I want to step out of the timeline for a minute. Most of these posts have been about what I built last week. This one is about something that happened earlier — back when the prototype barely worked, before any testers, when the only people I'd shown it to were friends.
The trap I might be standing in
There's a thing every founder gets warned about. Don't build for a problem nobody has. If people don't engage with the idea, that's the market telling you something. The polite nodding is data, and the data is not on your side.
I'm aware of this. I know what the trap looks like from the outside. I might still be standing in it. I want to name that plainly before I say anything else, because not naming it is how the trap closes around you.
The friends
I'd describe Bounty to friends. They were polite. The part of the description that got engagement was I'm using AI to build it — that's the thing people wanted to ask about. The grocery part got a different reaction. The grocery part got the same shrug every time. Yeah, prices are bad. What are you going to do.
It took me a while to read the shrug correctly. They weren't disagreeing that there was a problem. They had quietly decided there wasn't a solution. That's a different thing, and it's harder to push against. You can argue with someone who thinks the problem isn't real. You can't argue with someone who has accepted that the problem is real and permanent.
"There's apps for that already"
The other thing that came up, every time, was some version of isn't this already a thing. Flipp. Reebee. The store apps themselves. Flyer aggregators. Cashback apps. They'd name one and look at me like I should already know it.
I do know them. I've used most of them. And the honest answer — the one I'd start trying to give before realizing it wasn't landing — is that none of them actually do this. Flyer apps tell you what the store wants you to know about pricing. Store apps tell you what the store wants you to pay. Cashback apps pay you a few cents after you've already paid the full price. Every one of those tools is built from data the grocer chose to publish. None of them is built from what's actually on the shelf, on a Tuesday afternoon, in front of you.
That's the gap. The gap is that there's no record of grocery prices that exists independently of the grocer. And until there is, no app on either side of that wall can really help, because they're all working from the picture the grocer drew. That difference is hard to convey in one sentence at a dinner table. The shrug was usually faster than the explanation.
Why I kept going anyway
The honest reason: the technology kept working a little better each week, and the picture in my head kept getting more specific. I had nothing to push against the doubt except the prototype itself, and the prototype kept earning the next decision. Around seventy percent of scans came back right, and seventy percent was bad enough to feel like a problem and good enough to feel like something real.
Then, once I was building toward a version that might actually work, two questions arrived at the same door and got answered together.
The crowd, and the wall
The first question: who does the work. I couldn't be the only scanner — not even for the small version. The data had to come from regular shoppers, doing one extra second of work in an aisle they were already standing in. That's the only ask the app gets to make. Anything more and the friction kills the loop. Anything less and there's no data. The smallest possible effort, repeated by enough people, builds something none of us could build alone. If it's a little fun on top, even better. That's where the Bounty side of Bounty came from.
The second question: how does an app like this not become the thing it's trying to fight. The playbook says track the users, sell the data, monetize the attention. I sat with that for a long time. The version that runs on user surveillance is the version that turns the people I'm trying to help into the product. So that got walled off, early, in writing. The grocers can buy the picture the crowd builds. They can never buy the people in it.
The harder admission
This whole thing comes from wanting to help. I see prices going up, I see people getting quietly squeezed, I see the resignation in everyone I talk to about it, and I want to do something about it.
But the part the friends keep showing me — without meaning to — is that wanting to help people is not the same as people wanting help. Some of them would rather take the shrug. The risk of building something like this isn't that the technology fails. The risk is that you build the right tool and nobody picks it up because they've already made peace with the problem it solves.
I'm building it anyway. The picture in my head got too specific to stop. But the trap is real, and I'm trying to keep my eyes open about it.
— Elmer