Treat the paywall as the only honest product-market-fit test

I ran marketing and day-to-day operations for Cal AI as we scaled from nothing to $50M ARR in roughly 18 months. Here are the 34 things I wish more founders understood before their consumer app started scaling (everything we learned the hard way): 1. Your paywall is a lie detector A paywall does more than collect money. It tells you whether the promise was compelling, the onboarding increased desire, the value was understood and the offer made sense. A weak paywall conversion rate does not automatically mean “redesign the paywall.” The lie may have started in the ad, the App Store page or the onboarding. The paywall is simply where the user finally tells you the truth. 2. You CANNOT scale a broken machine When an app stalls, every founder asks the same question: “How do I get more users?” Wrong question. First figure out where the machine is broken. Not enough attention? It’s a distribution problem. People see the content but never visit the App Store? It’s a creative or positioning problem. They start onboarding but never reach the paywall? It’s an onboarding problem. They reach the paywall but don’t buy? Your offer is the problem. More traffic poured into a broken funnel just helps you lose faster. 3. Consumer apps are not tiny SaaS companies Consumer apps can have massive audiences and relatively low switching costs. That makes acquisition the #1 priority. B2B tools have fewer qualified buyers, painful implementation and much higher trust costs. That makes reliability and retention #1 priorities. Both businesses need acquisition. Both need retention. But the order of operations changes with the market. Stop importing B2B playbooks into consumer and wondering why the math doesn’t work. 4. Higher conversions can make the business harder to understand Founders optimize trial length for one thing: conversion. But a longer trial creates more time for other campaigns, organic content and outside events to influence the user before they pay. Your conversion rate goes up while your ability to identify what caused the conversion goes down. That is not automatically a win. The best metric in isolation can still create a worse decision system. 5. More app downloads can hide the fact that you don’t have a business A free app with a million downloads can still have zero product-market fit. Downloading proves somebody was curious. Paying proves the promise was valuable enough to act on. Staying proves the product delivered enough value to justify the decision. Those are three different signals. Only two of them can support a business. 6. Every new feature creates another place for the company to break Founders think adding a feature only changes the product. It also changes: The marketing promise The App Store page The onboarding The paywall The support burden The retention loop The number of things your team has to diagnose More functionality also creates more complexity. Earn the right to expand by making the core job undeniable first. 7. Power users are often seasonal A user can obsess over a fitness app while training for an event, then disappear when the event ends. They can live inside a budgeting app while paying off debt, then barely open it after reaching the goal. That does not mean the product failed. Engagement often follows the intensity of the underlying pain a user is trying to solve. Before building another retention feature, ask whether the user’s goal simply became less urgent. 8. Your most-used feature deserves more attention than your roadmap The fastest way to improve retention is usually NOT adding another feature. Find the action retained users already repeat, then simply: Make it faster Make it clearer Make it more reliable Make the result more valuable Founders love roadmaps because shipping new things feels like progress. Users stay because the thing they came for keeps getting better. 9. An influencer’s fee is not the actual cost A lower-fee creator can be more expensive than a higher-fee one. The invoice is only one line item. Add the time spent chasing replies, negotiating terms, revising content, rescheduling posts, fixing links, collecting analytics and managing usage rights. Then ask whether they can become a repeatable partner or whether your team has to start from zero again next month. Creator CAC includes management bandwidth, so price it that way. 10. Your influencer rates are set by brands outside your category Your biggest competitor for a creator deal may not be another app. It might be a high-margin physical-product company that can afford to pay far more for the same audience. A niche can have perfect product fit, trusted creators and great content formats…and still be a terrible acquisition channel. Don’t just ask whether the creator fits your app. Ask who else is bidding for their attention and what those companies can afford to pay. 11. Not every app should use seamless integration “Make the promotion feel organic” is incomplete advice. If the product is instantly understood and the magic moment looks good on camera, let the creator weave it into their normal content. If the value requires context, a subtle integration may leave the audience confused. Use a direct problem-first explanation or demonstration instead. Native is good. Confusing is not. The product determines how hard the creative needs to sell. 12. Your loudest users are not always your most important users Power users send detailed feedback because they care enough to think about the product all day. That makes their insight useful. It also makes them dangerous if you blindly build everything they request. Most users do not want 40 settings and an advanced dashboard. They want one outcome with as little effort as possible. Listen to power users, but build primarily for the quiet basic-user majority. 13. Manual work should come before automation You cannot automate a process you do not understand. Run it manually until you know: Which steps repeat Which exceptions matter What good performance looks like Where people routinely get stuck Which decisions require judgment Then automate the predictable parts. Founders who automate too early just scale their own messy setup. 14. A system is useful because it exposes the bottleneck A creator spreadsheet is not valuable because it looks organized. It is valuable because you can see where your systems are failing. Not enough replies? Fix the outreach or the list. Creators agree but never post? Fix onboarding, scheduling or incentives. Posts go live but do not perform? Fix selection, integration or the angle. Winners never renew? Fix follow-up and deal structure. 15. The person who makes the problem sound simple probably understands it best Be wary of anyone who turns your app problem into a 15-step proprietary framework before looking at the basic numbers. Complexity can be real. It can also be part of the pitch. The harder a problem sounds, the less likely you are to solve it yourself and the more likely you are to hire the person explaining it. Real expertise usually compresses the problem into the few variables that actually matter. 