How failing 7 times made me $1,000,000 (And how to replicate my results)

In late 2024 I quit my job as an engineer at Lyft to go fix my own doomscrolling problem. I put together a 4,000-word market-research report with 52 sources. Then I spent about 5 months building my first app and trying to market it. I gave it everything: filming TikToks, Product Hunt, ASO, and more.

It made $0,000,000.

But I learned.

The next app I worked on made $1,000,000.

The product wasn't really the difference. My first app was fine, even if very ugly. The people who used it liked it. What changed between $0 and a million was actually focusing on the things that matter.

You can verify the revenue here: https://verified.revenuecat.com/pushscroll

Who I am

I'm Mario, a software engineer working on B2C apps full time since I left Lyft.

The first one was SpeedBump, an intentionally annoying app to help you scroll less. I built it alone, and I made every mistake below with it.

The second is Pushscroll, which I build with my co-founder Alejandro (@skyirezumi). It makes you do pushups before you can doomscroll: 28,572,256 of them so far, and $100k+ MRR.

Before we start

Most apps fail. Mine did. I'm not a guru, and I don't really trust anyone selling a 7-step formula.

I'll also admit that luck is real. Years ago, as a clueless student, I made a silly Android app that got 5 million downloads. I had no idea what I was doing, the timing was just right (that's a story for another time). So I know some of this is random.

The 7 things below aren't, though. They're the part you actually control, and I got every one of them wrong before I got them right. You're probably making one or two of them right now.

Mistake 1: I built the whole thing before doing any marketing

I wrote the app first and thought about distribution second. Months of code before I had a real way to get it in front of people.

To be fair, I did talk to users a lot, it wasn't just coding. I just hadn't figured out how to actually distribute the thing, and B2C is ruthless if you don't know marketing. You can build something people love and still watch it go nowhere, which is just... sad. If a tree falls in a forest but no one hears it, did it really fall?

This is the easiest mistake to make if you're technical. Building is the fun part, so marketing becomes something you'll deal with later. before writing so much code, I should have figured out how my app would spread, not just whether people wanted it.

Mistake 2: I picked channels that suited me, not the user

I went hard on SEO and ASO. Of course I did. They appeal to the engineer brain: you set them up once and they compound, like building infrastructure. I paid 32€ for Ahrefs and burned through the credits the first afternoon.

The problem is that doomscrollers don't sit on Google looking for a screen-time app. They're on TikTok and Reels, which is exactly the thing I was trying to fight. I was optimizing for a place my users never went.

There's a second problem that took me a while to notice: the feedback loop is painfully slow, and early on that's exactly what you can't afford. ASO can take weeks to tell you whether a change did anything. Compare that to the video that first validated Pushscroll: it went up and within hours the comments made it clear people wanted it. At the early stage, that speed is worth far more than a channel that compounds slowly.

Mistake 3: I treated launches as distribution

Every indie dev does this, without fail. Hacker News, IndieHackers (does anyone actually use that site?), Product Hunt and its millions of clones, a launch post on X. I did all of them and even got to #7 of the day on Product Hunt.

It felt great for about a day. Product Hunt's own account even quote-tweeted the app, calling it "like paying a Gen-Z intern to annoy you". A spike of traffic, some upvotes, a handful of signups, it felt awesome. But then the graph went right back to flat.

For B2C this is a horrible strategy. Product Hunt is not a distribution method, it's just not what it's for. A launch is a one-time thing, you can't do it every Tuesday, and what you need is something that keeps bringing people in next week too. A spike won't do that.

The same goes for building in public on X. It looks like marketing, but you're mostly talking to other founders, not the people who'd use your app. Unless your customers actually live on X, it's the wrong place. Mine were mostly on Instagram and TikTok.

Mistake 4: my content was unauthentic, and too much of it was AI

One of my carousels got over 1.5 million views. It got me around... 1.5k downloads.

