the program and isn't guaranteed, and several numbers come from secondhand accounts without confirmation (https://youtu.be/uc0y94vNqlo, https://youtu.be/2dOCPr355TQ).

Pay affiliates recurring commission for as long as the customer pays

29 founders used itin 5 of 7 channels23 with revenue cited

What it is

You pay the referrer a share of each monthly payment, for as long as the referred customer keeps paying. The cases use 20% to 50%, and the most common is 30%. The affiliate earns more the longer the customer stays.

Why it works

The commission only goes out when revenue comes in, so acquisition cost is variable. The affiliate also has a reason to keep promoting and to help the customer stay. Over time, the affiliate portfolio becomes a sales channel that doesn't ask for a fixed salary.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

Your product is subscription-based and has the margin to pay ongoing commission. It works best after direct sales have already brought in happy customers and there is an audience reachable through people who have one.

Step by step

  1. Go with direct sales first, to have happy customers and proof. Affiliates speed things up but don't replace that stage.
  2. Set the recurring commission (start at 20% to 30%) and use an affiliate tool to track and pay.
  3. Calculate the maximum CAC from the payback you want (a customer paying 100 per month and a 3-month payback allows spending up to 300).
  4. Recruit people who already have an audience of your profile: bloggers, review creators, consultants, accountants.
  5. Give affiliates a dashboard with signup, payment, and cancellation of each referred customer, so they can help with retention.
  6. Review the commission over time (reduce it for new affiliates or swap it for a one-time amount) and always pay on time.

Best example

Surfer SEO pays 30% recurring to creators who do reviews. There were about 3,000 affiliates, managed by one person part-time, with about US$ 3.5 million in partnership revenue in the last year, more than 30% of the total (https://youtu.be/gmYaKC0u3aQ).

Numbers that came up

  • SEMrush: in 2018 it passed US$ 60 million in revenue, with about 40% coming from affiliates (https://youtu.be/gi39IvjfGRI).
  • system.io: 70% of sales come from affiliates, with more than US$ 100 thousand per month paid in commission (https://youtu.be/DjZQA7sCfI8).
  • ConvertKit: about 3 thousand paid affiliates in 2020 and about US$ 250 thousand per month in commissions (https://youtu.be/jersh1-1jxw).
  • SmartSuite: 50% of first-year revenue to affiliates brought about 325 affiliates in six weeks and US$ 85 to 90 thousand in MRR (https://youtu.be/mmgXFbcH6N4).

When it does NOT apply

  • Margin gets tight: with 40% commission the maximum margin is around 60% (https://youtu.be/gi39IvjfGRI), and with 20% it is below 80% (https://youtu.be/1N0wQ1j3T1g).
  • Results are concentrated: at ConvertKit about 100 affiliates make US$ 10 thousand per month or more (https://youtu.be/jersh1-1jxw).
  • Building the network takes time or contacts: Convertri says its network came from years in the industry and isn't easy to replicate (https://youtu.be/YtYDYd1O1bM), and system.io took years to grow (https://youtu.be/DjZQA7sCfI8).
10

Launch

Build a waitlist or audience before launching

23 founders used itin 6 of 7 channels14 with revenue cited

What it is

Before the product is ready, you collect emails from people who are interested. You can do it with a landing page, a simple form, or posts in your audience's communities. On launch day, you notify that list.

Why it works

People who already signed up raised their hand and are waiting for the announcement. That turns the launch into a sale to people who already want it, and the result stops depending on luck on Product Hunt or Reddit. The first sales show up on day one.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

You are still building, or you just validated the idea, and you can reach your audience before the product exists. It works best if you have some audience or access to niche communities.

Step by step

  1. Put up a landing page (or a simple form, like Google Forms or Tally) with the product's promise and an email field.
  2. Send people there through several channels: your content, niche communities (Facebook, Reddit, Discord), podcasts, conferences and, if you have budget left, paid ads.
  3. Offer a real reason to sign up: a discount in the first hours, a free first chapter or beta access.
  4. Talk to the people who signed up and ask what they use today and what they want to switch from. Use the answers to adjust the product.
  5. Launch to the list first, with a launch price or a short pre-sale.
  6. Plan what comes after launch day, so traffic doesn't drop off a cliff.

Best example

Drip gathered about 3,400 emails before launching: about 500 from its own audience, 600 from its network, 500 from press and lists, 900 from ads and 900 from the "powered by" link (MicroConf 2014, Rob Walling). Those 3,400 made launch day very strong (Validate Your SaaS Idea FAST).

Numbers that came up

  • Amy Social: Facebook group with over 1,000 people, a 72-hour pre-sale at half price, about 250 bought (about 25%) and revenue was already about 20k per month (video).
  • Tenon.io: about 1,000 people on the list before the beta, and the first paying customers came in within an hour of the announcement at the CSUN conference, in 2015 (video).
  • RightMessage: about 2k emails at launch, and 100% of the first sales came from the list (video).
  • 67 Water: 25k signups on the waitlist and 300 thousand dollars in revenue in the first 5 weeks (video).

When it does NOT apply

  • A list loses value every week. Steph Smith had a list of 1 million people for an app and, at launch, only 10k to 20k tried it (video). Test the product with 100 to 1,000 customers before growing the list.
  • A signup is not a customer yet: at Flight, billing hadn't even started at the time of recording (video). At Drip, the emails from its own audience converted poorly, because many just wanted to follow the process.
  • Fake limited spots: Beehive used them and the founder himself warns that the audience notices later (video). Waiting 18 months like SparkToro requires cash to live on during that period (video).

Launch to the list and audience that already trusts you

17 founders used itin 6 of 7 channels10 with revenue cited

What it is

You take the new product first to a base that already exists: newsletter, blog, channel, community, customers from another business. These people already know you and don't need ads to be convinced.

Why it works

The trust is already built, so the acquisition cost is close to zero. The base buys or tries it right on day one, and that generates the first numbers and testimonials for the rest of the growth.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

You already have your own blog, list, channel or community, or customers from a previous business. The more engaged the base, the better.

Step by step

  1. List the base you already have: emails from another project, newsletter, followers, course or consulting customers.
  2. Turn something this base already buys, or a system you use yourself, into a product.
  3. Ask the base what it wants in the product and tell the story behind it.
  4. Send the launch by email to the whole list on the same day and reinforce it with posts and lives.
  5. If you are going to post on Product Hunt, do it afterward, with testimonials already collected and without pointing all the links there at once.
  6. Prepare servers, support and cash for the first-day peak.

Best example

Thomas Frank made about 90 thousand dollars in one launch of Ultimate Brain to a list of 3,200 people, from a template he already used for himself (video).

Numbers that came up

  • system.io: a list of 25k contacts from courses sold since 2010 brought in about 400 first customers and close to 10 thousand euros per month right at the start (video).
  • Edgar (Laura Roeder): 100k in MRR within a few months after launch, with years of training and consulting customers as the base (video).
  • First Round Review: about 30k emails sent and 8k signups in the first week (video).
  • Bolt (StackBlitz): launched with a single post on X, about 60k in ARR on the first day and 20 million in ARR in two months (video).

When it does NOT apply

  • It only works with an engaged list. The system.io list came from years of selling courses (video), and the trust at Handshake took a decade to build (video).
  • Selling software straight to an audience that has never bought anything is hard. Selling a book or course first helps (video).
  • A course gives a big spike and then drops, and you have to keep creating another one. It doesn't generate recurring revenue (video).

Launch on Product Hunt to get a first spike in signups

23 founders used itin 5 of 7 channels15 with revenue cited

What it is

You publish the product on Product Hunt and take advantage of the day when curious visitors from the community arrive together. The goal is to get signups, trials and first customers without paid ads.

Why it works

Product Hunt concentrates people curious about new things into a single day. If signup is simple and the product is usable on first access, part of that traffic turns into trials and some become paying customers.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

The product is ready to use on first access, with simple signup, and you don't have another source of customers yet.

Step by step

  1. Test the platform beforehand by launching smaller products, to understand how it works.
  2. Keep signup simple and the product usable right on first access, ready for the spike.
  3. Write the message with a clear angle, even a provocative line about the buyer's pain.
  4. Ask your contacts for votes and avoid buying votes. Prepare the 24 hours of promotion.
  5. Keep promoting on social media all week, because part of the real traffic comes after the day.
  6. After the spike, take the same message to the communities where buyers hang out (Reddit, Hacker News, LinkedIn) and filter out low-quality signups.

Best example

Softr launched the simple MVP and got about 12k visits, hundreds of signups and about 10 paying customers on the day, with the product ranked first (video).

Numbers that came up

  • Salesflare: about 200 new trials and 4k unique visitors in two days, and traffic stayed above the previous level (video).
  • Vidboard: 426 votes and about 1,500 signups in launch week (video).
  • TrueFan (Surf): about 600 signups from Product Hunt on the first day (video).
  • Shipper: Product Hunt brought the first 50 dollars of MRR, and Reddit took it from 50 to a thousand (video).