16. Speed without risk judgment is recklessness Move fast on reversible decisions. Test the angle. Ship the landing page. Try the creator. Change the onboarding. But legal, safety, financial and high-cost decisions deserve a different threshold. “Move fast” is not permission to ignore downside. Speed is a weapon when you know which mistakes you can survive and how high your tolerance for failure is. 17. Your calendar should expose the current constraint Not every completed task moves the business equally. If distribution is broken, another product meeting will not save you. If people click but never pay, more impressions will not save you. If the core feature fails, a new logo or design definitely will not save you. The founder’s time should flow toward the bottleneck tied to survival, revenue or learning. Look at your calendar. Does it match the problem you claim is most important? 18. Hunger and skill have different minimums by role For an early marketing hire, I will often bet on the hungry person who learns quickly and refuses to let the opportunity slip. Marketing has a lot of surface area where reps, taste and effort can compound fast. Technical work still requires a baseline no amount of motivation can replace. Do not use one hiring rule for every function. 19. The inbox preview is your first landing page Creator outreach has to earn the open before it can earn the deal. The recipient should understand four things immediately: 1. This is paid. 2. You chose them for a reason. 3. The ask is clear. 4. The next steps are easy. If someone needs a call just to understand the opportunity, you created friction before the relationship even started. 20. The visible tactic is almost never the moat People copied Cal AI’s influencer strategy constantly. I never worried much about it. “Scale using influencers” was not the hard part. The hard part was finding creators every day, sending the outreach, negotiating deals, reviewing the content, tracking delivery, renewing winners, cutting losers and repeating the cycle faster than everyone else. Information is easy to copy, but sustained execution volume is not. 21. Flat creator fees buy upside AND variance I love flat-fee creator deals because one breakout post can create insane asymmetric upside. But two things can be true. The same structure also means you pay the full amount when a post dies. That is why follower count is not enough. You need realistic expected views, a read on performance volatility and a price that makes sense across many attempts, not only the one outcome you’re daydreaming about. 22. Freemium only works when paid value survives the free feature Ask yourself one question: If I give away the feature that makes people download the app, is there another outcome they will confidently pay for? If yes, freemium may work. If you have to invent six weak upgrades to justify the subscription, it probably won’t. Free should create demand for paid value. It should not eliminate the only reason to buy. 23. The customer relationship does not end when they close the app Please stop using email marketing like a receipt printer: Install confirmation Trial reminder Cancellation notice Done Meanwhile, email can be used to teach the user how to get value, introduce another use case, reconnect them to a goal and bring back someone whose timing changed. You paid to acquire the relationship. Stop acting like it disappears when the user closes the product. 24. Winback should reconnect the user to the original dream outcome A lapsed user already knows what the product does. They downloaded it because they wanted a result. So do not start with “WE MISS YOU” and a random discount. Start with the outcomes: Is the goal relevant again? Did their context change? Did your product improve where they got stuck? Winback works when the reason to return is stronger than the reason they left. 25. Community does not need to be permanent Your utility app does not need a feed. It may not need comments, followers or another half-built social network either. Sometimes community should exist only around a high-intent window: A shared challenge A small accountability group A team goal with a clear start and finish But still keep the core product simple and individual. 26. Badges and streaks amplify value; they do not create it A streak gives users another reason to repeat a behavior that already matters. A badge makes meaningful progress visible. A challenge adds accountability to a result they already want. But if the core action is useless, you’re just gamifying disappointment. Build in this order: Value → repetition → progress mechanic. Never the other way around. 27. A perfect system can make ambitious work too easy Two early Cal AI marketing team members left after we had systemized enough of the work that their jobs stopped challenging them. That taught me a painful lesson: Pay can keep someone comfortable. It cannot always keep an ambitious person engaged when the learning stops. Efficiency is great for the company. But if your best people outgrow the work, you need to create a new problem worth owning. 28. Complementary founders > identical killers The goal is not to assemble four versions of the same cracked founder. You need different forms of leverage: Someone who sees opportunities nobody else sees. Someone who reduces messy decisions to first principles. Someone who can build almost anything. Someone who turns the work into a system and gets it in front of people. The team wins when the strengths multiply. Not when everyone fights to own the same job. 29. Being a little successful can be more dangerous than being a beginner Beginners know they do not know anything. So they listen, work hard and ask questions. Then they win a little and ego creeps in. They stop taking feedback, assume the first win proved the whole worldview and protect the identity of being successful. The best operators I know eventually look like beginners again: Curious Hardworking Open to being wrong Remember that. 30. Understanding luck makes your success more repeatable Acknowledging luck does not make the work less impressive. It makes the next decision more intelligent. Great operators can separate: What they controlled What they earned What somebody else contributed What happened at the perfect time What could easily have gone the other way If you credit every outcome to your own brilliance, you will eventually try to repeat the wrong part of your “success”. 