Founder Playbook · Sub Club by RevenueCat
9 tactics from Eric Seufert
The Post-Attribution Playbook for Growth
Watch the full episode“There is no point for you to try to interpret why an ad won or didn't win. What you should be interpreting is: when you get a win, or the win rate increases — the process worked. The output is irrelevant. That output is utterly random. Why that worked was utterly random.”
Stop trying to decode why a winning ad won — focus on the process that produced it
Seufert's most counterintuitive insight: with black-box platforms like Advantage+ and Pmax, the winning creative is selected by an algorithm you can't see, targeting an audience you don't know. Post-hoc rationalizations are noise. The only learnable signal is whether your creative-input process improved. Focus on raising the win rate, not reverse-engineering individual winners.
“Diversifying for the sake of diversifying is often times a bad idea. A new channel's ROAS doesn't just have to meet the ROAS of the other channels that could have absorbed that budget — it has to exceed it, because you're supporting a new channel.”
Diversifying UA channels for its own sake destroys performance — use the waterfall method instead
Seufert's waterfall model: max out the biggest channel until it hits your ROAS threshold, then move to channel two, then three. Adding a second channel before saturating the first adds overhead without proportional return. Diversify only when you've hit true saturation or when a brand-oriented channel lifts blended performance across the portfolio.
“The thing I see most commonly is just a very chaotic approach to measurement. Your measurement model is essentially the heartbeat of the company. Everything flows from that. You need to be doing it correctly, in a way that's credible, and in a way that everyone understands.”
Measurement disorganization — not technology — is the #1 growth blocker
Seufert names measurement disorganization as the single most common failure he sees: competing tools no one knows how to reconcile, finance and UA with different definitions of success, and LTV models the product team rejected and rebuilt. The fix is a people problem — getting finance, UA, and product in a room and aligning on one operational model before touching any new technology.
“You can create hurdles for the user to clear that potentially are good proxies for LTV. The most high-intent users will do it — and that could be a very strong signal of ultimate value. One host posted on Twitter after they created a CAPTCHA for no reason other than to test user intent — and it drove a 40% increase in ROAS.”
Signal engineering: create intent hurdles to send higher-quality signals to ad platforms
Signal engineering inverts conventional UA thinking: instead of lowering friction, deliberately add friction at the ad-to-install handoff to filter for intent. The users who clear the hurdle self-select as high-LTV; that signal gets fed back to the ad platform to optimize acquisition. Seufert envisions Meta/Google eventually doing this automatically, personalizing landing pages per user in real time.
“Creative generation for ads is probably the least valuable place to apply AI. What you really care about is the concept. What I see people using AI well for is creative prospecting — pulling competitors' ads and using an LLM agent to synthesize what's working. That would have been a full-time job three years ago.”
AI's real marketing value is creative prospecting at scale, not variant generation
Generating 200 variants of the same 10 concepts adds near-zero value — the 190th variant is not meaningfully different from the first. The leverage is in concepting: finding angles that wouldn't emerge from a human brainstorm. AI-powered competitive synthesis (scraping ad libraries, agent interpretation, hypothesis generation) automates what used to require a full-time analyst.
“Why would performance bomb on a channel? Probably because a competitor came in and is outbidding you everywhere. If they're outbidding you on Facebook they're probably outbidding you on Google and Snap and TikTok. The risk is not per channel — it's structural.”
Structural channel declines are correlated — diversification rarely protects against the real risk
The popular rationale for diversifying across paid UA channels is risk mitigation: if one channel tanks, the others hold. Seufert points out this is mostly wrong. Any structural performance drop — competitor entry, macro sentiment shift, category saturation — will affect all channels simultaneously. The real risk is a business-level headwind that no channel mix fixes.
“The real answer to 'what's the biggest opportunity for growth?' is just that your measurement doesn't support true growth. It's broken or flimsy. You're just trying to replicate what you've done in the past. If you don't believe your measurement can adapt to new channels, you just won't grow.”
Measurement paralysis: broken attribution freezes teams into replicating the past
Teams that don't trust their measurement framework default to running the exact campaigns they've run before — they can't evaluate anything new. Building truly robust, incrementality-focused measurement is the unlock for every other growth lever: influencer, OOH, CTV, podcasts. Without it, the growth ceiling is baked in.
“I'll spend $5K a day, hitting 150 ROAS. Let those cohorts age. Understand the day 20 ROAS, the day 30 ROAS. Build more cohorts. Then push that frontier out. I iteratively progress that frontier, starting from a place where if I can't hit it at very low spend, it's back to the drawing board.”
Iteratively push the ROAS frontier instead of chasing terminal LTV on day 10
Early campaign optimization should not attempt to project terminal LTV from 10 days of data. Seufert's process: establish a baseline ROAS floor at minimal spend, let cohorts mature, then incrementally raise budget only as cohorts prove out downstream value. This prevents both premature scaling (before economics are validated) and over-cautious under-investment.
“The app industry — and the proliferation of subscription apps specifically — Meta and Google deserve almost as much credit as Apple for the variety of apps today. They empowered app developers to reach those audiences more effectively. What we see as the app industry owes a debt of gratitude to Meta and to Google.”
Personalized advertising created the subscription app economy — the data is the product
Without granular audience targeting, niche subscription apps can't reach their first 10K users profitably. TV and radio are economically impossible for a $10/month wellness app. Personalized advertising is the infrastructure that made the subscription app economy viable — protecting it matters for anyone building in the space.