When it does NOT apply

  • Few people pay. Tranul had 448 votes and few paying customers, but gained affiliates and influencers (video). At Leon, the result came more from communities than from Product Hunt (video).
  • Low quality: at GrowSumo, most signups were hackers and seed-stage startups (video). SingleLink wasn't prepared for the volume and lost people (video).
  • There is a lot of luck involved, and today it is harder to stay on top. SubMagic's founder doesn't recommend repeating it (video).

Run a beta with acquaintances, fans or your personal network before opening to the public

13 founders used itin 6 of 7 channels6 with revenue cited

What it is

Before the open launch, you put the product in the hands of a close group: friends, followers, your own team, a waitlist. They test it, point out what gets stuck and later speak well of the product.

Why it works

Close people give honest feedback and accept a product that is still broken. You fix the errors before the big day, and the people who took part in building it defend the product and invite others.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

You have access to acquaintances, a community, a team or followers willing to test, and the product hasn't been opened to the public yet.

Step by step

  1. Gather the group: 10 to 150 people from your network, followers on X or the waitlist.
  2. Invite them by post ("does anyone have this challenge?") or DM, or send a short personal email, one by one.
  3. Show them in person how to use it and watch where they click and get stuck.
  4. Reply to the group's emails and let them know when a bug is fixed.
  5. Ask them to share it with people they know.
  6. Keep improving until the day before and only then open to the public, or start charging.

Best example

Creator Buddy gave the product to 100 to 150 followers on X, followed up with each one and reached 100k in ARR within 15 minutes and 300k in ARR within 2 weeks after launch (video).

Numbers that came up

  • Fastlane: beta sent to about 2,000 people from the waitlist, with the public launch on March 23 (video).
  • X.ai: 50 friend beta testers brought about 500 users through referrals (video).
  • Pill Up: from 10 people to 250, through word of mouth alone, before the full launch (video).
  • Visably: about 1,000 free signups in about three weeks, coming from the founder's networks and B2B outlets (video).

When it does NOT apply

  • If the product breaks on the day, the beta doesn't prevent everything: the Creator Buddy app broke and was fixed on the spot (video).
  • A long free beta needs a reason. X.ai stayed in it for three years, and it only worked because the problem was hard to solve and hard to explain (video).
  • It works when a loyal base already exists to form the group, as Polyvore shows (video). A beta signup still doesn't show how many will pay (video).
11

Outbound

Ask for space in third-party media (podcast, press, blog) with a specific request

16 founders used itin 6 of 7 channels4 with revenue cited

What it is

You find podcasts, journalists and blogs that already talk to your audience and send a short, personalized pitch to appear there. It can be an interview, an article or a guest post. You pay with research and email work, not with money.

Why it works

Podcasts and publications need guests and story ideas all the time. When the email shows you know the content and already hands over a ready-made topic, the host only has to say yes. The audience they built starts seeing you, and each appearance opens the door to the next one.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

Early stage, no audience and no budget, when you want reach and contacts. It works best if you have something truly useful to teach the publication's audience.

Step by step

  1. List 5 to 20 peers or competitors who reach the same audience. Google each name and note the podcasts where they have already been guests.
  2. For press, check the publication's site to see who writes about your topic (the author of your competitor's articles). Confirm whether they are a freelancer and find their portfolio on Google.
  3. Engage with the person for a few weeks first: reply to stories, comment, send short messages.
  4. Write a short email: mention a specific episode or article, propose 3 topics that teach something new and ask them to reply with just the number of the topic they pick.
  5. In a press pitch, tell the problem you faced and how you solved it, without bringing up the product right away.
  6. If they don't reply, send a short follow-up ("did you get my last email?") and keep up the volume: 10 emails a day, for months.

Best example

Courtland Allen, of Indie Hackers, emailed 200 people asking for an interview and almost all of them said no at first. The first 10 published interviews made many more people say yes afterward (https://youtu.be/MHZU4fC7LnA). Another strong case is SubmitHub, which hand-wrote more than 1,000 emails to blogs, every day for 5 months, and reached US$ 55 mil per month (https://youtu.be/MHZU4fC7LnA).

Numbers that came up

  • Erik Davis pitched 7 sites, got a 100% response rate and 57% publication rate, and is now paid to write for 4 big sites (https://youtu.be/oXACKWi6QYU).
  • Buffer: about 10 emails a day for 3 to 5 months, with a 30% to 40% response rate; 3 to 4 out of every 10 emails turned into an installed button or a guest post (https://youtu.be/p5Yoyh90nTM).
  • Rooted got into Cosmopolitan, Women's Health and Time with cold email alone (https://youtu.be/VfNRd5Rk0cM).
  • Danny Miranda started in September 2020 and reached 30 to 50 thousand downloads per month by asking for interviews by message; he says 100 messages is the minimum (https://youtu.be/Fk2J2U7669o).

When it does NOT apply

  • A generic request ("can I be a guest?") doesn't work, and neither does pretending to be a fan: the journalist notices right away (https://youtu.be/LZLdoAq3R7Y).
  • It is almost full-time work and doesn't scale, according to Buffer's founder (https://youtu.be/p5Yoyh90nTM).
  • Big-name guests almost always say no; the gain comes from message volume (https://youtu.be/Fk2J2U7669o).

Outbound as the channel for your first customers before inbound

23 founders used itin 5 of 7 channels16 with revenue cited

What it is

Before you have a brand, content or traction, you build a list of your ideal customer and send cold emails, messages and calls to that list. It is the grunt work that brings in the first customers while inbound doesn't exist yet.

Why it works

Without a brand, nobody shows up on their own. Talking directly to the right audience takes the product to the people who have the problem and brings back fast feedback, which tunes product and price. Content takes months to pay off, while outbound answers within the week.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

At the start, with no brand or traction, usually before the first 10 to 100 customers. Once the product catches on, it is worth testing inbound and referrals to bring costs down.

Step by step

  1. Define a small, specific audience (for example, founders, CMOs, real estate photographers or bank decision-makers).
  2. Build the contact list, with a dedicated tool, LinkedIn or by hand. Lists of 100 to 1,000 people are enough to start.
  3. Send an email, call or DM to each one. Do it yourself at first, because the conversation teaches you the product.
  4. Use the feedback from the replies to tune product and price.
  5. With your first customers in hand, ask for referrals to sell to the next ones and go to industry conferences.
  6. Once outbound is running, test inbound and double down on what works.

Best example

Rippling built a list of founders, CEOs and CTOs at growing companies and sent cold email; the first 100 customers came mostly from that, and the company reached about 2,000 customers (https://youtu.be/DEaTgKg8WA4).

Numbers that came up

  • ServiceForm: about 95% of revenue comes from outbound, with MRR going from 65 mil to 150 mil in 12 months (https://youtu.be/KlLZtICs3lI).
  • Synthesis: 60% of customers come from outbound and 30 to 40% from inbound; CAC of about US$ 5 mil for customers paying US$ 2 to 3 mil per month, with a 2 to 3 month payback (https://youtu.be/jOMySrH2H3Q).
  • BuzzBuilder: CAC under US$ 200 per customer and payback under two months, with an average ticket of US$ 4.900 per year (https://youtu.be/6ftSe2FZImQ).
  • Gong went from 12 customers to about 200, and 20 to 30% of the business comes from referrals by existing users (https://youtu.be/vdHXt_xa4Gw).

When it does NOT apply

  • It is manual work, and the founder himself made the calls; to scale he needed more people (https://youtu.be/QK3LSsW9GVY).
  • In most cases no conversion rate is cited, so the return is uncertain (https://youtu.be/Ki3qBvIRo6A).
  • At a large company the sales cycle can take 6 to 10 months (https://youtu.be/voWKlhA7jV4).

Deliver value or ready-made proof before the ask

18 founders used itin 6 of 7 channels10 with revenue cited

What it is

Instead of asking for a call, you send a ready-made demo for free: a report, prototype, analysis, clip or lead list made for that specific customer. The target sees the result before spending any time on you.

Why it works

The customer sees the value without having to do any work or trust a promise. Someone who receives something made for their own company tends to reply and book the meeting on their own. Most competitors send only text, so whoever sends the actual deliverable stands out.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

High ticket and the ability to generate something quickly, or cheaply, for each target. It helps when you have your own asset (a database, a scraping tool, an AI that generates the demo).

Step by step

  1. Pick the target and find a concrete problem of theirs (a review report, a marketing gap, an ad not yet measured, outdated posts).
  2. Build the proof fast: a dashboard with 10 insights, a short report with screenshots, a video made for the company or a sample generated tour.
  3. Send it by cold email to the decision-maker (find the right person inside the company), saying "I made this for you".
  4. Offer to discuss the material, or deliver just a sample (for example, 50 free leads) and let the person reply.
  5. If there is interest, propose the paid pilot or the contract.
  6. Send in small batches (for example, 500 a day) so you don't land in spam.

Best example

A 9-page report with employee review data, sent on a Sunday to a Facebook executive, got a reply within hours and a meeting the next day; Facebook became a big customer (https://youtu.be/pYZ0S7a72po).