31. Money cannot manufacture long term interest I spent years in management consulting making more money than I had ever made before. And I hated almost every day of it. Then I joined Cal AI, worked longer hours for almost no money and somehow had MORE energy. You can force yourself to work hard for a while. But genuine interest is what gets you through the repetitive, unglamorous work after the novelty disappears. If every week requires a new productivity system, the problem may be the thing you chose to work on. 32. Reinvest after evidence, not ego Early traction is not permission to start acting rich. It is evidence that the company may deserve more resources: Strengthen the product Upgrade the technical foundation Bring in great designers Pay for the legal and financial help you were surviving without Fix the parts that break under scale The lifestyle rewards can wait. The market window may not. 33. AI multiplies skilled operators more than it replaces judgment Give the same AI tool to a beginner and an elite operator and the elite operator usually creates more leverage.... They know what to build They recognize weak output They know which detail matters and which one is noise They can connect the result to distribution AI lowers the cost of execution. It does not remove the value of taste, context or reps. In many markets, it will widen the gap. 34. Nobody is coming to identify the bottleneck for you You can read every playbook, copy every funnel, test every AI tool, and hire every consultant. None of it replaces an operator’s actual job: See the business clearly, find the current constraint, and pull the right lever. Then do it again when the constraint moves. That is the game. @John0t1jJohn you're welcome @codymclain facts If you enjoyed this article, you'll love my newsletter. Join 1,700+ weekly readers of the Operator's Notebook below (it's free): https://operatorsnotebook.beehiiv.com/ @oliverpour You just gotta be honest in your assessment and see where the incongruence lies

The old, outdated consumer app growth playbook looked like this: Build an app Market it Optimize conversion That playbook worked when building the product was hard and distribution was cheap, but the opposite is true in 2026. Only 4.6% of new consumer apps reach $10,000 in monthly revenue within two years of launching. The biggest problem for app founders today is building a system that can repeatedly turn attention into retained customers. I call this the consumer app growth stack, and it has eight layers: Demand creation Intent capture The promise The magic moment The access model The checkout Retention The feedback loop If one layer breaks, everything below it gets harder. Here is how to build each level, brick by brick. 1. Demand creation The majority of users are not scrolling TikTok hoping to discover a new calorie tracker, screen-time app, or meditation tool. You need to make the problem feel important before you can sell the solution. That is the job of distribution. Influencers can work well because the audience already trusts the person explaining the problem. Paid creative works because you can test dozens of angles and pain points quickly. Founder content works because it gives people a reason to care about the story behind the product. The mistake is treating distribution like a place to just talk about your app features. Nobody cares that your app can do 27 different things. They care that it can save them four hours, help them lose 15 pounds, or stop them from checking their phone 200 times per day. Sell the change in the customer's life. Then show how the product expedites that process. 2. Intent capture Apple says 65% of App Store downloads begin with high intent searches. This is why demand creation and demand capture need to work together. A customer sees three UGC or influencer videos about your product but doesn’t install your app. Two days later, they search your name in the App Store. If your competitor is bidding on the keyword, has stronger screenshots, or makes the outcome easier to understand, they can steal the customer you spent time and money educating Your intent-capture layer should include: Brand protection in Apple Search Ads Exact-match category keywords Competitor campaigns Custom product pages that are congruent with your different ads App Store screenshots that repeat the original promise and are outcome based, not feature based Web and AI search results that make the product easy to verify Creators create intent, an optimized funnel captures it, and your app store page converts it. Do not let those three teams operate like they have nothing to do with each other. 3. Keep the promise consistent Imagine an ad promises to help someone scan a meal in seconds. They click and see an App Store page talking about meal plans. Then onboarding asks about their long-term fitness goals. Then the paywall sells macro tracking. All four screens may describe real features. But the customer feels like they’re being sold four different products. Every handoff should repeat the same promise with more proof: The ad creates curiosity The store page confirms the outcome Onboarding personalizes it The product proves it The paywall sells continued access to it If click-through rate is high but installs are low, do not immediately blame the store page. The ad may be making a promise the product cannot support. 4. Get to the magic moment The magic moment is the first time a user experiences the reason they downloaded. For Cal AI, it happened when someone took a picture of their food and watched the calories appear seconds later. For an AI photo editor app, it may be the first successful edit that saved a user 15 mins and endless headaches. For a screen-time app, it may be seeing the first day of usage data and realizing how bad the problem is. Your marketing has one job: Demonstrate that moment with the least confusion possible. 5. Match the access model to the product RevenueCat found that hard-paywall apps convert downloads to paid users at a 10.7% median rate by Day 35 vs freemium apps that convert at 2.1%. How easily someone can access your app needs to match how the product creates value. Use a hard paywall when the outcome is clear, urgent, and expensive to deliver. Use a free trial when the customer needs to experience the result before trusting the promise. Use freemium when free users build a habit, create network effects, lower acquisition cost, or move naturally toward a second paid outcome or a network of upsells. The decision rule is simple: How much product does a serious user need to experience before they can make an informed buying decision? Then give them that amount. Not more because a competitor does it, and not less because you are scared they will leave. 6. Route the checkout based on intent Web-to-app funnels are exploding because they give operators more control over onboarding, pricing, payment methods, and customer data. They can also lower platform fees. But lower fees do not automatically mean more profit. Dipsea tested native in-app purchases against a nearly identical web checkout in an experiment reported by RevenueCat. The native flow generated 42% more initial conversions. Too many operators look at a 30% store fee and automatically assume the web will make them 25-30% more money, but they forget to factor in funnel friction. An app store customer can confirm the purchase with a face scan. A web customer may need to enter their email, card details, billing address, and then return to the app. The right question is not: Should we use web or in-app purchases? It is: Which customer should see which checkout at which point in the journey? High-intent App Store traffic may convert best through native checkout. Cold web traffic may need a quiz, personalized result, paid trial, or different offer before the install. Test the full path by audience and compare take-home revenue, not just fee percentage. 