The Week I Actually Told People About My App
For months I have been very good at building and very bad at the one thing that matters more.
I can lose a whole day to a caching bug. I can rewrite a homepage headline eleven times. I can archive Stripe products and rename database columns and read my own token logs at midnight. All of that feels like progress, and some of it even is.
None of it puts a single person in front of the thing I made.
This week I finally did the part I'd been avoiding. Here's what happened — the bugs, the fear, and the strangers who said yes anyway.
---
The thing I kept not doing
The math was never complicated. I had a product that worked. I had zero customers. The gap between those two facts is not closed by writing more code, and I knew it, and I kept writing more code.
Because building has a scoreboard. The cache hit rate goes up. The checkout clears. There's a green checkmark at the end. You can feel the progress.
Telling strangers about your app has none of that. It has silence, mostly. And the small, specific fear that someone will look at the thing you spent months on and not care.
So I built instead. For a long time.
This week I ran out of runway on that excuse.
---
What I found the moment I looked
The first thing that happened when I started letting people in was that they immediately found things I could not see.
Not because I'm careless. Because I'm too close. I know where every button is. I know you hit return to see the map. I know the recording works if you do it the way I do it. None of that knowledge survives contact with a person who has never seen the app before.
Here is an incomplete list of things real testers hit in the first thirty seconds, all of which I would have sworn were fine:
- A save prompt that popped up before they'd made anything worth saving. One tester said it "confused" her — and she was being kind. - The app defaulting to the wrong language for someone in an English-speaking city, with no obvious way to switch. She signed up, loved the design, and then couldn't use it. She told me she'd try again "when there's an English option." That's a polite way of saying she left. - Voice recording that took "a few tries" to start. For a product whose entire pitch is just start talking, the talking part was the confusing part. - Content running edge-to-edge on an iPhone with no margins.
Four different people. Four different bugs. One theme I couldn't ignore: the first thirty seconds of my "no setup" product were full of setup.
Every one of them pushed through and liked it anyway. But these were my generous testers — the ones who care enough to send feedback. The people who bounced silently didn't push through. They just left, and I'll never know how many.
That's the thing about only ever testing something yourself. You optimize the part you can see and you never find the wall everyone else walks into.
---
The bug I'm least proud of
One of them wasn't cosmetic.
Someone typed a real, personal, difficult thing into the app. A sentence about a family member they were worried about. And the app said nothing. Blank. No reply, no error, just silence — at the exact moment a person had made themselves a little vulnerable.
The cause was boring and technical: a response got cut off mid-way through generating, which broke the part that parses it, which triggered a cleanup step that was too aggressive and deleted the actual reply along with the broken bit. Three small mistakes stacked into one bad silence.
I fixed the stack. But the lesson wasn't technical. It was this: a product people talk to about their lives can never, under any circumstance, respond with nothing. Silence is the one output that isn't allowed. I hadn't designed for that, because I'd never watched it happen to someone who wasn't me.
You don't find that bug in your own testing. You find it in the transcript of a stranger trying to use your app for the thing it's actually for.
---
Then the part I was scared of
Here is the fear, said plainly: I did not want to message people.
Not cold strangers, not even warm-ish acquaintances. Sending someone a link to the thing I made and asking them to look at it felt like standing in a doorway holding something and waiting to see their face. I put it off for months.
This week I sent the messages. Not a campaign — just individual people, one at a time, in a community I'm part of. No pitch. I said some version of: here's the thing, try it if you want, and tell me the one part that didn't work — I'd rather hear that than anything kind.
And people replied.
One tried it and came back with, roughly, "I came away with new ideas — this was super cool." A stranger. Who used it cold, for ten minutes, and left with something they didn't have before. That's the whole promise of the product, happening to someone who wasn't me, for the first time I've ever watched it.
Another — someone technical, someone who builds cinematic AI tools and thinks in systems — called it "gentle, like a kind friend." Which was funny, because I'd spent a full day fighting to make sure the homepage didn't sound like a therapist. Turns out the warmth I was worried about was the thing a stranger noticed first and liked most.
Another offered — completely unprompted, on the same day we met — to put her name and face on a testimonial. Two days earlier the only testimonial on my site was a fake one I'd written and attributed to myself, which I'd deleted out of embarrassment. Now a real person was volunteering to be one.
Nobody roasted me. Nobody looked at the thing and shrugged. The fear had promised humiliation and delivered, instead, a small pile of bug reports and a few people saying this is good, and here's what's broken.
Which is exactly what I needed and could not get any other way.
---
What I actually learned
Every generous tester is worth more than a hundred anonymous visitors. The visitors bounce and say nothing. The testers hand you the exact sentence that explains why the visitors bounced. One of them essentially diagnosed my whole funnel leak for free.
The fear was wrong, not just unhelpful — wrong. Reaching out didn't cost me anything. It gave me four bug reports, three encouragements, and a real relationship with people building alongside me. The discomfort was pointing me away from the single most valuable thing I did all week.
Being useful is the whole strategy. The connections that worked didn't come from pitching. They came from me testing their apps properly, giving real feedback, saying the honest hard thing when their landing page needed it. You give first. The rest follows, or it doesn't, but you gave first either way.
I judge my own work harshly, and it leaks. At one point I opened another founder's app with a mental checklist — does it look amateur, are the colors right — and realized halfway through that the checklist wasn't a quality standard. It was my own impostor anxiety, pointed outward for once. Her app was substantive. I felt bad for a second, and then I understood something: the critic I run on myself is not a measure of anything. It's just a habit.
---
Where I actually am
Product-wise: it works. The bugs from this week are fixed. It's faster and cheaper than it was seven days ago in ways I can measure. The front door still sticks a little, and fixing that stickiness is the most valuable thing on my list — because it's the difference between the people who love it and the people who leave before they find out they would have.
But the real change this week wasn't in the code.
It was that I stopped hiding behind the code.
I have a deadline at the end of the year. For months I've treated that deadline as a building problem. This week made it clear it was never a building problem. It was a telling people problem, and the only way through it is the part with no green checkmark at the end.
So that's the plan now. Keep making it better. But mostly: keep telling people, one at a time, and keep asking them the question that actually matters —
What didn't work?
The answers are the roadmap. I just had to be brave enough to ask.