Numbers that came up

  • Massive: first big contract with Procter & Gamble, about US$ 150 mil per year, after delivering an analysis the buyer asked for; revenue went from US$ 40 mil to US$ 83 mil per month (https://youtu.be/CbchIc1EUEQ).
  • Niche.com: monthly report per school, sent in batches of 500 a day; paying customers went from 78 to 336 in one year (https://youtu.be/z1v9en87nP4).
  • Munch: a base of 800 mil creators and email with 4 to 5% conversion; reached 2 mil users in under two months (https://youtu.be/bu3v7lE9gYI).
  • Hyper: CAC near US$ 20 mil, payback of about 8 months and annual contract above US$ 30 mil (https://youtu.be/k2NiaR8acqM).

When it does NOT apply

  • The demo has to be made for that customer; a generic resource they can find on Google or in an LLM doesn't work (https://youtu.be/OWvDbX5Fmps).
  • With a low ticket, the cost of producing each proof may not pay off; the method of Hyper and Ace Metrix depends on having your own database (https://youtu.be/04p15wnARwc).
  • Watch the ethics: the Uber case involved exploiting the referral program, and the video itself recommends copying only the approach of showing the flaw and offering the fix (https://youtu.be/rG4zxf0CAv0).

Ask for feedback or a short chat instead of selling

13 founders used itin 5 of 7 channels5 with revenue cited

What it is

You send a short, personal message asking for the person's opinion on a problem or on something you are building, with no sales pitch. The ask is a 10 to 15 minute conversation.

Why it works

When you ask a question, the person doesn't feel they are being sold to and lowers their guard. The conversation reveals the real problem in the market, and whoever replied often becomes a relationship, then a tester, then a customer. Also, asking for one small thing makes it easier to say yes.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

Discovery phase and first customers, with a simple or incomplete product, to reach decision-makers (CMOs, founders, product leaders).

Step by step

  1. Find the right audience on LinkedIn, Product Hunt or through Sales Navigator (for example, cloud professionals, CMOs or SaaS founders).
  2. Send a short message naming the problem you solve and asking for feedback. Without the problem in the message, nobody replies.
  3. Ask for a single commitment, with a clear size and format, like "15 minutes by phone".
  4. On the call, spend 60 to 70% of the time on the person's problem and, at the end, ask if they want to see what you are building.
  5. Send the signup link for the trial and, after they use it, book the demo.
  6. Ask for the sale explicitly afterward, only to those who fit the product.

Best example

Brainboard messaged cloud professionals on LinkedIn asking for feedback on a tool that solved a specific problem. That produced more than 100 free POCs in 6 months and 10 enterprise customers converted, about 10% (https://youtu.be/QDxZL6Ysn8s).

Numbers that came up

  • NeoReach: about 200 to 300 messages to CMOs asking for feedback; the first big customer was Zappos, with a US$ 5 mil per year license, referred by the CMO himself (https://youtu.be/M14CDWgRV7E).
  • Tally: 15% to 20% response on DMs asking for feedback from people who upvoted similar products on Product Hunt (https://youtu.be/Dt6HAEuwEug).
  • Tutor: response rate of about 30% and open rate above 90% on LinkedIn and email sequences (https://youtu.be/Rdz8GVnZsBc).
  • QuanticMind: the first enterprise sale, with HomeAdvisor's CMO, came in 2014 after about six months of conversation that began in 2013 (https://youtu.be/URaP3FNgPM0).

When it does NOT apply

  • It is laborious and doesn't scale; it worked early on, while the MVP was still being built (https://youtu.be/Dt6HAEuwEug).
  • The cycle can be long: from 2014 to 2018 QuanticMind's prospecting was almost entirely outbound (https://youtu.be/URaP3FNgPM0).
  • Conversation alone doesn't close: afterward you need to ask for the sale, with a special offer, and only to those who fit (https://youtu.be/OWvDbX5Fmps).
13

Sales

Land and expand: start small in a big account and grow from the inside

56 founders used itin 3 of 7 channels45 with revenue cited

What it is

You close a small contract with one team, one division or one brand of a large customer. Once the results show up, the product spreads to the other areas of the same company, with more seats, more use cases and more budget.

Why it works

The first working case becomes internal proof. The other areas already know the product and have seen the results, so you don't sell from scratch to every division. The account grows on its own, and net revenue retention (NRR) rises with it.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

Your customer is a large company with several divisions, brands, countries or departments. The product needs to deliver visible results with a small team before any expansion.

Step by step

  1. Pick one team, division or brand as your entry point, with a small contract or pilot (10 users, 10 to 20 units, or one use case).
  2. Install fast and teach the team to use it well. Radius got the platform live in about 18 days.
  3. Measure the results before and after (Rising Team measures engagement before and after three sessions) and keep the numbers.
  4. Use that data to propose the expansion: the whole division, then the other regions, brands or use cases.
  5. Ask customer success to look for expansion opportunities inside the account, instead of waiting for the customer to ask.
  6. Protect the internal champion, the person who backed the product, because at Hyper the customers who leave tend to be the ones who lost that champion.

Best example

Calibra closed an initial contract of about US$ 150 thousand and expanded to other business units and use cases. Customers went from US$ 150 thousand to US$ 1 million in ARR, with net retention of about 130% and churn of about 4% a year (vídeo).

Numbers that came up

  • Hyper: customers who started on US$ 20 thousand packages now pay close to US$ 200 thousand when the whole group comes in (vídeo).
  • Badger Maps: one customer started with 10 seats, went to 200 in one division and reached 1,000 seats (vídeo).
  • Rising Team: dollar net retention of 145% and about 60% of demos turn into a contract or pilot, according to the founder (vídeo).
  • Stella Connect: NRR of about 130%, with gross churn near 13% and expansion of about 43% (vídeo).

When it does NOT apply

  • When the challenge is turning the pilot into a large rollout: the Rising Team founder said he is still working on that.
  • In accounts with one or two seats (SMB): at One Mob, churn in that group is high.
  • With few customers, losing one already hits revenue hard (Door Boost).

CAC ceiling: a spending limit per customer tied to what they pay

54 founders used itin 2 of 7 channels41 with revenue cited

What it is

You decide up front how much you're willing to spend to bring in a customer, as a fraction of the first-year contract, of LTV or of a few months of subscription. Each channel or campaign only gets in if it fits under that ceiling.

Why it works

The ceiling turns "how much to invest in sales" into a simple number to check. If the customer pays the spend back in 12 months or less, every dollar invested comes back fast and cash flow can sustain the growth.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

You know your annual contract or monthly subscription value and the cost of sales weighs on cash flow. Works best with a ticket high enough to pay for the acquisition spend.

Step by step

  1. Calculate the customer's ticket: ACV (annual contract value) or monthly subscription.
  2. Choose the ceiling rule: up to 50% of first-year ACV (CloudHealth), equal to the first year (Vena), 80% of the first invoice (Cookie Information) or one third of LTV (noCRM.io).
  3. Add up everything you spend to win the customer (commission, travel, conferences, ads) and divide by the new customers.
  4. Calculate payback in months and track whether it falls over time.
  5. For LTV, use a conservative horizon (3 years at Help Scout, 5 years at Social Tables) instead of 1 divided by churn.
  6. Only approve the channel or campaign that fits under the ceiling.

Best example

Vena Solutions has customers worth about US$ 50 thousand a year and accepts spending US$ 45 to 50 thousand per customer, which gives a 12-month payback, with about 50 new customers per quarter (vídeo).

Numbers that came up

  • Active Demand: ticket of about US$ 400 a month and accepted spend of up to US$ 800 per customer, payback of about 2 months (vídeo).
  • noCRM.io: maximum CAC of 500 euros for a customer paying US$ 75 a month, payback of 7 to 8 months, 2,500 customers (vídeo).
  • Consensus: CAC of about US$ 18,500 for a US$ 25 thousand a year contract, with payback under a year (vídeo).
  • ECRS: cost of sale of about US$ 1,700 per customer who pays US$ 500 to 600 a month, payback of about 3 months (vídeo).

When it does NOT apply

  • With small customers, the math doesn't close (Consensus), and with very fast growth, payback can turn into a cash hole (ECRS).
  • The "1 divided by churn" LTV is misleading when churn is high, according to the Help Scout founder; Pure Chat calls its LTV a rough estimate.
  • High-CAC ideas take longer to pay off, so the ceiling requires patience with tests (Cookie Information).

Discover the problem before the demo and do a short demo only on the pain

23 founders used itin 6 of 7 channels8 with revenue cited

What it is

Before showing the product, you talk to understand the buyer's problem, in their own words, plus the timeline and budget. Only then do you run a demo of a few minutes showing just the screen that solves that pain.

Why it works

Qualifying beforehand filters out people who won't buy and gives you the script for the demo. Without the full feature tour, the buyer can't pick apart the story and the sale moves faster. Whoever sells features ends up in price negotiation (Churnly).

