Product Playbooks
Decisions that shape the product itself — what to build next, when to say no, and how founders used real feedback to steer the roadmap without losing focus.
277 tactics · page 3 of 10
“at RevenueCat we acquired an app called Dipsy we acquired this app specifically to run these kind of experiments to beta test our revenue cap features within days of the ruling my colleagues built out a system to take advantage of this”
Acquire an app specifically to run live monetization experiments on your platform
RevenueCat bought Dipsy not for its users or revenue, but as a real production test bed — a live app fully instrumented with their own tools. This let them ship experiments within days of a legal ruling without convincing external partners. For founders building monetization tooling, owning a first-party app eliminates partner dependency and produces credible data to share publicly.
“you really want to be smart about what each individual user is seeing and making sure that it's not too much... if you're running like you know four or five different like packages in terms of like weekly monthly quarterly annual um and then you're also doing like three different monetization methods it can become a lot”
Serve one monetization model per user segment, not all at once
As hybrid monetization grows, the risk is overwhelming each user by showing every model simultaneously — weekly, monthly, quarterly, annual plans plus ads plus usage credits. The fix is to run many variants but ensure each user sees only one clear, well-matched offer at a time. One well-chosen offer for the right segment converts better than three mediocre options fighting for attention.
“the company that did that the best nothing to do with subscription but Credit Karma... You don't even know it's an ad right you're just using their website... you are actually getting matched with a product and service that is good for you”
Credit Karma ads model: helpful matching beats banner spam
Life360's aspirational ad model is Credit Karma: match driving-score data to insurance offers, or surface an Uber tap-to-book when the app detects a user just landed at an airport. Their internal bill of rights forbids any ad that interrupts the job-to-be-done. Ads that add genuine contextual value convert better than banner spam and preserve trust with the 87.5% of Life360's user base who are permanently on free.
“It's almost like any like portfolio of companies or products like there's a power law to subscription features where one or two features are probably going to drive the majority of your new subscriptions and everything else is nice to have but it's incremental.”
Subscription Features Follow A Power Law — Two Features Will Drive The Majority Of New Subs
After years of building Tinder's revenue roadmap, Jeff observed that regardless of how many features you put behind a paywall, one or two will dominate subscription conversion. Everything else is incremental. His dashboard showed click-through and conversion per feature daily — knowing which features drove subs was the north star for what to build next.
“There's a like a really clear zeitgeist right now the anti-slop zeitgeist and people wanting to use things whether it's in the physical world or digital world that are high quality and thoughtful and doubling down on product and design.”
The Anti-Slop Zeitgeist: Design Quality Will Rise As AI Lowers The Floor
Jeff's contrarian take on AI-driven app creation: the bar for design won't fall, it will rise. As AI-generated slop floods the market, users will gravitate toward products with intentional design and distinctive point of view. This creates an opportunity for founders who double down on craft — the same dynamic that made Flighty worth $40/year despite free alternatives.
“We put a paywall on the number of swipes you could do on a daily basis because we found there were people who were just swiping their way through literally like their entire city in you know like a very small amount of sessions... we put up that paywall to also create a better ecosystem but it also turned out to be a great thing for monetization.”
Tinder's Swipe Limit Paywall Fixed Ecosystem Health And Made Money At The Same Time
One of Tinder's most successful paywall decisions was limiting daily swipes — motivated not primarily by revenue but by ecosystem health. Power users were swiping through entire cities, degrading match quality for everyone. The paywall solved a product problem and became a major monetization driver. Jeff's lesson: the best subscription features often protect the product's core value, not just extract money.
“Most people do talk to their tolen via voice and we tried to make that like the best quality voices possible the fastest response times possible the best memory system... once you actually start trying to build a awesome voice AI experience there's a reason that there are like two or three on the app store.”
Voice AI With Memory Is The Core Technical Defensibility — Most Competitors Won't Replicate It
Tolen's real defensibility isn't the cute alien art — it's the engineering underneath: low-latency voice AI, parallel prompt streams for conversation classification, and a memory bank that resurfaces personal details across sessions. Ajay notes that while competitors can copy the visual design, replicating the actual voice + memory experience is what keeps the category sparse.
