The MVP of Everything

You can't optimize an app nobody uses. Every founder nods at that sentence, and then most of them spend their first budget trying to do it anyway.

Here's what that looks like in practice. A new client comes to us with a list. Gamification, streaks, a smarter paywall, push notification re-engagement, referral loops. The stuff that makes Duolingo Duolingo. At best, they want to invest in getting users and keeping users equally. In our experience, that's under-investing in getting them. Usually badly.

Quick disclosure before I make this argument. We're a studio. We get paid to build things, and a client who wants five systems instead of two is better for our invoice. So when I tell you to build less, I'm arguing against my own revenue. Weigh that however you like.

Walk through how anyone actually meets an app. Take Duolingo. First you discover it somewhere - an ad, a friend, a Reddit thread, a search for language learning apps. Then you hit the App Store page. A title, some screenshots, a description you'll skim at best. You think about your trip to Spain this summer and download it. You go through onboarding. You pick Spanish, create an account, accept push notifications, commit to a daily goal. Only then do you complete your first lesson. That's step five. You made a dozen decisions before you touched the actual product.

Once you're inside, though, Duolingo is a machine. Animations, sounds, streaks, leagues. It's like learning Spanish in a casino. Miss a day and the sad owl haunts your home screen until you come back. Stick around long enough and you'll probably pay, because paying is easier than the hoops. Language learning is a graveyard of apps, and that machine is why Duolingo survived it.

But notice where the machine lives. Everything it does happens after step five. It only works on people who already showed up. A brand new app has no one inside the casino. There's no churn to reduce, no streak to protect, no paywall traffic to A/B test. The engine is real. It's just not your problem yet.

Here's the math that convinced me. Say you have $5k to invest in your product. You could split it evenly - a thousand each into marketing, the App Store listing, onboarding, the paywall, and gamification. A thousand dollars buys a mediocre version of each. Or you could put all $5k into gamification and do it properly. Suppose you have 100 users and you cut monthly churn from 10% to 5%. That's a genuinely great result. It's also five users. At $10 a month, that's $50 of new revenue, which means 100 months to earn your $5k back. More than eight years, before any actual return. The same $5k spent on ads, the listing, and onboarding gets you users this quarter. The exact numbers matter less than the shape: optimization pays you a percentage of your users, and a percentage of almost nothing is almost nothing.

Clients push back here, and it's a fair push. "If I don't have gamification, why would anyone download the app?" But look at the journey again. At the moment someone decides to download, they haven't used the app. They can't want a feature they've never seen. If your category has a table stakes feature - a specific workout, converting an image to a PDF - then yes, you need it. For everything else, no one uses apps because of gamification. They use apps more because of gamification. You need a product worth returning to before the return mechanics matter.

Which brings me to what we actually do about this. We call it The MVP of Everything. We build every part of the funnel - onboarding, the core product, payments, push notifications, even some light gamification - but we build each part thin, flexible, and simple. Minimal because we don't know what users want until we have users. Flexible because we expect all of it to change. Simple because complexity is optimization, and optimization comes later. Payments are the clearest example. We usually recommend charging from day one, since nothing validates a product like a stranger paying for it. But clear and simple pricing. Don’t make pricing a difficult user decision on day 1. 

Two honest caveats. This isn't a license to ship a bad product - thin and bad are different things. And it won't fit every app the same way. Some categories live or die on one deep feature, and we'll weigh the budget differently there.

The counterargument I hear most is "we're going to build it eventually, so why not now?" Because budget isn't the only constraint. Focus is the other one. A small team can only go deep in a few places at once, and every dollar spent tuning a casino with no one in it is a dollar not spent getting someone through the door.

So build the MVP of everything, and only the MVP of everything. Get the users. Then you'll have problems worth optimizing.