7. Build retention before buying more traffic AI apps generate about 41% more revenue per payer than non-AI apps but retain worse after one year. This is the perfect example of why conversion data can lie. A great ad can get someone to try a mediocre product. A great paywall can get them to subscribe. But only the product can make them renew. You need to be tracking the behaviors that happen before the first renewal: Did they hit the magic moment? Did they save data or create something valuable with your product? Did they turn on notifications that support the outcome? Did they invite somebody to try your product or share a result? Then build onboarding and lifecycle messaging around those behaviors. 8. Shorten the feedback loop This is the layer that turns an app funnel into a real growth system. Every customer should stay attached to their acquisition source as they move through: Ad view → click → store impression → install → activation → paywall → purchase → renewal. If installs are weak, you may have a creative or store-page problem. If trial starts are weak, the customer may not experience enough value before the paywall. If trial conversion is strong but payback is slow, the offer or acquisition cost may be wrong. If first renewals are weak, stop changing the paywall and fix the product. Too many founders change the creative, onboarding, price, and trial length at the same time, then revenue moves and nobody knows why. Test the biggest structural decision first, keep every other variable as stable as possible, and wait long enough to measure the outcome you actually care about. The goal is not more data. The goal is a usable answer before you make the next decision. Now what? Pick one meaningful acquisition path into your app. It could be: Creator post > App Store search > install > onboarding > trial > first renewal. Write down the conversion rate at every handoff. Then ask: Where are we losing the most qualified customers? Fix that transition before you redesign everything around it. Consumer app growth is not just great ads, a clever paywall, or a cheaper payment processor. The best operators understand how every layer affects the next one, and they can find the broken layer before burning another $100,000 trying to scale it. P.S. If you enjoyed this article, I go deeper with more tactical insights and lessons on AI and scaling consumer apps in my newsletter. It's called Operator's Notebook, and you can sign up for free at the link in the comments. - Jake Join 1,700+ weekly readers of the Operator's Notebook below (it's free): https://operatorsnotebook.beehiiv.com/ @themustachedad appreciate it! @vanshdevx you know it @blog2video_ai tried to simplify it as much as possible for everybody @SergeiFonov yep, but why did you automate this reply with AI, cmon now @benou_irl create demand or die lol @cwayneshuck 👍

I ran the marketing and operations for Cal AI that took us from nothing to #1 app in Health & Fitness in 18 months. In that time we signed hundreds of influencer partnerships, tested every niche imaginable, and built a system that ran most of it without us. When scaling a consumer app to any level, distribution is the #1 bottleneck you’ll encounter. And I’m going to help you fix this once and for all in the next 5 minutes. By the end of this you'll know: The systems that let four people manage hundreds of influencers on autopilot How to identity scalable creators vs duds The 3-to-5 rule for testing niches without burning through your budget Why your campaign briefs are killing performance and creator relationships How the influencer marketing pricing curve screws you at both ends, and show you the gap in the middle where the money is What you're actually buying with each marketing channel (most operators get this wrong and overpay for years) It's everything I'd tell a founder who asked me, "how do I scale my app?" Let's get into it. 1. Know what you're actually buying I've written about the three distribution channels (UGC, paid ads, and influencers) and how to evaluate which one fits best for your business and niche. Here’s the short version: With UGC, you're buying content and swings at the algorithm. Cheap videos used as a volume play, hoping a few go viral. With paid ads, you're buying distribution. Predictable, scalable, and you get exactly what you pay for…nothing more. With influencers, you're buying trust. The audience relationship already exists, and you're essentially renting it for the duration of your creator deal. The majority of failed influencer programs die because the operator behind it had no rhyme or reason for the decisions they made. These are some of the most common mistakes I see being made with brands who try influencer marketing: Paying for follower count instead of views Going after the biggest name instead of influencers with the right audience that will convert Forcing a scripted "ad" instead of organic-feeling videos that slots seamlessly into the creators natural content Running the channel in the wrong niche (i.e. B2B or Saas products don’t feel like a natural promo for most influencers) Never building a system for onboarding, deliverables and creator analysis Influencer marketing worked for Cal AI because: The health and fitness niche already embraced the before-and-after/transformation content format and could easily go viral We could sign fitness influencers to reasonable creator deals because everyone and their mom is trying to make a name for themselves (creator availability outpaced demand) The app felt like an actual product these influencers would use in their day-to-day lives. We partnered with the right influencers who had real trust built with their audiences, which helped us scale faster than anyone else. And the main reason we could consistently find creators who converted for Cal AI is thanks to the 20-second Influencer Test. 2. The 20-second test that makes or breaks influencer deals Too many operators spend hours analyzing whether a creator will convert by obsessing over followers, engagement rates, and audience demographics. If you need hours to figure out if an influencer would work for your brand, I can tell you right now they won't. After evaluating over 10,000 creators for Cal AI, these are the are only three questions you need to answer: Question 1: What do their views look like? Looking at average views for all their videos is a good indicator of sponsored post performance. And I am more interested in their baseline view metrics than the random few videos that popped off. I've seen creators with millions of followers who get 10,000 views a video, and creators with 15,000 followers who get millions of views a video. You're paying for the video, so the only number that matters is what a video actually does once it hits the algorithm. Question 2: What does the comment section look like? The sign of a healthy, engaged audience that will listen and buy anything that an influencer puts in front of them can be seen by simply looking through their comment