Building for the corner shop, not the dashboard

I'm 3 months into building Poppify — an AI marketing team for solo founders and small businesses.

The plan when I started was the usual one: ship features fast, find PMF, scale. Then two things made me change it, and I wrote about both at [poppify.ai/research/before-poppify](https://poppify.ai/research/before-poppify) before I'd written a line of marketing copy.

The short version:

I kept walking past the corner shop on my street — the one with the handwritten window display, the owner who knows your name — and watching it lose ground every month to the chain three blocks down. Not because the chain is better. Because it's louder. The corner shop posts twice a month if it's lucky. The chain has a marketing team and an algorithm.

Then this: the AI tools getting built around me are mostly aimed at people who already have an audience. The loudest getting louder. Same shape as every technology shift before it.

So before shipping the next feature, I sat down and wrote a whitepaper. Then four more. Five whitepapers now — on [the agentic marketing landscape](https://poppify.ai/research/agentic-marketing), on where SMB marketing value actually gets created (spoiler: ~80% of it is the owner's time), on what platform algorithms reward in 2026, on the competitive tooling landscape. All free, no email gate.

What surprised me: writing the research changed how I designed the product. The agent architecture wasn't marketing — it was the shape the research kept pointing to. The pricing model — pay per post, no subscription — fell out of the SMB economics paper. The Reels editor came directly from the algorithms paper.

It's slower than shipping features. But I keep coming back to the question my essay asked: do we need an AI revolution,or do we need something more?

Anyone else here writing before they ship? Curious which is working for whom — and for what kind of customer.