That should have felt like a win, and instead it taught me the opposite. It was a carousel made to rack up views, not to convert anyone. The topic was education, where it's easy to pull in a bunch of students who can't afford to pay for an app. Your goal should be to get customers, not views.

The deeper issue was that too much of it was AI-driven and average. The whole point of virality is that you can't be average - that's what social media users scroll past. AI is fine as an assistant, to speed things up or get unstuck, but the moment you let it drive everything you end up with generic content that gets no reach and convinces no one to download. Authenticity is key.

Mistake 5: I wasn't brainrotted enough

A confession: when I started doing marketing in early 2025, I didn't even have TikTok installed.

You can't really write hooks for a feed you don't live in. You pick it up by using the app a lot, not by reading about it.

The people who win on these platforms are natives. Their own scroll gets stopped every day, so they know what stops a scroll. Turns out brainrot is part of the job. Looking back, I should have started on Instagram, which has a vibe I'm more used to. Being less stingy and paying for a course would've helped, too.

Mistake 6: I was spread too thin

I'm Spanish, I was living in Germany, and I spent most of my day speaking English (the European starter pack). So posting in all three languages sounded like an easy way to 3x my views. It wasn't.

Localizing the app in a few languages can be reasonable. Doing marketing in three is a different story. I kept them on separate accounts, so the problem wasn't the algorithm getting confused. The problem was focus. As a solo founder it's already hard enough to stay focused on one thing, and proper localization is always harder than it looks. Not to mention that one of these languages was German, which I don't even speak that well. Genius move.

So pick one language and one audience, and put all that energy into making it good. If you're worried that English won't fit a mostly-EU audience, don't be. With Pushscroll our audience is mostly EU too, and our English posts worked fine (especially on younger audiences).

Focus, focus, focus. I still fail at this all the time.

Mistake 7: my app was free

SpeedBump was free because I assumed that would help it spread.

A free app with no revenue has no flywheel. Fifty thousand users from a viral video sounds great, but you can't turn it into more growth if there's no money to put back into ads or creators. You get a wave of users and then it's gone.

The "free, grow first, monetize later" plan is for VC-funded apps whose whole job is growth at any cost. That wasn't me, and it's probably not you either. Unless you already have a million in the bank to burn, the revenue is what pays for the next round of distribution. So charge from day one.

There's another reason to have money coming in. You don't want to be doing manual marketing forever, especially if you don't enjoy it. Once there's enough revenue you can switch to paid ads or hire someone to handle it, and go back to the part you actually like.

What changed

Then I met Alejandro on YC co-founder matching. I'd been failing at this alone for months, drowning in marketing I couldn't crack. He had the opposite problem: shortly before we met, he'd done the one thing that inverts mistake #1.

Before any code existed, he'd posted a fake demo of Pushscroll: an AI pushup-detection clip he found, a phone propped against a wall, some scrolling footage, all stitched together. No app yet (he never claimed there was one).

It got around 80,000 views and roughly 500 comments asking for it to be built. He'd cracked the viral side and needed someone to build it. Perfect match.

That's basically the whole article in a single move. We'd proven people wanted it, and that it could spread, before building it.

The rest came from getting the 7 things right. We launched by promising free access to the first users (honestly not sure I'd recommend that part, but it worked for us), put up a hard paywall so the app could fund its own growth, and went all-in on authentic, single-audience content on platforms we actually used. If you want the full tactical playbook, Alejandro wrote it up here.

The takeaway

My first app didn't fail because the product was bad. It failed because I got distribution wrong in 7 ways, and all 7 were things I could have fixed. It cost me months to learn them, but at least it was only once.

Trust me, it's a lot more fun to build a great product for thousands of users than for three people and your mom. Leave the engineering mindset and get your app to more people.

Next article: how I went on Fox News (the Bay Area one) live on TV and got exactly 0 downloads out of it. It's my favorite fail.

Follow me @null_deref so you don't miss it.