section. If people are posting actual comments, asking questions, arguing with other followers, basically anything other than a bunch of emojis, then you're on the right track. Question 3: Could you imagine being friends with this person? How well does this influencer talk to the camera? Do they have a personality? Does the audience feel like they know them? Riche Lovelace aka daddywellness on IG is a great example of a creator who crushes these criteria. If you can't picture someone feeling like this creator is their friend, the parasocial relationship isn't there. And the parasocial relationship is the entire product. If the answer to any of these questions is no, move onto the next creator. If you're still deliberating after 20 seconds, that's your answer too. 3. You get screwed at both ends of the deal pricing curve Here's something nobody tells you about influencer pricing: the market is broken at both ends of the spectrum, but for opposite reasons. The top of the market is overpriced because big brands can't attribute. When a big brand signs a massive creator, they have no idea how much revenue that deal drives. The brand is doing so much volume that no single video moves the needle that much. So they take a shot in the dark, anchor a random price against what they'd pay Meta for the same reach, and usually overpay. Those inflated deals set the asking price for every other big creator in the market. The bottom of the market is overpriced because nobody knows what they're doing. When it comes to small creators, they usually only get DMs from inexperienced brands. And most small brands have never run a successful influencer deal. So when a creator with 5,000 followers quotes $2,000 for a post that'll likely get 3-4k views, the brand pays it because they don't know any better. Multiply that across the whole bottom of the market and tiny creators end up with prices that make zero economic sense. The money is in the middle. Mid-sized creators are what kept our dashboards up and to the right at Cal AI. This category of influencers are too expensive for the brands that don't know what they're doing, and too small to be worth a giant brand's time. That's where the entire Cal AI playbook lived. One more pricing rule, and it's non-negotiable: pay flat rates, never per view. A flat fee means the video can go mega-viral and you don't pay a cent more. That asymmetric upside is the whole reason organic content beats paid ads and traditional spray-and-pray UGC when done the right way. The second you pay per view, you've converted your influencer program into expensive Meta ads. 4. You're hiring the audience, not the influencer This is the lesson that cost us the most money to learn, so pay attention. We once paid a mukbang creator a serious check. Their videos hit huge views and always showcased tons of food, so on paper it made perfect sense for a calorie-tracking app. Spoiler: it barely converted. Because people watching someone eat 8,000 calories for entertainment do not want to count their calories. We made the same mistake with UFC fighters. Fighters track their weight obsessively, they cut for every fight, and the product fit seemed obvious. But we only realized after losing thousands in deals that the audience watching UFC fights are drinking beer and eating pizza on the couch. Comedy creators, same story. Yes they get massive viral reach, but when those videos go viral, it's a different audience every time. They might have a strong following, but they don’t have trust built up to start promoting a health and fitness app that clearly feels like a paid promotion. On the other hand, one of our best performers ever was a TikTok dancer girl. She never talked to the camera, had no previous fitness content, and every instinct I had said it would flop. What I didn’t factor in was that her audience had followed her for years without ever getting to know her personally, and the Cal AI video was the first time she let them into her actual life. They finally got to meet her through her Cal AI videos, and we reaped the rewards of a teenage Tiktok cult following. So how do you test a niche without lighting money on fire? The simple rule we followed was the 3-5 test method. Test a niche three to five times with different creators and different angles. At any sign of life, keep pushing. But if you get crickets across every attempt, kill the niche and move on. 5. Let creators cook The creator brief is where most influencer programs quietly sabotage the entire marketing channel. Operators send creators a 10-page brand deck with mandatory talking points or strict word-for-word scripts. Then put the influencer through three rounds of revisions and multiple approval checkpoints. The brands are then confused and angry when the content performs like an ad (hint: because they turned it into one). Think about what you're actually paying for. This person built an audience of hundreds of thousands of people by being good at making content. The moment you over-engineer their video, you've removed the only thing that made the deal worth doing. Our briefs were deliberately minimal. Here's the product Here's what it does Here's the one thing you can't say for legal reasons, aaaaand go. I worked in management consulting for years, so this was painful for me at first. I thought everything needed to be "client ready" aka detailed decks, perfect formatting, and every detail spelled out. It took me embarrassingly long to accept that creators don't want your boring PowerPoint. They want a Google Doc they can read in two seconds, and then they want you to get out of the way. And from a logistical standpoint, a lean team of 4 co-founders couldn’t physically micromanage hundreds of creators. Heavy briefs don't just kill performance, they also kill your ability to grow past 10-20 partnerships. 6. The Systems Everything I’ve already covered works because of creative testing, taste, and intuition. But none of it can scale without proper systems. When I joined Cal AI, we had two or three influencers and weren’t optimized yet for aggressive scaling. Signing a creator meant someone manually sending a contract, someone remembering to follow up, and checking view counts and comment sections multiple times a day. I rebuilt our whole system as an automated pipeline, and by the end, we were signing 10 new influencers a week and managing hundreds of active partnerships with a tiny team. These are the systems that winning apps use to scale properly: Step 1: Volume outreach. Virtual assistants running multiple outbound channels like clockwork. Step 2: Instant calls. Jump on a meeting the second a creator responds. By the end of that call it was a yes or a no, never "let me run it by the team." Step 3: One-click contracts. A yes triggered a set of automations: contract sent with the deal terms, post count, payment, and expected deliverables. No lawyer ping-pong or week of back and forth. Step 4: Auto-updating tracker. The signed contract updates on a master sheet automatically…so you can easily see every active creator, every scheduled post date, every deliverable, all in one place