“we were super lucky at some point to hire someone dedicated to monetization mik... he has really like this data mind... all experiment and this is working very well”
One dedicated monetization person beats a committee
Mojo grew to $1M MRR with a team of 30, allocating just one person full-time to monetization, paywall, pricing, and onboarding experiments. That focus enabled quarterly iteration cycles: paywall position, then full redesign, then localization, then price testing. Spreading monetization across multiple part-time owners produces none of these results.
“you're going to have complexity choose your complexity and you want to choose that complexity based on leverage”
Choose your complexity — prioritize high-leverage hard problems
Mojo skipped CRM entirely at $1M MRR — not because email is unimportant, but because CRM across a multi-language user base is low-leverage complexity. Instead, resources went into the high-leverage complexity of building AI into video creation. The principle: every decision adds complexity; deliberately choose complexity where the potential leverage is highest, and be explicit about what is being deferred.
“hiring people who are into fishing makes a huge difference right like hiring people who are a target customer you shouldn't just hire your target customer when it comes to marketing and business it's actually pretty helpful”
Hire target-customer employees to close the empathy gap
Fishbrain is headquartered in Stockholm but its core market is American fishermen. Lisa deliberately hired US-based team members who actually fish — for business development, social media, and marketing. When a product team is culturally or geographically removed from its users, hiring inside the customer base gives permanent, cheap market research that no agency can replicate.
“we'll get like one of our amazing customers... a guy from rural mississippi who loves fishbrain and he will... call in and just take questions from the team for 20 minutes and that in itself is like a really great”
Invite a real customer to your all-hands to create company-wide empathy
Fishbrain periodically invites a real user to take live questions from the entire company for 20 minutes at their all-hands meeting. A product manager's customer interview notes reach a few people; a fisherman from rural Mississippi talking directly to engineers, designers, and finance changes everyone's mental model. One unfiltered customer voice compounds across every future decision.
“all of those 13 million registered users they don't all of a sudden just start buying just because you put a shop in their face which in retrospect makes sense but like it's been much more of a challenge”
Commerce layered on social did not convert — even when it made logical sense
Fishbrain added a marketplace three years before this episode with a logical thesis: users see what gear someone used to catch a fish, one click to buy it. In practice, 13 million registered users didn't convert to shoppers at any meaningful rate. Most commerce revenue came from the web, not the app. The mental mode of using a utility or social app doesn't automatically include buying — that purchase habit has to be built separately.
“we put up a registration wall... the backlash that we got of you're going to put up this registration wall when it was completely free before... we just took it all down... I think we earn goodwill by being willing to say we're going to walk back on what we ultimately did”
Walk back a bad monetization decision before resentment compounds
When Pitchfork added a registration wall the day of an unrelated editorial controversy, the backlash was immediate. Michael pulled it the same day rather than defending the decision. Walking back a monetization test quickly signals respect for the community and recovers goodwill faster than any apology campaign. Stubbornly defending a bad experiment costs far more than reversing it.
“a lot of people want to launch a premium product and I'm like no... prove to me there's going to be enough value and like there is another cohort here that isn't just like 1 or 2% that we're going to do a lot of work for maybe not a lot of return”
Prove the premium cohort exists before building a premium tier
Adoption rates for premium tiers run in the low single digits unless the product already has a clear superfan segment. Michael's test: before building a high-touch tier, identify whether the cohort is large enough to justify the operational lift, or whether the same LTV can be captured more cheaply through IRL events, merch, or post-purchase upsells. Most teams launch premium tiers too early and should optimize the base product first.
“if you're talking to a product team and they really want to advocate for one part of this should absolutely be free because we want the most people using it... I would love a show of hands of whose company's KPIs are completely aligned... it's zero”
Align KPIs across teams before running free-vs-paid experiments
The hardest part of optimizing a freemium paywall is not deciding what to gate but getting product, revenue, and marketing teams to agree on which metric to optimize. When product wants MAU and revenue wants conversions, every paywall test surfaces as a political conflict rather than a data question. Aligning the team on a single north-star KPI before launching experiments is the prerequisite most companies skip.