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

B2B with internal buy-in and a longer cycle, or SaaS with a ticket high enough to justify a call before purchase. Qualification has to come before the demo.

Step by step

  1. Call every trial signup within minutes (Rubberstamp calls within 5) and ask what problem the person wants to solve, the timeline and the budget.
  2. In discovery, ask what they do today, what works, what doesn't and what the consequence of not changing is (Moraware). Then stay quiet and listen.
  3. Only take to the demo those who fit, and use the call to disqualify early.
  4. Keep the demo short: 7 to 9 minutes on the screen that solves the pain, in the customer's language (Steli Efti, Rubberstamp).
  5. End the demo with a next-step question and wait through the silence.

Best example

Rubberstamp started calling every trial, qualifying and only then running the demo, which dropped from 40 to 50 minutes to about 8. Trial conversion went from 3% to about 7% in three months, with half the demos and more closed deals (vídeo).

Numbers that came up

  • RapidFunnel: 20 quick, qualified demos in two and a half weeks, with about 70% closing (founder's estimate) (vídeo).
  • Client with a script lacking gap questions: after adding questions about where the customer wants to get to, where they are and what's in the way, it generated US$ 1.5 million in pipeline in 3 months (vídeo).
  • Moraware: a sale that took 30 person-hours over months got shorter, and the founder has already talked to about 1,957 countertop shops (vídeo).
  • Jen Abel: in about 1 out of 4 calls there is no maturity to buy (vídeo).

When it does NOT apply

  • Without qualification beforehand, the short demo doesn't solve it (Rubberstamp).
  • On an expensive product with too small a ticket, calling doesn't pay off; ScrapingBee and Moraware say it works with urgent pain or a ticket that justifies the call.
  • The 7 to 9 minute rule applies to software for small teams; for other products the demo can run longer (Steli Efti).

Focus on big customers and let go of the small ones

46 founders used itin 2 of 7 channels36 with revenue cited

What it is

You stop serving, or stop pursuing, low-ticket customers and start selling to large companies that pay much more for the same product. Some founders refuse contracts below a minimum value.

Why it works

A small customer generates almost the same support volume as a large one but pays a fraction, and tends to have high churn. At HappyCo, one customer with 110 thousand units was equivalent to more than a thousand small ones, with about 1 support ticket a month in both cases. With large accounts the ticket goes up and churn goes down.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

The small ticket doesn't grow and the product truly solves the problem for a large company. You need to withstand longer sales cycles and build a sales team.

Step by step

  1. Compare revenue, churn and support by customer tier and see who pays little and costs the same.
  2. Write a minimum-size rule: employees (Rallyware, more than 10 thousand), open job postings (ConveyIQ, 100 or more) or minimum ticket (US$ 50 thousand a year).
  3. Close a few large contracts first and use each one as proof to raise the price of the next ones (TrueFan).
  4. Hire sales for long cycles, with legal and procurement in the middle (Appify).
  5. Leave the old small customers on a free plan or with basic support, instead of cutting them off all at once.

Best example

Appify realized the product only truly solved the problem for large companies and aimed at enterprise. The average monthly ticket went from about US$ 400 to about US$ 10 thousand, with about 15 enterprise customers and one contract of US$ 336 thousand a year (vídeo).

Numbers that came up

  • GUS: from 35 to 45 customers, average ticket of about US$ 7.5 thousand a month and revenue from about US$ 40 thousand to about US$ 100 thousand a month in one year (vídeo).
  • TrueFan: average ACV rose from about 800 to about 1,800 after closing 2 or 3 customers above 100 thousand a year (vídeo).
  • Crystal: B2B went from almost 0% to 60 to 65% of revenue, with a ticket of about US$ 8 thousand a year against US$ 600 for self-service (vídeo).
  • Way In: 152 customers with a ticket of about US$ 65 thousand a year and net retention of about 125% (vídeo).

When it does NOT apply

  • With few customers, each one weighs heavily on revenue (Nahas).
  • It requires a sales team and, often, an operation close to an agency, as at GUS and Crystal.
  • Small customers can still pay a relevant share of revenue: at AlgoPix they remain at about 40%.

Headline and top of the page talking about the customer's result and problem

18 founders used itin 6 of 7 channels1 with revenue cited

What it is

You rewrite the headline and the start of the page to say what the customer gains, in plain language, instead of listing features or internal terms. Results, before and after, and numbers take the place of the long explanation.

Why it works

The visitor decides in a few seconds. When the headline shows the result they want, they understand what they get and keep reading. A headline that describes the product or uses internal terms forces the visitor to translate it on their own.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

Your site lists features, uses internal terms or has a generic headline. It's one of the cheapest changes to test, with an effect that should be confirmed with real traffic.

Step by step

  1. Research what your best customer wants to hire the product to do (job-to-be-done) and pick just one.
  2. Write the headline with that result, in one sentence (example: "Wave goodbye to unexpected AWS costs", from CloudForecast).
  3. If the subheadline says better what the product does, swap the two (Versely).
  4. Leave only the headline, one or two explanatory sentences and a button at the top (Swipe Well). Cut the rest of that area.
  5. Make the button repeat the headline's benefit, instead of "Submit" or "Sign up".
  6. Test versions. With little traffic, run an A/B with only the best version against the current one.

Best example

On a social media tool, Georgiana Laudi swapped the site's message for a job and shortened the trial from 30 to 7 days. Site conversion rose 89% and trial-to-paid rose 40%, without touching the product (vídeo).

Numbers that came up

  • Late Checkout: the new page, with results in place of text, converted 20 to 30 times more than the old one, according to Neville (vídeo).
  • YNAB: after bringing the four-rules method into the headline, sales went from about US$ 100 to US$ 300 in a month, with US$ 450 in April and US$ 870 in May (vídeo).
  • Copy Hackers: optimized headline and button gave 10.3% more completed forms; in another test, the button "show me outfits I'll love" brought 124% more clicks to signup (vídeo).
  • SmugMug: swapping the feature list for benefits on the pricing page gave a double-digit improvement in revenue (vídeo).

When it does NOT apply

  • The promise has to be kept, otherwise it becomes a joke (Drip, Superhuman).
  • With little traffic you can't measure small differences, and testing only the headline brought little improvement in the Copy Hackers case.
  • Georgiana Laudi's gain also came from a more qualified audience coming in, with changes to the site and the trial.

Split service by size: self-serve for the small, salesperson for the large

33 founders used itin 3 of 7 channels23 with revenue cited

What it is

Small customers sign up on their own, pay by card and follow the self-service flow. The sales team handles only large accounts, those who ask for a demo or those above a certain size.

Why it works

A human salesperson only pays for themselves above a certain ticket. Self-serve keeps acquisition cost low on the small customer, and sales effort concentrates where the revenue is. At Algolia, about 300 enterprise customers account for about 80% of revenue.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

You have many small customers and few large ones. Self-serve must already work before you put salespeople on top of it.

Step by step

  1. Keep self-serve at list price and put a demo button on the site.
  2. Score the lead by team size and by what they connect (at Alo, if they connect Salesforce the sales team calls; if it's Notion, nobody calls).
  3. Define the segment cutoff (Help Scout: 11 users or more go to assisted sales; Showpad: up to 2,000 employees is mid-market, above that is enterprise).
  4. Only hire a salesperson with a quota if the ticket justifies it: SalesQualia cites about US$ 10 thousand a year as the floor.
  5. Take the self-service base, sort it by who bought more than one seat and call offering licenses for the whole team (Pipedrive).
  6. Organize the team by segment (SMB, mid-market, enterprise) as you grow.

Best example

At Pipedrive, the self-service base was sorted by who had more than one seat and the salesperson called from the top down. AOL had 20 seats and went to 100 after the call, and the salesperson earned hundreds of thousands of dollars in commission (vídeo).

Numbers that came up

  • Alo: with the first salesperson ACV rose 3x, with about 750 signups a week, over 35% trial-to-paid conversion and about 50% of demos becoming customers (vídeo).
  • Flipsnack: 28 thousand paying customers, a team of 94 people with about 6 salespeople, and the largest customer pays about US$ 200 thousand a year (vídeo).
  • Mention: small plan of about US$ 60 a month with no salesperson, and mid-market of about US$ 400 a month with 12 inside sales people (vídeo).
  • Help Scout: 92% of customers are self-service, and the group with 11 users or more drives expansion (vídeo).

When it does NOT apply

  • A cheap entry attracts many unprofessional users and churn gets high (SEMrush, TapClicks).
  • Enterprise takes months to close and support grows with the volume of small customers (video about Wufu and Drip).
  • With a low average ticket, like Kontentino's (about US$ 130 a month), a salesperson with a quota doesn't pay off.
14

Retention and expansion

Sell extra modules and products to existing customers

88 founders used itin 3 of 7 channels72 with revenue cited

What it is

You offer new products, modules or integrations to the base that already pays. The customer buys more without you spending anything to win them again. The account grows inside the base itself.