nobody has to maintain by hand. Step 5: Automated reminders. On post day, the creator got an automatic text: you're scheduled today, submit your video for review. Step 6: Track and renew. Every post got logged against views and comment quality, and that data decides who you renewed and who you cut. Your roster gets stronger every month on its own. Notice what's missing from that list: meetings, status reports, and project managers. Each step existed for one reason, to remove a human bottleneck between "creator says yes" and "their video is live." Build the systems once and every deal after that costs you almost nothing. That's the difference between running scalable influencer marketing and being run into the ground by it. 7. Speed is the only moat left For a long time, the smartest people in consumer apps said taste was the moat. They claimed AI can build anything now, but it can't make it look good, so UI/UX design was the last defensible edge. I believed that too, but I don't anymore. AI design has caught up. Every copycat app looks sleek now with just a few simple prompts. There are Cal AI clones trying to run our exact playbook as you’re reading this. Same features, same influencer strategy, some of them studied us closely enough to copy almost everything. But none of them have come close to our level of success, because the playbook was never the moat either. The only moat left is speed. How fast you implement How fast you test How fast you kill what's not working and double down on what is. And distribution is just speed applied to mindshare. The goal of everything in this playbook is to create one specific feeling in the market: "why do I keep hearing about this app everywhere?" That feeling is manufactured. It comes from concentration, hitting a niche from every direction at once until you're unavoidable. Now What? Nothing in here requires a Stanford degree or a million-dollar budget. Cal AI was built by four guys with obsession, some automations, and a willingness to test faster than everyone else. The apps that win the next few years are the ones that win on speed and distribution Now go make some money. P.S. If you enjoyed this article, I go deeper with more tactical insights and lessons on AI, scaling consumer apps, and business in my newsletter. It's called Operator's Notebook, and you can join for free at the link in the comments (it's free). -Jake @bucky_MT glad you found it helpful! @themustachedad 👍 @tim_gremp you got it Tim @lyuda_mekhed let's go, glad it's been helpful! @AndyTheGiant @swolematesgym nice andy @BurkhartLatam you got it Join my newsletter here: https://operatorsnotebook.beehiiv.com/ @alejandro_ff1 thanks man @TradfiBaby no problem @doclkk 2018-2024 If you like this tweet, you will love my newsletter. Join here: https://operatorsnotebook.beehiiv.com/ @ceokhalife thanks man! @BurkhartLatam 👍 @WPConvert_ai read this for rates https://x.com/jakecastilloooo/status/2069030138357112867?s=20 @MrjBond1968 glad you liked it @degenrsc thanks for the kind words @Blum_OG that was the hope @hussain__sh98 thanks Hussain @sun_choi8 hahaha appreciate it @jakkvega appreciate it Jakk @ipivto thanks IIya @Mr_Salio appreciate it Salio, glad it was helpful @everett770 you got it man @TheCyberPost1 yep always got to test everything @ArsiHoxha_ yep yep @VibeMngmnt for sure @deanfluence appreciate it dean

Two years ago while building Cal AI, we did the opposite of what every app growth blog and guru on the internet said to do: Reduce “friction” by shortening your onboarding by cutting steps to get users into the app as fast as possible. We made our onboarding flow LONGER, and went on to make millions a month. Over the life of Cal AI, we ran over 100 paywall A/B tests each with countless variants. Many of these experiments were on the onboarding paywall alone. I’m always getting DMs and comments about our influencer program, but almost nobody asks about the machine that actually collected the money: Paywall and onboarding optimization. The influencers got people to the app store, and the onboarding and paywall took care of the rest Here's the full playbook: 1. Get the order of operations right Most founders build in this order: Fancy UI/UX Product features Free trials galore Onboarding Paywall That's entirely backwards. At Cal AI we built: Paywall first Onboarding second Product features In-app experience Focus first on the paywall, because that's how you make your money. Focus on the onboarding second, because that's how you get people to the paywall. We didn't seriously optimize the in-app experience until we were already crushing it. Because who cares if you have 100% retention and 100% customer satisfaction if you have no customers? 2. Your paywall is a lie detector There are two versions of product-market fit. Version one: people are willing to download your app. Version two: people are willing to PAY for your app. Only one of these is a business. I talk to founders every week who are "growing fast" with a free app, acquiring users for $20 and only making $5 off each of them. I’m not good at math, but that's a money pit, not a real business. The industry data backs this up hard. RevenueCat's latest report across tens of thousands of subscription apps found hard-paywall apps convert trials to paid at 10.7% versus 2.1% for freemium (that's a 5x difference). By day 60, hard paywall apps generate about 8x more revenue per install ($3.09 vs $0.38). And at 12 months, retention between hard paywall and freemium apps is basically identical (27% vs 28%). Charging up front doesn't just make more money. It tells you the truth about your product faster than any customer survey ever will. 3. The sacred rule we broke (and why it worked) Back to the long onboarding. The reason "reduce friction" is bad advice for the onboarding-to-paywall flow is because friction isn’t always a bad thing. Friction is bad when someone is trying to USE your product. Friction is an asset when someone is deciding whether to BUY your product. Every additional step we added to Cal AI's onboarding was another opportunity to do three things: Learn something about the user we could use to personalize their plan Get the user to invest effort, which made leaving feel like losing something Remind them of their painful issue and connect their problem to our product as the solution they needed By the time a Cal AI user hit the paywall, they'd told us their goals, their body type, their habits, and their ideal timeline for change. They'd watched the app build a plan for THEM. That's a big reason why it converted. 