I built a phone security app for android. solves a real problem, decent market size. product works, live on google play, shipped 14 versions in 110 days.
building was the fun part. marketing was... educational.
The Automation Bet
I'm a developer so naturally I solved my marketing problem by building more software. wrote a custom automation system from scratch — native twitter and meta APIs, cron jobs, rate limiting, token refresh, the whole thing. no buffer, no hootsuite. spent about 40 hours on the infrastructure. it was genuinely impressive.
Native platform APIs (not third-party wrappers), automated token refresh, rate-limit handling with exponential backoff, multi-platform scheduling from a single JSON config. A proper piece of engineering for a problem that wasn't actually engineering.
The AI Bet
I leaned into AI for the content side too — generated copy, optimized timing, A/B tested formats. thought this would be the cheat code. turns out you can have a perfectly functioning machine delivering perfectly generated content to nobody. AI didn't solve the distribution problem. it just made it faster to confirm I had one.
the tools were ready. the distribution problem wasn't.
The Numbers
15-day campaign over the holidays. 81 posts across twitter/ig/facebook. 31 actually went through (rest got blocked by the platforms fighting back).
the parking lot actively told people to leave.
Best performing tweet? "the 9-to-5 was designed in 1926" — 87 impressions, 24% of the entire campaign reach. it had nothing to do with my app. that ratio tells you everything about posting generic content vs. posting content people already want.
Why It Failed (obvious in hindsight)
posting to 26 followers means roughly 0-5 people see each post. the algorithms literally have nobody to show your content to. then the high volume + zero engagement trains the algorithm you're spam, so each post gets suppressed harder than the last.
I was actively teaching twitter to hide me. 81 posts out, zero comments on anyone else's stuff. broadcasting to empty rooms on a schedule.
Four Pivots
Broadcast campaign
The 81-post failure documented above. Built the infrastructure. Delivered to nobody.
Engagement-first
went into communities, replied genuinely, but with 3 karma everything gets buried. Same gatekeeping problem, different layer.
Embedded growth
built interactive tools and SEO content. genuinely believed AI had finally matured enough to unlock a different distribution path for solo founders — no team, no audience, just leverage the tools. turned out the tools were ready. the distribution problem wasn't.
Sovereign distribution
stop chasing attention and let it come. made the core product free — emergency wallpaper, QR code on lockscreen, SMS activation. built it so every use naturally exposes it to new people. creating content people are already googling. showing up where phone security is an actual concern, not a pitch.
too early for results. but the mental shift is different. I stopped building systems to push content at people and started thinking about where attention already wants to go.
The Interesting Part
the part that messes with my head: documenting this journey honestly might be the distribution. this post is more interesting than anything I wrote in those 81 scheduled posts. the failure story is better content than the product pitch ever was.
The Two Takeaways
content without distribution is invisible. the cold-start problem is a distribution problem. and distribution at zero is fundamentally different from distribution at 1,000 or 10,000. every piece of advice out there assumes you have some baseline audience. when you have literally nothing, the rules are different and nobody tells you that.
I made the core product free — not freemium-free-but-actually-useless, genuinely free. emergency wallpaper, lockscreen QR code that links to your contact info, SMS activation — all included. the premium is just the web dashboard for remote control. turns out "it's free" is the most effective marketing copy that exists.
what actually worked for getting your first handful of users? not growth hacking, not marketing frameworks. the actual messy thing that got real humans to try your stuff when you had nothing.
This is part of the Distribution Lab — a running public record of every marketing experiment we run building FINDERR from zero, with real data. No survivorship bias. Pivot #4 is currently live. The breakthrough hasn't happened yet. We're documenting until it does.