Your App Is A Funnel
A new mobile app is a funnel optimization exercise. Not a product problem. Or more precisely, the product problem doesn't earn your attention until the funnel proves someone wants to walk through it.
Founders have a unique ability to see the future state of a product while ignoring its present one. We picture version four, the one with the community features and the Watch app, and skip past the uncomfortable fact that version one has almost no users. Most of the leads that come to Flywheel think this way. They want to talk about what the product will become. They don't have a product or users right now.
I do it too. I'll come back to that in a minute.
Disclosures
Flywheel's new engagements start with a paywall optimization sprint or a revenue leakage audit. App monetization work, in other words, funnel work. This essay argues that funnel work is the highest and best use of an early stage founder's investment,
Second, Primus is my Navy SEAL fitness app. It has a six month training plan built to prepare users for BUD/S. Throughout this essay, I’ll be using my own app as an example.
What’s The Right Question?
When developers talk about a new app, the conversation defaults to the core experience. For Primus that sounds like: how do we make the workouts more compelling? Should we add GPS and maps for live tracking? What about a Watch integration? Should we add AI, since everyone else is?
With zero users, those are the wrong questions. Not because the product doesn't matter. You're selling a product and it should be good. But every one of those is an enhancement, and we're still in the "is this a good idea?" phase. The answer to that question is top of the funnel metrics - advertising, store impressions, downloads. You don't have users to ask, so you have to read behavior instead.
The Feedback Objection
A client pushed back on this recently, and it's the strongest counterargument: get a handful of users first, then build from their feedback. Real usage beats guessing.
He's right, under one condition. If I had a controlled group of 50 users who wanted to train for BUD/S, my assumed ICP, I could watch their analytics, collect their feedback, and improve the in app experience with confidence. That's a focus group. But that's not what early downloads are. Early downloads are ten anonymous people from the App Store, and you have no idea if they're your people. The feedback isn't wrong, but it isn’t actionable. You can't weight a signal when you don't know who's sending it.
Primus taught me a second version of this problem. I assumed the ICP was people preparing for BUD/S. I now think it's broader and messier - people who want to train like a Navy SEAL, most of whom will never ship out to Coronado. I couldn't have learned that from ten users. That distinction only shows up at volume, when you can see who the ads attract versus who actually converts.
Which brings me to where I've landed: when your user group is small, the highest and best use of your time is increasing the size of the group, not optimizing for it.
Where the Funnel Ends
The funnel doesn't stop at the download. It runs all the way to first conversion - install, first open, onboarding, paywall, purchase. Your app onboarding is the middle of the funnel, not a separate project. Fixing drop off in the onboarding is still funnel work. What waits is everything after that first conversion: habit loops, feature depth, the Watch app, retention mechanics. Acquisition and activation now. Engagement later.
So there are two dashboards to watch, and only two. In the App Store you're reading four numbers: impressions, product page views, conversion rate, and first time downloads. Remember that browsing users only see your screenshots, title, and subtitle. They don't see the description until they've already tapped in. The description sells the curious. It doesn't attract anyone.
Inside the app, the exact events vary by product, but normally looks like this:
- First open
- Onboarding drop-off per screen
- Paywall view rate
- Trial start and paid conversion rate
That's it. Don't build reporting beyond this yet. The metrics matter less than the discipline of stopping there.
Back to the Future State
My working threshold is a few hundred users before real product decisions - roadmap bets, feature investments, the fun stuff. That number is an estimate, not law, and it varies for every app but it’s important that it isn’t a dozen users. We need more than that.
In Flywheel, we’re putting this into practice too. It’s similar to the Minimum Lovable Product (MLP) work we’ve done in the past, but now we’re using the funnel exercise to add a more scientific approach to development and launches. Long term retention features are backlogged until we have concrete evidence we have the right idea, the right funnel, and we’re attracting the right users. Then we invest in the future.
Next in this series: turning the funnel into a loop, and what a loop costs to build.
The future state of your app is real. I believe in yours and I believe in mine. But the future is downstream of a funnel that works today, and today, the funnel is where the work is.