Why it works

A customer who already saw value in the first product trusts you. Selling to them costs far less than attracting a stranger. When the revenue coming in from people who stay passes the revenue leaving from people who cancel, the base grows on its own and churn stops weighing on you.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

You already have a second product or module ready and customers who see value in the first. It works best with mid-size and large accounts.

Step by step

  1. Start by selling an essential package, the simplest or cheapest product, and let the customer really use it.
  2. Track usage per account and revenue per cohort over time.
  3. When the team likes it, offer the next module, product or integration, preferably to mid-size and large accounts, which expand the most.
  4. Train the sales team to visit the existing base, and not only hunt for new customers.
  5. Build plans (essential, professional, enterprise) so the customer moves up a tier as they grow.
  6. Measure gross churn and expansion separately, so you know whether one covers the other.

Best example

Market Factory (about US$ 12 mi in ARR) sells more modules as the customer opens new markets. Customers use 20 to 45 of the roughly 120 modules, and net retention stays between 125% and 138% a year, even while losing about 10% a year (vídeo).

Numbers that came up

  • Qualified (US$ 210k MRR): a customer paying 30 thousand a year reaches about 40 thousand in the second year, with NRR of about 115% (vídeo).
  • Pendo: net retention above 130% with about 200 customers, selling a second product to coworkers at the same company (vídeo).
  • Asaas: net retention of about 170 to 180% a year, with almost all growth coming from people already in the base (vídeo).
  • Copado (40 million run rate): net retention above 140%, and 50% of last quarter's new revenue came from customers who came back to buy again (vídeo).

When it does NOT apply

  • If the base has high churn, expansion only plugs the hole: Numerator has a scraping line with high churn that pulls the average down (vídeo).
  • Small accounts expand little, so the expansion focus goes to the larger ones (vídeo).
  • Selling everything at once makes the customer pay for things they don't use and cancel (vídeo).

Land and expand: enter small and grow inside the same account

75 founders used itin 2 of 7 channels60 with revenue cited

What it is

You close a small entry contract, with few users, one unit or a pilot. Then, with usage and results, the account grows with more licenses, teams and products. The expansion happens without a brand new sale.

Why it works

A smaller entry lowers the purchase risk and shortens the decision. When the product shows fast results, other teams at the same company ask for access. The value delivered justifies the bigger packages, and the renewal becomes a growth conversation.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

B2B sales to companies with several teams, units or departments, where the product proves results within a few months and the customer has budget to grow.

Step by step

  1. Define a small entry: a pilot, one team, one use case, one dataset or a phase 1 rollout.
  2. Get usage to take off until all purchased licenses are active.
  3. In about 1.5 to 2 quarters, expand the same use case and then bring a second one to the same account.
  4. Open other units, countries or departments at the same customer.
  5. Track users, licenses or spend per customer, and measure net retention, not just cancellation.
  6. Have someone dedicated to adding value in each large account.

Best example

Sumo Logic (about US$ 100 million in revenue) starts the customer with one use case and one dataset, expands in 1.5 to 2 quarters and then brings another use case. Average net retention was above 150% a year across 20 cohorts, with about 1,300 paying customers (vídeo).

Numbers that came up

  • Pipefy: Accenture started with 4 users, passed 400 on the platform in six months and grew without a new sale (vídeo).
  • Birdzi: customers who came in at US$ 10 to 20 thousand a month moved to US$ 50 to 60 thousand a month within a few years (vídeo).
  • Yellow.ai: customers who came in at 30 to 40 thousand in ARR in the US reached 600 to 700 thousand, with NRR of 150% to 160% (vídeo).
  • Rapid7: initial contract of about 20 to 50 thousand dollars and expansion of about 120% a year (vídeo).

When it does NOT apply

  • If the customer has only one app or one unit, there is nowhere to expand (vídeo).
  • The sale is slow and cash suffers at the start: Birdzi closes about two sales a year per executive (vídeo).
  • Doubling or tripling spend is not sustainable forever, and Jebbit's own CEO admits growth tends to flatten (vídeo).

Get new users to their first win by cutting onboarding friction

45 founders used itin 6 of 7 channels16 with revenue cited

What it is

You find the shortest path to the moment the user sees value (connect, import, create, share) and push them there on first use. Everything that doesn't lead to that moment gets out of the way, like setup screens and too many choices.

Why it works

A user who doesn't see value early disappears, and churn in the first days can be huge. Someone who completes the key action builds a habit and switching costs. That is why the product asks right away for the action that creates value, instead of leaving the person stuck in front of an empty screen.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

A product with a short essential setup, where there is a clear, measurable "wow" moment. Worth doing right at launch and in any product with high early churn.

Step by step

  1. Map 4 to 5 key moves the user needs to make in the first week (or the 2 minimum steps to the "wow").
  2. Cut everything between signup and that first result: setup screens, choices and confirmations.
  3. Ask for the right action on first access: connect the account, import the data, create the first item.
  4. Measure time to first value and the full activation rate.
  5. Track who still hasn't done the key action and send a reminder.
  6. Shorten the time gradually, with a clear target (Rise Vision aims for 5 minutes).

Best example

At Drip, the path to the "wow" was two steps: install the tracking code and create an email sequence, guided by an assistant inside the product. Of 3,400 people, 60% completed full activation (vídeo).

Numbers that came up

  • Chatmatic: with a ready-made trigger as the first action, 80 to 85% of users got at least one subscriber and about 60% were active after 7 days (vídeo).
  • Wave: churn below 2% a month, with no salesperson in onboarding, by defining the actions that show success on day one (vídeo).
  • Coda: forces the user to share a document during the two-week trial, and over 80% of active users use it 3 or more days a week (vídeo).
  • Omie: churn of about 1.7% a month, focused on the moment the customer stops typing (vídeo).

When it does NOT apply

  • If the product requires long setup or the "wow" depends on something hard to force (Everlance needs location permission and the person has to go out driving) (vídeo).
  • Cutting setup screens only works with a single-function product (vídeo).
  • Charging for the setup service right at the start was a mistake at Drip: the value is in getting the customer to use it (vídeo).

Founder's personal support for the first customers

25 founders used itin 6 of 7 channels12 with revenue cited

What it is

In the first months, the founder personally answers chats, tickets and direct messages and calls customers. Every problem becomes a lesson for the product. The customer feels they are talking to the person who decides.

Why it works

Talking directly to the customer reveals real problems before they turn into cancellations. Well-served customers renew, buy more and refer others. Each manual help also becomes an idea for a new feature, and the roadmap is born from what real customers ask for.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

Few customers and a new product, until revenue grows a lot. One example in the material did this until each product reached US$ 10 thousand a month.

Step by step

  1. Put your direct contact (chat, DM) on the support link and answer every paying customer.
  2. Message each new customer: introduce yourself as the founder and ask what they thought.
  3. Reply fast and fix things within minutes when possible, even small problems.
  4. When you help someone manually, write down the steps and turn them into a product feature.
  5. If something is delayed, call each affected customer and explain the new deadline.
  6. When support passes about 4 hours a day, hire someone and measure response time and satisfaction (CSAT).

Best example

Supergrow sold a US$ 65,000 LTD in 3 days. The founder answered customers in chat for 1 to 2 hours a day and turned every manual help into a feature. About 10% of buyers used the product, and 20 to 25 customers a week helped shape the roadmap (vídeo).

Numbers that came up

  • Linjer: called the 350 customers of a delayed bag batch, and most were understanding (vídeo).
  • Teachable: response time dropped from almost 8 hours to under 1 hour, and CSAT went from 85 to the 90s (vídeo).
  • Yaphone: direct message to every paying customer, with over 10 thousand signups and 4,500 paying customers in 7 months (vídeo).
  • Block to Pin (US$ 16K MRR): over 400 active subscribers with churn of about 10% (vídeo).

When it does NOT apply

  • It doesn't scale: the founder makes the contact and the team still can't (vídeo).
  • It eats too many hours: Teespring's founder dropped direct email when it passed 4 hours a day (vídeo).
  • Teachable doesn't prove that support caused the churn drop, only that the two moved together (vídeo).

Human onboarding with individual contact and follow-up

35 founders used itin 7 of 7 channels19 with revenue cited

What it is

A person (consultant, customer success, coach) takes the new customer to their first result, with a call, training or setup meeting. They cover what self-service lets slip.

Why it works

Products with hard setup lose customers before they see value. When someone guides the setup and sets goals together with the customer, they reach results faster, leave less and even refer others. Human time pays for itself when the ticket is high enough.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

A product that requires setup, data connection or process change, or a ticket that pays for human time. It also helps when the pilot needs to become an annual contract.

Step by step

  1. Include an onboarding call or session in the plan (Swiftpage gives 20 minutes to every new customer).
  2. At signup, ask for whatever the setup needs (credentials, connections) and book the call.
  3. Define with the customer 5 to 10 KPIs or the first leader's goal, and build the dashboard together.
  4. Follow the first steps until the first result (first campaign, first closed deal).
  5. Stay in touch afterward, with webinars and content, or with a relationship manager per account.
  6. Work out the cost per customer before scaling the team.

