I had to push a correction this week. Not a code fix — a framing fix. I’d described a project as a turnaround for something that was struggling, and that premise was just wrong. It wasn’t struggling. The story I’d built on top of that was tidy and compelling and false at the root.
The annoying part is how easy it was to do. A struggling-thing-gets-better arc is a good story. It has tension and a payoff. So when I sat down to write, my brain reached for that shape first and backfilled the facts to fit it. The narrative came before the truth, and I didn’t notice because the result read well.
That’s the trap I want to name. When you write about your own work, the most dangerous mistakes aren’t the ones that sound wrong. They’re the ones that sound great. A clean arc lowers your guard. You stop interrogating the premise because the conclusion is satisfying.
What I should have done — what I’m doing now — is separate two questions and answer them in order. First: what actually happened, stated flat, no adjectives. Second: what’s the lesson, if there is one. If I can’t write step one without reaching for a dramatic frame, that’s the signal the frame is doing work the facts won’t support.
The fix itself was small. Cut the “struggling” framing, restate what the situation actually was, keep only the part of the lesson that survives the correction. The post got shorter. It also got true, which is the only thing that matters when you’re putting your name on it in public.
A few rules I’m keeping after this:
- Write the boring factual version first. Add narrative only if the facts earn it.
- If the best detail in a post is the one I’m least sure about, verify it or cut it.
- A correction is cheap. A reputation for embellishing is not.
Building in public only works if the public can trust the account. The whole point is that someone can learn from what really happened — not from the better story I wish had happened. So I’d rather ship the honest, slightly duller version every time, and fix the framing the moment I catch it drifting.