Year one is mostly a long argument with yourself about which mistakes were the bad ones. Here are the three we eventually conceded.

1. We built integrations before we had retention

Six weeks of engineering on Slack, HubSpot, and Zapier. We thought integrations would unlock retention.

They didn't. Users who didn't return after week one didn't suddenly come back because we connected to their CRM. Integrations were rewarding the users who already loved us. We needed something that pulled in the ones who'd left.

Lesson: Build for the user you don't have yet, not the one who's already happy.

2. We priced too low

$9 per month felt founder-friendly and accessible. It was. It also signaled we weren't a serious tool, attracted users who churned at higher rates, and gave us no margin to invest in support.

Six months in we tripled the price. We lost about 12% of new signups and gained roughly 40% in revenue per customer. The remaining users were more engaged, not less.

Lesson: Cheap doesn't mean accessible. It often means disposable.

3. We didn't say no to a big customer

A 50-seat enterprise wanted us to build a custom permissions model. The deal was real. The money was real. We said yes.

Three months later we had built features no other customer wanted, on a contract we couldn't profitably renew. The product roadmap shifted around their needs and the rest of the customer base felt the absence.

Lesson: A bad fit at scale is just slow damage. Politely declining is a real option.

What ties them together

All three felt like the obvious right call at the time. The pattern in retrospect is the same: we optimized for the loudest signal in the room. The quiet signals — the users who left without complaining, the customers who'd pay more, the small healthy growth — were always the ones we should have followed.