Best example

Swiftpage (about US$ 90 million in ARR) includes a 20-minute call with an onboarding consultant in the plan and then keeps in touch with webinars and content. With about 85 thousand customers, churn is about 1.5% a month, and the success team has about 70 to 80 people (vídeo).

Numbers that came up

  • Patreon: connecting the creator with a person at the right moment improved first and second month revenue by 25% (vídeo).
  • GetAccept: personal onboarding even on a two-user contract, with logo churn below 1% a month and about 4,700 customers (vídeo).
  • WorkBoard: coaches align the first leader's goal, with an average ticket of 125 thousand dollars a year and NRR near 140% (vídeo).
  • Inspection Expert: with a dedicated success person, annual revenue churn dropped from about 10% to under 8% (vídeo).

When it does NOT apply

  • With a low ticket, the human cost doesn't pay off: Swiftpage has ARPU below US$ 150 and an expensive operation (vídeo).
  • If the customer doesn't connect the data or answer the call, there is nothing to do for them (vídeo).
  • In some cases it requires a data team and coaches, and the human part may be only the testing phase (vídeo).

Focus on the ideal customer, filter out who doesn't fit and let the rest go

41 founders used itin 5 of 7 channels26 with revenue cited

What it is

You look at customers one by one, not just total revenue, to see who stays and who leaves. Then you define the ideal profile, filter out who doesn't fit before selling and stop serving people who will never work out.

Why it works

A customer outside the profile cancels fast and consumes team time. When you find out who stays, you concentrate marketing, sales and product on that group. Churn drops and customer lifetime value rises, even if the base shrinks at first.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

Logo churn above what you want or a base with mixed profiles, with enough customers already to cross-reference data (the top 10 or 100, for example).

Step by step

  1. Once a year, pull data on your best customers and look for patterns: industry, size, who referred them.
  2. Classify each cancellation as ideal or non-ideal customer, and break the data down by segment.
  3. Write the ideal profile (ICP) and create an entry filter: confirm traffic, a minimum marketing team or income to pay.
  4. Raise the price floor or charge a setup fee, and require a demo to get in.
  5. Stop investing in segments that don't fit, without needing to cancel anyone.
  6. Track net retention by cohort, not just gross churn.

Best example

Market Maker Leads (about US$ 200 thousand in MRR) started finding out whether the real estate agent had the income to pay and explaining how long it takes to see a return. Logo churn dropped from 18% to 20% a month to about 3% (vídeo).

Numbers that came up

  • Infusionsoft: monthly churn dropped from 8% to about 2% after a setup fee came in (vídeo).
  • Platform OS: annual churn of 40 to 60% dropped to about 5% in 12 months, by encouraging customers who didn't fit to leave (vídeo).
  • AppSumo: over 80% of top customers were marketing agencies in 2017, and focusing on them tripled LTV (vídeo).
  • Tiny Seed: Gather went from US$ 29 to about US$ 200 to 250 at the floor in 18 months, and Cart Hook started charging US$ 300 a month with a mandatory demo (vídeo).

When it does NOT apply

  • It reduces sales at the start, and short-term revenue falls: it is a conscious trade-off (vídeo).
  • It needs a customer base to analyze. Without a base, there is nothing to cross-reference (vídeo).
  • Part of the churn drop comes from cutting customers, and Platform OS's founder lost short-term revenue to get there (vídeo).
15

Operations and team

Six fundamentals, ordered by strength of evidence. Every number comes from real interviews, with the video link next to it.

Focus on a few priorities and cut the rest

31 founders used itin 5 of 7 channels5 with revenue cited

What it is

Pick the few things that move the result, go deep on them, and leave the rest out on purpose. The short list applies to the company, to each person, and to each week.

Why it works

When there are more ideas and projects than people, every extra item splits the energy of the items that matter. Cutting the rest concentrates time and attention on the few decisive items, and execution on them gets faster.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

When you have more ideas, projects, or requests than capacity to carry. It works from solo founders to large companies.

Step by step

  1. Gather everything on the table and cut until one or two ideas remain that truly increase revenue.
  2. Write at most three company goals and choose the one that wins when the others conflict. Each goal has a single owner, with a name.
  3. Every Monday, gather the leadership and define one to three things that must ship that week. Agree beforehand which fires can keep burning for now.
  4. When someone asks for everything, show the menu with the cost of each item in time and focus, and ask the person to choose and record the decision.
  5. Once a quarter, ask the team: "if we had no problems, how would we advance the strategy?". List a few priorities and the obstacles for each, on one page.
  6. Before accepting a service project, ask whether it leads to recurring revenue. If it does not, turn it down.

Best example

LCB Tracker runs a six-month planning cycle. In June the board sets three priorities, in July each department sets its own, in September it locks four to six objectives with KPIs, and in November everyone signs off on execution. The company went from US$1 mi in 2011 to about US$25 mi a year and stayed within 1% of its revenue and expense targets in each of the last six years (video).

Numbers that came up

  • Revolut: about 100 projects in progress, and the executive goes deep on 7 to 10 (video).
  • Facebook: three goals for five years (growth, engagement, and revenue) (video).
  • Mixpanel: 10 to 12 bets per half-year, instead of 50 (video).
  • Ramp: the argument was to execute 4 items faster rather than 8 (video).

When it does NOT apply

  • It only works if the whole team agrees on what to let burn. Without agreement, everyone interprets it their own way (video).
  • It requires autonomy and trust between teams. When a team is struggling, it becomes the next focus (video).
  • Running the strategic review more often becomes overhead. Roughly every 100 days the focus gets lost and has to be redone (video).

Build the team in a lower-cost country

47 founders used itin 5 of 7 channels35 with revenue cited

What it is

Hire the technical team (and sometimes support and sales) in regions where salaries are lower or the talent pool is deeper, working remotely or from a local office. Headquarters keeps leadership, sales, and marketing.

Why it works

Salary and talent availability vary a lot from region to region. The same budget pays for more people, or more senior people, where the cost of living is lower. The result is the same technical team for a fraction of the cost.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

When the work can be done remotely and the local market is expensive or shallow. It shows up from pre-revenue companies to 100k+ MRR.

Step by step

  1. Choose the region by cost, English level, and time zone. Eastern Europe and India show up several times in the evidence.
  2. Find the first technical leader there through someone who already knows the country: a local cofounder, a CTO with a network, a country dev portal, or Upwork.
  3. To probe the market, list the largest software houses in the country, write to each CEO, and ask for devs under a team augmentation contract.
  4. If you hire directly, open a local entity first. Count on about 6 weeks per paperwork step and nothing productive in the first months.
  5. Give equity or a bonus to those who stay and, if retention weighs heavily, open a local office.
  6. Keep sales, marketing, and leadership close to the customer and manage the junior team closely.

Best example

Cognism chose an Eastern European region with fluent English and a similar time zone. It has about 60 people in Macedonia and Croatia, at about 500 euros per person per month, out of a total of 179 people. The CEO says it comes out cheaper than India (video).

Numbers that came up

  • BlueSky HQ: 17 of about 20 people in India. The founder's off-the-cuff estimate is about US$1 million a year there versus about US$10 million in New York (video).
  • Design Pickle: about 470 full-time designers, paid US$3 to US$6 per hour, with a gross margin of about 68% in a good month (video).
  • CrawlQ: 10 people for about US$7.500 per month (with AWS), for revenue of about US$88 mil per month (video).
  • Neon CRM: the budget for two junior devs in the US (US$45 to 50 mil each) became six devs with more than ten years of experience in Chengdu (video).

When it does NOT apply

  • A dev working from home tends to be less loyal, which is why Cognizant opened local offices (video).
  • A junior team demands strong management, and a distributed team requires management care (video, video).
  • There is an entry cost: paperwork, months without output, and lost efficiency from time zones (video, video).

Small fixed core and a variable rest by contract

37 founders used itin 5 of 7 channels23 with revenue cited

What it is

Keep only a small group full time, tied to the product and the customer. Support, design, marketing, prospecting, and maintenance go to freelancers, agencies, or vendors paid per deliverable.

Why it works

Fixed payroll cost weighs on you even when revenue drops. With a contract per project, per task, or per service delivered, cost follows revenue and the team goes up or down with demand. Outsourcing stays reversible: if the vendor does not work out, you swap.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

Before outsourcing, separate the tasks: whatever is close to the customer or touches revenue stays in-house. It works well in a lean operation, with the process already written down and trusted vendors.

Step by step

  1. List the tasks and separate what touches customers and revenue (stays in) from what can be delivered by contract (support, accounting, design, editing).
  2. Write the process (checklist or SOP) before handing the task off.
  3. Test: give the same task to 5 to 10 people with your SOP and pick the best deliverable.
  4. Prefer a few fixed vendors that work as part of the team, instead of many small ones. Assign one person responsible for each vendor.
  5. Pay per service delivered, per project, or per consultation, so cost varies with revenue.
  6. If outsourcing does not work in a function, hire directly and build culture.