“inside of 11 Labs if you were to take a look at it we really look like and feel like maybe 10 or 12 individual startups that are all building shipping and iterating all at once... our mobile team is about 12 people right now... each team that we kind of set up as a pod or a speedboat is their own kind of self-contained development design growth product”
Speedboat teams: 12 people, full autonomy, monthly ship cadence
ElevenLabs organizes its 300-400 person company into 10-12 fully self-contained speedboat teams, each with its own design, engineering, growth, and product responsibilities. The mobile team ships major features monthly, far faster than the typical once-per-quarter cadence. This structure lets the company operate as many small startups simultaneously without the coordination overhead that kills speed at larger companies.
“we take a good hard look at public GitHub profiles to see who's engaged in what... we look for those that build side projects on their own that can identify a user pain point and like really build and be obsessed with that... it just becomes a referral flywheel”
Screen GitHub profiles before interviews to hire product-minded engineers
ElevenLabs screens engineering candidates' public GitHub profiles before the first interview, looking specifically for developers who independently build and ship side projects. This filters for product-minded engineers who own outcomes rather than just close tickets. These hires then attract similar people through referrals, compounding the culture without diluting the founding character of the team.
“it's so powerful to be the first best customer of your API... having those consumer products can be a superpower for the API... it's a better API when internally you can be testing against it and know what the experience is of building against that API”
Consumer apps are the best dogfood for your API product
ElevenLabs builds consumer apps in part because they make the API better. Internally using the product surfaces quality issues, latency problems, and UX friction that external API customers experience but rarely report clearly. The consumer product and API product become mutually reinforcing: shipping a better consumer experience reveals model improvements needed, which compounds the technical moat for API customers as well.
“we saw about a third of our leads every day were saying they wanted traditional bodybuilding and at that moment it was mostly women who were coming through our funnel and saying that and we didn't have a traditional bodybuilding program so the conversion rate for that user versus someone who didn't say they wanted bodybuilding was half”
A 15-question quiz routing 1M users became the product roadmap for which coaches to hire
A million people completed Ladder's 15-question web quiz in 18 months. When a third consistently chose 'traditional bodybuilding' but Ladder had no such program, conversion rate for that cohort was half the average — a clear signal. The quiz became a demand-forecasting tool: identify the missing modality from unmet quiz responses, model the revenue, then recruit the exact right coach to fill it. Ladder's last two launches came directly from this data.
“our sort of like a umbrella metric for all those monetization experiments was the average revenue per user in the first seven days So this kind of RPU 7D”
Use ARPU-7D as your monetization north star
Mojo chose 7-day ARPU as their single unifying experiment metric because it covers the full 3-day trial window and directly feeds the unit economics needed to scale paid acquisition. Optimizing for trial conversion alone can miss revenue quality signals; ARPU-7D captures plan mix, price, and conversion simultaneously. It also shortened the feedback loop so each experiment cycle concluded in 1-2 weeks per geo segment.
“we mostly run experiments We typically kind of broke down the audience to sort of three buckets according to GIO It was like a US or kind of US Australia Canada Then there was the Europeans and then there was Latin America and we typically had like this streams of tests for each”
Run three geo-segmented experiment streams simultaneously
Mojo ran three parallel experiment pipelines covering US/Canada/AU, Europe, and Latin America so each test reached statistical significance in 1-2 weeks per segment. A winner in Europe sometimes failed in LatAm and rolled out only where it worked, preventing revenue dilution in mismatched markets. Segmenting experiments by purchasing-power region rather than language avoids misleading blended results.
“the number one thing which allowed me to do that was having kind of a third party payable platform allowed me to being autonomous and and iterate fast and and really kind of shorten the cycle of actually like coming up with experiment developing it launching it from yeah months or weeks to to essentially days”
A paywall platform compresses experiment cycles from weeks to days
Using a third-party paywall platform let Michal run experiments without engineering dependency, cutting the cycle from months or weeks down to days. This autonomy was the enabling factor behind the 60% ARPU increase: without it, the same experiments would have taken far too long to compound in five months. Experiment velocity is a growth asset; if your paywall requires a sprint to change, your testing cadence is artificially capped.
“you want the what the customer is buying like the heart of like why do they want this thing to be part of what they're getting and they're paying for that $300 or they'll feel sort of tricked and duped right”
Keep the product's core promise free or customers feel tricked
Skylight draws the paywall line at the product's core promise: photo-sharing to grandma stays free, videos go behind the paywall. Calendar management stays free, meal planning is a subscription. When the feature a customer bought the device for is paywalled, the emotional response is betrayal rather than aspiration. Define the core promise of your product explicitly, and never put it behind a paywall regardless of conversion impact.