4. Match your onboarding length to pain, not to "best practices" This is the framework nobody gave me, so I'm giving it to you. The right onboarding length isn't a design decision. It's a function of how much pain your app solves. Low-pain apps (productivity tools, utilities, nice-to-haves): Run a SHORTER onboarding. There's no deep wound to press on. Your user is mildly curious, not desperate. Every extra screen taxes someone who was barely motivated to begin with. Your only jobs: get them through fast, show them the value the app brings to THEM immediately, and give them one clear reason not to leave. High-pain apps (weight, money, addiction, dating - anything they've failed at before): Run a LONGER onboarding. Your user has tried to solve this before and lost. In this situation, the onboarding does the job of agitation and connection: Remind them why they're here. Make the pain specific again. Ask specific questions that prove you understand their exact situation. Show them their numbers, their timeline, and their projected result. Connect your mechanism to their problem so directly that by the paywall, your app is the obvious solution. Cal AI is a high-pain app. Traditional calorie tracking is tedious, frustrating, and most people have quit it at least once. Our onboarding made that pain vivid, then presented scanning a photo as the escape from it. The emotional target we aimed everything at: you can get the body you've always wanted, and it's as easy as taking a picture. Pain agitated, solution connected, paywall friction removed. Trying to copy Cal AI's 30-screen onboarding for your low-pain app is the easiest way to end up in the app store graveyard. 5. The anatomy of an onboarding that converts Steal this sequence. It's the skeleton under basically every top-grossing consumer app right now: The Promise Screen: One dream outcome, stated in the user's language (not features). The Demo: Show the magic moment in the first 10 seconds. For us: point camera at food, get calories. If they don't instantly get what your app does, nothing after this matters. The Questions: Goals, current state, habits, and constraints. Every answer should either personalize the plan or deepen their investment. No filler questions. Progress You Can Feel: Step indicators, small confirmations. Onboarding should feel finite and like it's building toward something. The Mid-Flow Review Ask: Counterintuitive, but asking for a rating BEFORE the paywall, right after a moment of delight, loads your app store page with social proof. (Time this carefully, a rating prompt at the wrong moment kills momentum.) The Personalized Plan Reveal: This is the "aha" moment. Their numbers, their timeline, their plan. The app already did work for them. Now leaving means abandoning THEIR plan, not skipping YOUR product. And finally, the Paywall: Their goal restated in the offer with 2-3 pricing options max, anchored with the annual option. Make the value of paying obvious and the cost of leaving concrete. One rule to follow across the board: every screen either teaches you something about the user, or teaches the user something about their problem. Anything else gets cut. 6. Free trials are a poor man's conversion tool If you can't convince someone to pay for your product right away, a free trial won't save you. You have one of these problems: No "wow WTF" moment within 10 seconds of someone learning about your app No unique mechanism (it's just like the others) The wrong audience entirely 84% of 3-day trial cancellations happen by the end of day one, meaning people decide almost immediately whether or not they’re going to continue using your product. The trial isn't buying you consideration time, it's mostly giving you false signals about whether this app is viable in the first place. And if you're running paid ads, a 7-day trial makes performance increasingly hard to attribute. You pay Meta today and find out in a week whether the cohort was any good. The longer the trial, the more polluted the data. My rule: never run a free trial until you've proven people will pay for your app out of the gate. Earn the right to use trials as an optimization, don't use them as a crutch for a product that can't close consumers from the product itself. 7. The benchmarks to memorize These are the numbers I used to diagnose Cal AI's funnel every single week: View-to-download ratio: 5 downloads per 1,000 views. Below that, your marketing isn't communicating enough value, or people just don't want the product. Paywall rate: 75%+. Three out of four people who open the app should reach your paywall. If they don't, your onboarding flow needs tweaking. Paywall conversion: 10%. One in ten people who see the paywall should pay. Hit all three consistently and you're on the right track. Miss one, and you know exactly which part of the machine to fix: Bad view-to-download = content/positioning problem. Bad paywall rate = onboarding problem. Bad paywall conversion = offer problem. Most founders can't tell you which of the three is broken. Not being able to identify leaks in your funnel and how to fix them is one of the biggest killers of consumer apps. 8. Price for mindshare, not margin Everyone thought we were insane for undercharging with Cal AI. But cheap enough becomes an impulse buy, an impulse buy becomes a habit, and habit becomes the default and obvious choice. You can fix margins later. You cannot fix being the app nobody chose. Two more pricing rules to steal: Skip lifetime deals. You can run into legal headaches, they complicate when it comes time to being acquired, and they rarely generate meaningfully more revenue than a well-priced annual. Mind the cashflow gap. It's hard to get CAC on Meta under $30, and for a lot of apps $30 IS the annual price. "We're breaking even" usually means you're lying to yourself: you pay Meta today and collect that subscriber's value over months. Just keep that in mind, cash flow is always king. 9. Test more (no, more than that) The majority of my consulting calls involve me telling founders to test 10x more of everything. My co-founder Zach said it best: "We treat paywalls the same way we treat the product: ship something, measure it, and improve it tomorrow. If you're spending weeks debating what your paywall should look like, you've already lost. Just test it." Your first paywall will be wrong, just accept that. The founders who win are willing to test and fail fast and fix their mistakes faster. 10. Stop paying rent on your own revenue Apple takes their cut of every app store transaction, and they keep tightening the rules: free trial toggles on paywalls, discount offers after the onboarding paywall, and rating requests during onboarding have all been hit. When you don't own the distribution channel, you're always at the mercy of whoever does. That's why the next wave of big consumer apps runs web2app funnels: acquire customers on a website you control, convert through any payment processor, keep roughly 95% of revenue, and test dozens of funnel variations in a day. Glam AI reportedly rode web2app funnels to $70M ARR. Flo tests trial-upgrade offers on the web that Apple would never allow in-app. You don't have to move everything to web funnels tomorrow. But if 100% of your revenue flows through a channel where someone else writes the rules, you don't fully own your business. The whole playbook, TL;DR edition: Build the paywall first, because that's the business. Build the onboarding second, because that's the bridge. Charge up front, because the paywall is your only honest PMF test. Make onboarding long when pain is high and short when it's low. Personalize until leaving feels like quitting their own plan. Memorize the three benchmarks and fix the one that's broken. Price for mindshare. Skip lifetime deals. Respect the cashflow gap. And test five things a month until the day you sell the company. There's no magic bullet in here. Just a machine most founders never bother to build. If you enjoyed this breakdown, you'll love my newsletter. Join 1,500+ weekly readers of the Operator's Notebook below (it's free): https://operatorsnotebook.beehiiv.com/ Link to the RevenueCat report: https://www.revenuecat.com/pdf/state-of-subscription-apps-2026-sosa.pdf?_gl=1*1tu6wgt*_up*MQ..*_ga*NTg3OTM1NzEyLjE3ODQ5MDM1MDA.*_ga_0MLNVKXFGB*czE3ODQ5MDM0OTckbzEkZzAkdDE3ODQ5MDM0OTckajYwJGwwJGgxMjc1ODU4MTM2 @John0t1jJohn thanks john @mrhanzhang correct @DiptamoyBarman FunnelFox is great