Best example

ConnectAndSell operates with 33 full-time employees and about 500 people inside vendors that charge per service delivered. The CEO says that this way growth does not require new capital, and the company had a profit of about US$166 mil in 2019 (video).

Numbers that came up

  • DevStride: 7 full-time people and about 15 contractors. Accounting for about US$600 per month, versus US$4 to 6 mil for a fractional CFO (video).
  • Chrome Image Labs: 200 editors and designers paid US$5 to 10 per image, with a margin of 55% to 60% versus about 85% for a SaaS (video).
  • Verblio: passes about 50% of revenue to freelancers. With about US$260 mil per month, it keeps about US$130 mil, with a team of 12 people (video).
  • Real Defense: 300 people, 200 of them outsourced, for US$70 million in revenue (video).

When it does NOT apply

  • It depends on reliable vendors. Hiring a random firm did not work out for YouMail, and Rapid Funnel tested outsourcing in India and the Philippines and went back to hiring directly (video, video).
  • The agency can leave. The SaaStr founder took 4 years to find the right production agency (video).
  • Lower margin than SaaS, and less control over schedule and code (video, video, video).

Revenue or capacity trigger before opening a role

26 founders used itin 6 of 7 channels6 with revenue cited

What it is

Only open a role when there is a concrete signal: revenue covers the person's cost, the team says it cannot keep up, or a milestone was hit. Hiring in anticipation or to hit a headcount target is off the table.

Why it works

Hiring beyond what cash can support leads to layoffs later, and more people increase coordination and slow delivery. With a clear trigger, the team grows at the pace the product and the balance sheet can sustain. When the team has to choose between two things, prioritization gets tougher without new people.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

Whenever you are about to open a role, especially after raising money or during a period of fast growth.

Step by step

  1. Compare net monthly revenue with the new person's cost. If it covers it, hire. If it does not, wait another month or sell more.
  2. Calculate real capacity: add up meeting hours, vacations, events, and projects already promised, and compare with the working days in the quarter.
  3. Define a results milestone the team has to hit before the role opens.
  4. Write down the hypothesis the role will test and why AI cannot do that work. Only then hire.
  5. Hire in small waves (of 3 roles, open 2, see the result) and leave 3 to 4 months between one hire and the next.
  6. Whoever made the hire also decides the exit: if the hire does not improve after being warned, act fast.

Best example

Brumate did almost everything with agencies and freelancers and only had a full-time employee after US$20 million in revenue. It reached US$100 million in 2020 (video).

Numbers that came up

  • Pavilion: entered 2023 with 62 full-time employees and now has 27, after raising US$25 million and hiring at an accelerated pace (video).
  • WebinarNinja: one bad hire cost the company about US$500 mil (video).
  • Contentino: went from 18 to 45 people in about 10 months, with net recurring revenue falling over the period (video).
  • Distro (the interviewee's hypothetical example): with US$20 mil in net sales per month, a US$12 mil hire pays for itself. With US$8 mil, you wait (video).

When it does NOT apply

  • If there is a short market window (6 to 8 months in Compliable's case), doubling engineering can do in six months what would take twelve (video).
  • The trigger numbers are rules of thumb. Molly Graham says they are guidelines, not law (video).
  • Windsurf's rule only works with a team that trusts each other and accepts hard prioritization (video).

Recruit through your network and active outreach

26 founders used itin 6 of 7 channels5 with revenue cited

What it is

Instead of waiting for applications, you look for people who already do the work: on LinkedIn, Twitter/X, GitHub, niche communities, products you admire, and the networks of team members, partners, and investors.

Why it works

People with a rare profile are almost never looking at job listings. Searching where they already show their work (open source, screen redesigns, a blog, a profile with metrics) brings qualified candidates without waiting. Referrals from people you know also filter out those who just know how to pass a form.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

When you need a specific or rare profile. For junior roles, the volume of a large job post may still be necessary.

Step by step

  1. Define the exact profile (for an engineer, the language) and the list of who already does it: the 200 best in the market, those who built products you admire, those who maintain a popular open source project.
  2. Search where the person shows up: LinkedIn (filter by title and use terms like MRR and ARR), GitHub and Stack Overflow, Twitter/X, Google (pages 2 to 4).
  3. Write a short message: who the company is, what the role does, and the salary range in bullet points. Avoid generic recruiter text.
  4. Ask your network: post that you are looking for people and help those who are job hunting, ask partners, investors, and the team for referrals, and consider a referral bonus.
  5. Create a Slack channel for the team to post impressive profiles, and run sessions every two weeks to map each person's network.
  6. When you have cash, hire a good recruiter. The commission cited is about 20% of first-year salary.

Best example

The founder of dpad posted on LinkedIn that he was looking for people and helping those job hunting. The post got about 650 comments from candidates, four roles were filled at no cost (two had been open for months), and he introduced the candidates to founders in his network (video).

Numbers that came up

  • One recruiter's message templates (clients Rippling, Keeper, Gorgias): a response rate of 60% or more, about 2.5 times more replies than the standard (video).
  • Quantagonia: 7 of 12 employees are engineers, recruited through the partners' network of professors and researchers (video).
  • Clutch: the cofounder came from an open source repository with about 10 thousand stars, and the team of 7 joined without a jobs page (video).
  • JT Barnett (agency model): 1,500 interested people in one night, a base of 7,000 creators (video).

When it does NOT apply

  • For engineers, LinkedIn does not work well. The recruiter points to GitHub and Stack Overflow (video).
  • It depends on network and trust. Those without a network need a talent platform with paid vetting (video, video).
  • It is manual, slow work that eats the founder's time, and AI helps but does not replace personal contact (video, video).
16

Fundraising and exit

Delay fundraising until you have traction and know what the money is for

34 founders used itin 5 of 7 channels25 with revenue cited

What it is

You grow on your own revenue and only go to investors once you have paying customers and a clear plan for the money. The check arrives after the main risks have already been resolved.

Why it works

Every risk you eliminate before raising reduces what the investor needs to charge to take on the risk that's left. With less risk, you sell a smaller slice of the company for the same money and negotiate as an equal. Founders who raise without traction give up more equity and end up hostage to the pace the investor demands.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

When the business can grow without outside capital, from the first customer up to somewhere between US$ 500 mil and US$ 2 milhões of ARR. It works best in software, which doesn't need much cash to run.

Step by step

  1. Build the product and pre-sell to customers before looking for an investor (BetterAgency reached US$ 1M of ARR this way).
  2. Reinvest the revenue and work out your unit economics until the model repeats.
  3. Write down what each round is for: the first pays for a prototype and a small team, the second proves the business is predictable before you hire salespeople in volume.
  4. Raise a small round, giving up little equity, once you already have revenue (examples below: 500 mil for 10%, less than 20% for US$ 5 mi).
  5. Revisit the decision to raise every six months, as Backblaze did, and only raise when you're short on cash for something concrete.
  6. Prefer to turn down a round that demands a pace before you know where product-market fit is.

Best example

Arkadium built a profitable business (revenue of about US$ 12 mi, EBITDA margin of 10% to 20%) and only then sold less than 20% for US$ 5 mi, with a pre-money of about US$ 20 mi (vídeo). Because it already had profit, it negotiated better and sold a small piece.

Numbers that came up

  • BetterAgency: US$ 2,1 mi raised at a US$ 15 mi post-money valuation, with about US$ 1 mi of ARR and 260 customers (vídeo).
  • EyeContact: bootstrapped to about US$ 1 mi of ARR, then raised 500 mil, less than 1x ARR (vídeo).
  • SwagUp: every risk eliminated before the seed saves 5% to 10% of the company, because the investor was asking for 20% to 25% with nothing proven (vídeo).
  • Rob Walling: valuation of 3 to 4 milhões with 20 to 30 mil of MRR, and a 500 mil round for about 10% (vídeo).

When it does NOT apply

  • Businesses that need a lot of cash, like infrastructure. Backblaze was forced to raise when the price of hard drives jumped after a flood in Thailand (vídeo).
  • Physical products: Andy Dunn warns that the rule of selling to the network up to US$ 50 mil of run rate does not apply to SaaS (vídeo).
  • Founders who have what it takes to play the venture game: Patrick Campbell says he should have raised earlier to grow faster (vídeo).

Prepare the sale with finances, contracts and a data room ready

20 founders used itin 5 of 7 channels4 with revenue cited

What it is

Before putting the company up for sale, you gather contracts, numbers and documents in one place and check that everything adds up. The buyer will ask for all of it in due diligence, and whoever already has the folder ready answers in days.

Why it works

Diligence turns the whole company inside out. A missing document, an unsigned contract or revenue that doesn't match the bank statement becomes a delay, a request for a discount or even the end of the deal. Showing numbers and risks before the price is locked removes surprises from the table and the deal moves faster.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

Before listing or selling the company. The videos recommend starting early and updating every quarter, even with no plan to sell.