“today about 20% of people who reply say that it's life-changing and our hypothesis is those are the people who are getting us dozens and dozens of new customers just by virtue of they can't stop talking about it”
20% of customers calling you life-changing are your real growth engine
Skylight polls customers and finds roughly 20% describe the product as life-changing rather than just useful. Michael's hypothesis is that this cohort drives a disproportionate share of organic growth through word-of-mouth, each generating dozens of new customers. The product goal is therefore not marginal improvements across all users but engineering specific moments that move customers from very useful to life-changing, which is what turns paid growth into organic momentum.
“The important part here is going to be measurement overall itself so having a place where you're able to see that performance in an unbiased way so essentially we want to be able to compare meta with Google with TikTok we also want to be able to see your influencers or your programmatic buys Etc so having all that in one place is very important.”
Central Measurement Platform Is the Foundation — Compare All Channels in One Unbiased View
Each ad network reports its own attributed conversions with an inherent bias toward overcounting. Shane's top tactical advice: use a neutral third-party measurement platform (like AppsFlyer) as the single source of truth so Meta, Google, TikTok, and influencer spend are evaluated on identical terms. Without this, budget decisions are distorted by whichever network shouts loudest about its own performance.
“The biggest tip here is to make sure that we have that data pipeline set up so you're able to see this measurement end to end. Where did they come from did they actually subscribe and is there a particular channel that maybe we should defocus because a majority of the users are churning.”
Build the Full Data Pipeline: Install Source → Trial Start → Conversion → Churn by Channel
Shane's most critical piece of advice for subscription apps: don't stop measuring at the install or even the trial start. The full pipeline must track install source → trial start → paid conversion → renewal → cancellation, all linked to the originating campaign. Without it, a campaign that drives massive trial starts but near-zero conversions looks like a winner in network dashboards but bleeds money in reality. The churn signal by channel is what enables real budget optimization.
“For a year or so I just switched as a founder as we were building the app to Android and so I was like only using that to make sure it's important and you pay attention to details.”
Switch To Android As Your Daily Driver For A Year — The Only Way To Actually Care About Quality
Photoroom's most practical Android QA advice: Matt Rouif removed his SIM card from his iPhone and ran Android as his sole device for a year during the Android build. Dog-fooding is the only thing that creates genuine attention to details — you notice the rough edges you'd otherwise never see in a formal QA pass. This is especially important in companies where most engineers and PMs are iPhone users.
“What we do give is like anyone that wants an Android as a Photoroom employee who on iOS and wants Android going to use it or test it or try it or is a PM which is you can have two device if you're going to help the Android team.”
Give Any Employee A Free Android Device If They Will Test It — Offset The iOS-Bias In Tech Culture
Tech company culture is heavily iOS-biased — most engineers and PMs own iPhones and never naturally encounter Android issues. Photoroom's fix: offer a free Android device to any employee willing to use or test it. PMs and developers who support Android are allowed two devices. It's a cheap, practical way to increase internal Android usage and catch bugs that would otherwise only surface in user reviews.
“The share icon — I had so much debate with the design team about the designs and you see this little arrow compared to the three circles like the branching on Android that is the share icon. People on Android don't understand the iOS icon and it doesn't look native.”
Adapt Native UI Language Per Platform — Using The iOS Share Icon On Android Signals "Foreign"
Cross-platform apps frequently use iOS UI patterns on Android out of convenience, but native users notice immediately. The share icon debate at Photoroom — iOS arrow vs Android branching circles — sounds trivial but represents a deeper truth: each platform has its own visual language, and violating it makes users feel the product is a port rather than a native app. Allow the Android team freedom to adapt platform-specific idioms.
“We basically ask for ratings after we've delivered value to the user. That multiplied by the number of downloads... helps overall. We don't try to do too much gymnastics with trying to get more ratings.”
Ask for Ratings Only After You've Delivered Value — Organic Velocity Beats Artificial Prompts
Super Unlimited has 2 million+ high-quality ratings while explicitly avoiding rating-prompt tricks. Tanuj's rule: trigger the request only after the user has received real value from the session. At million-downloads-per-day scale, even a modest honest rating rate produces massive ASO velocity. Over-prompting or ill-timed requests kills both UX and rating quality.