I was the operator behind Cal AI and scaled it to $30M ARR and #1 in the health and fitness category. Here's everything I learned building a consumer app people actually pay for: 1. A viral app does one thing We built Cal AI to be the best calorie tracker in the world. No gut health checks or workout plans. Be the master of one, not the jack of all trades. 2. A viral app has a “WTF” moment Build one core feature that feels like magic in the user's hand. Make the product so good people can’t help but share it organically. 3. Building the app is not the impressive part The barrier to building an app is now basically zero. The marketing and the strategy should get 95% of your time. If you won’t spend time learning how to market your product, don’t waste your time building it. 4. A viral app removes the most friction The business that wins is the one that reduces friction the most. Traditional calorie tracking is tedious and annoying, we made it as simple as snapping a photo. Find the painful step and delete it. 5. A viral app sells the dream, not the features Nobody wants a calorie tracker. They want the body they've always wanted. People buy more money, more time, better health, and less pain. The feature is just the vehicle. 6. A viral app is understood in 5 seconds If someone has to think about what it does, you've lost them. If your app needs a 60 second intro, cut until it doesn't. 7. A viral app is built on three metrics View-to-install ratio. Paywall rate. Paywall conversion rate. Master those three and you won't have a user acquisition problem. Everything else is a distraction until these are dialed. 8. Know your targets on those three 5 installs per thousand views = you're doing fine. 75% of installs should reach the paywall. 10%+ of them should pay. If you're hitting those, you’re on the right track. 9. Your metrics tell you product vs. distribution issues Bad view-to-install ratio? You're not communicating the value well enough, or nobody wants it. Good view-to-install but weak paywall conversion? The idea is fine, the product or the paywall isn't. 10. A viral app builds the paywall first Optimize the paywall, then onboarding, then everything else. The paywall makes the money, so build the thing that takes payment before stressing about the little details. 11. A viral app uses onboarding to sell Onboarding has one job: get them to the paywall ready to convert. Everyone preaches short onboarding flows to lower drop-off. But a longer flow hammers the pain point harder, and the lift in conversion beats the drop in paywall rate. 12. A viral app proves people will pay, not download A free app with a million downloads can have zero product-market fit. Being willing to download and willing to pay are different signals. Only one makes you profit every month. 13. A viral app acquires users profitably Of course you can find people to download, anyone can buy downloads. But if it costs $20 to get a customer who pays $5, you don't have a business. You have a money incinerator. 14. A viral app ignores in-app polish until it has customers We didn't touch the in-app experience until we were doing millions. Who cares about 100% retention if you have no customers? Retention is a problem you earn later. 15. A viral app markets while it builds Build the waitlist from day one and test angles, niches, and creators in parallel. Then you’ll launch with proven marketing strategies and a list that already wants what you’re selling. 16. A viral app is born in an expanding market Build in wellness, productivity, micro-finance, education, AI, women’s health and creative utility. All are growing markets, have problem and solution aware buyers, and a flood of creators making content. Build where the tide is already coming in. 17. A viral app validates the niche on social, not the App Store Search your niche on Instagram or TikTok. Seeing diverse creators on your feed? Green light. Seeing the same creator five times? Competition will be fierce. 18. A viral app knows speed is the only moat left Taste used to be the moat. The only moat left is speed. How fast can you build it and ship it? 19. A viral app uses first-principles thinking Don't solve a problem the way it's always been solved. Find the fastest, most logical way to do it right now. 20. A viral app distributes like a blitzkrieg A slow build is a death sentence. You want to be everywhere at once as fast as possble. The goal is getting people to say, "why do I keep hearing about this?" 21. A viral app does the unsexy work Getting early traction means hundreds of manual DMs and cold emails a day. Multiple accounts, multiple phones, boring, exhausting, and it works. If you're hunting for a cheat code, that's proof you won't do the work. 22. A viral app respects efficient markets There are two paywall providers everyone uses because they work. There are a thousand influencer-sourcing tools because none of them work. If everyone in a category avoids the obvious hack, there's usually a reason. 23. A viral app keeps Apple Search Ads simple Use keywords for your app name and your search terms. Done. Agencies make it sound complicated so you hire them. Don’t fall for it. 24. A viral app moves before it's ready The founder who breaks out stops asking whether it'll work and builds. Not a 30-day plan you debate. A 15-minute conversation, then move and iterate live. Fail fast and scale faster. 25. A viral app forces its own speed Set deadlines that are intentionally unreasonable. Fall short of an aggressive one and you're still fast. Fall short of a comfortable one and you're slow. The pressure is the point. 26. A viral app runs on a flat org No hierarchy in the decision-making. Anyone can make a call that moves the business, no chain of approvals. Everyone's a contributor and every idea is heard and tested. 27. A viral app is generous with equity 10% of a watermelon beats 100% of a grape. Align the incentives, hand out real ownership, and people build like it's theirs. Because it is. @themustachedad appreciate it @ZoharMosseri you got it Zohar @jsarthak110 no problem man! @kurorosage you got it Jose @ContrabandHQ 😉 @NpBelief glad you liked it @girlmoneyguide 100% @griffenshu thanks G If you liked this breakdown, you'll love my newsletter. Join 1,200+ weekly readers here (it's free): https://operatorsnotebook.beehiiv.com/ @J7K_dev love it @taylorfrankel_ Thank you!