Step by step

  1. Gather contracts for all employees and contractors, with signed IP assignment and NDA, plus the cap table and documents for every investor, even the small ones.
  2. Run a code scan to find open source licenses (look for GPL) and check sales tax nexus by US state and the revenue recognition policy.
  3. Reconcile revenue with Stripe, PayPal and bank statements, and give the buyer direct access to Stripe, the books and Google Analytics.
  4. Build a data room of about 10 files: deck, 3-year financial model, financial history, cap table, top 20 customers and why they buy, team, roadmap, key contracts, IP.
  5. Register the trademark for the name and separate the business from your other accounts (email marketing, hosting).
  6. Show the buyer the risky revenue sources before locking the price, and hire a broker to hold the hard line in the negotiation.

Best example

In the ScrapingBee sale, diligence was "issues 24/7 basically for four weeks". One contractor had no IP agreement and the founder had to invite him to dinner to get it signed (vídeo). The contrast is MeetEdgar: 100% of revenue came from Stripe and diligence took 30 days, with no forensic accountant (vídeo).

Numbers that came up

  • Preliminary fundraising diligence asks for 7 to 10 items, an exit one about 50 to 60 (vídeo).
  • Sale of a 40-employee company: three months of 12 hours a day to find the documents, after switching payroll providers three times (vídeo).
  • Bingo Card Creator: the buyer found a US$ 29,95 discrepancy and the process took four months (vídeo).
  • ZyraTalk, with great metrics: 22 meetings with PE funds and zero LOIs, and the sale went to a strategic buyer (vídeo).

When it does NOT apply

  • Small businesses: Patrick McKenzie says a US$ 500 per month company doesn't pay for the work of selling (vídeo).
  • The list of documents and payroll vary a lot by country and structure, so adapt the American checklist (vídeo).
  • The smaller the business, the lighter the process: selling for 1 to 2 milhões is less work than for 20 to 30 milhões (vídeo).

Price the company with a profit or revenue multiple

24 founders used itin 5 of 7 channels14 with revenue cited

What it is

You arrive at the price by multiplying a base (adjusted profit, recurring revenue or net revenue) by a multiple the market pays for your type of business. Small companies are usually valued on profit, and larger SaaS on recurring revenue.

Why it works

The buyer pays for the cash the business returns and for how predictable it is. When you split revenue by type (SaaS, services, media) and adjust profit for one-off expenses, the number becomes defensible, and every item removed has an explanation. A price with math behind it holds up in negotiation.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

When buying, selling or raising a round with a valuation. Multiples change with size, growth and type of revenue.

Step by step

  1. Calculate SDE (revenue minus operating costs, without your salary) and remove one-off expenses, explaining each item to the buyer.
  2. If revenue is recurring, multiply MRR by 12 to get ARR and apply the multiple for your range.
  3. If revenue mixes types, value each part with its own multiple and add them up (SaaS and services, or SaaS and e-commerce).
  4. Compare with the median multiple of public software companies and with the buyers in your niche before fixing the number.
  5. Ask for the price at the top of what the market accepts and expect the buyer to push back.
  6. Run the market math (possible customers x ticket) to check that the plan supports the valuation.

Best example

Derrick Reimer sold Code Tree in 2016 asking 3,5x SDE. The buyer thought it was high, but closed at the asking price of US$ 128 mil (vídeo). He himself notes the data is old and thinks the multiple today is close to 5x.

Numbers that came up

  • Median of public software companies: 5,5 to 6 times recurring revenue. PAR bought at about 5x ARR (vídeo).
  • PE buying B2B SaaS with 1 to 2 milhões of ARR: about 3x with 0 to 20% growth, 4x to 5x at 20% to 50%, 5x to 7x at 50% to 100% (vídeo).
  • Unilog: SaaS valued at 4 to 6 times revenue and services at 1,1 to 1,2 times, added together to reach the pre-money (vídeo).
  • Flippa: average multiple of about 3x SDE on SaaS aged 2 to 4 years, and a listing with US$ 8.800 of profit per month at US$ 450 mil, about 4,2x (vídeo).

When it does NOT apply

  • A specific buyer's multiple, or one from the guest's memory, is a reference, and the final price depends on demand in the process (Unilog, iView) (vídeo).
  • Businesses that depend only on the owner or need heavy equipment investment: real cash drops, and the adjusted profit multiple misleads (vídeo).
  • For selling a small company, the revenue multiple doesn't work; use the profit one (vídeo).

Choose an investor for what they bring beyond the money

31 founders used itin 3 of 7 channels22 with revenue cited

What it is

When raising, you compare investors by the network, experience and access to customers each one brings, and look at valuation afterward. It can be a strategic investor from your industry, an operator who has already scaled a company or a fund with operating partners.

Why it works

An investor with a network opens doors to distribution, sales and credibility that the check alone can't buy. Whoever sits on the board follows the company for years and helps avoid mistakes. That's why several founders accepted a lower multiple with the right investor.

›Full playbook: steps, best example, numbers, caveatsHide details

When to use

In any round where you have more than one offer, or when your sales channel depends on relationships in the industry (enterprise, healthcare, logistics, finance).

Step by step

  1. List what you need beyond cash: customers, contacts, a board, industry knowledge.
  2. Look for investors who are already in your customers' ecosystem (corporate VC, a regional software company, a big supplier in the industry).
  3. Ask who exactly will sit on your board and ask for references from other founders that person has worked with.
  4. Compare offers by their terms (like liquidation preference) and by who comes in, and choose a multiple you can sustain in the next round.
  5. Ask about time horizon and liquidity before accepting, and prefer whoever doesn't talk about selling in the first few years.
  6. Align expectations after the round: not every investor will help, so stay in frequent contact with the ones who do.

Best example

LemonEdge ran a proof of concept with a large company in the industry and, when it was impressed, extended the seed with Blackstone as a strategic partner: about US$ 4 mi extra, total of about US$ 6,5 mi, with 10 to 20% sold (vídeo).

Numbers that came up

  • Celigo: series B at 6,3x, with offers from 5 to 10 of the 25 to 30 investors approached. Salesforce paid 21x for MuleSoft (vídeo).
  • Clipfolio: US$ 12 mi round at a US$ 45 mi pre-money (7x to 8x recurring revenue), turning down offers with a higher multiple and heavier terms (vídeo).
  • Sudo.ai: 42 nos before the first yes in the first round, US$ 2,3 mi seed at US$ 8 mi pre-money (vídeo).
  • B2Chat: US$ 750 mil pre-seed in a convertible note with a cap of about US$ 7,3 mi, with a large software company from Argentina (vídeo).

When it does NOT apply

  • Even when hand-picked, half of the investors will be of little use. Paul Tran, of Native, set expectations because each one has their own life (vídeo).
  • A strategic investor charges a lot: the owner of Booty Bands gave 50% for US$ 100 mil and called it a "big slug", an example of high cost, nothing to follow (vídeo).
  • The benefit comes from the person on the board, and the person changes. The fund's name alone guarantees nothing (vídeo).

Isolated bets

Cited by a single founder, but with strong proof and revenue on record. Not confirmed by other sources, which is exactly why they are worth a look.

A 4-week study plan

  1. Today, 30 min: read the "interview your audience about today's problem" card and watch its first video. Done when you have written the 4 questions you would ask.
  2. Week 1 · Idea and product: your own problem, pre-sales, simple offer, search demand, manual service, lean MVP. Done when you have one validation method with a success bar and an MVP described in 3 features.
  3. Week 2 · Pricing and sales: annual billing, stepped price increases, discovery before demo, outcome-led headline. Done when you have a price table with monthly and annual.
  4. Week 3 · Organic acquisition and launch: communities, audience before product, building in public, niche SEO, marketplaces, waitlist. Done when you have 2 channels with a weekly cadence and a launch date.
  5. Week 4 · Paid, retention and operations: CAC ceiling, payback, stepped budgets, first user win, founder-led support, hiring triggers. Done when you can compute your max CAC and your product's win moment.
  6. Structural: fundraising, exit and large-SaaS tactics (add-ons, per-seat or usage pricing), once you have a product with revenue.

What not to do

  • Asking "would you use this?" or pitching the idea in a validation call: the data comes out biased.
  • Scaling paid media without knowing your CAC ceiling and payback.
  • Opening new fronts before the core sells.
  • Hiring before a revenue or capacity trigger.
  • Raising money before proving traction and knowing exactly where the money goes.

Limitations

Numbers are self-reported by founders in interviews and were not audited. Tactics were extracted from transcripts by a language model; a 60-video sample was checked by hand. Nathan Latka's show asks every guest the same pricing and operations questions, so those topics are over-represented in raw counts; the ranking weights cross-channel repetition to correct for it.

Sources: long-form videos from Starter Story, Greg Isenberg (Startup Ideas Podcast), Nathan Latka (Top Founders), MicroConf, Rob Walling, Lenny's Podcast and My First Million, published up to October 22, 2026.

We acquire and scale online businesses

LTV applies this playbook to the companies we buy. If you run or own a SaaS and want to talk, reach out.

Get in Touch →