Everyone Is Shipping. Very Few Can Finish the Sentence.
Success metrics aren't a reporting exercise. They're a thinking tool — and most teams reach for them too late.
“We’ll know this worked when…”
Try finishing that sentence about the last thing you shipped. Not the OKR. Not the dashboard tile. The actual sentence: when this is live, what observable thing will be different?
Most teams can’t finish it. That’s not a metrics problem. It’s a thinking problem.
Metrics as a thinking tool
The conventional view treats success metrics as something you set up after the work is scoped — a reporting layer for executives. That framing misses what metrics are actually for.
A success metric is a forcing function. It makes you commit, in advance, to what you believe is true about the world and how this change affects it. If you can’t write the metric, you don’t have a model yet. The model is the part that matters.
The pattern I see
Teams that ship without a clear “we’ll know this worked when…” tend to do one of two things after launch:
- Declare victory by activity. “We shipped on time, adoption is rolling out, the team is moving.” None of those measure the thing.
- Retrofit a metric. Find a number that moved and claim it as the win. This is how you end up with a portfolio of features that all “worked” while the business doesn’t.
Neither is malicious. Both come from skipping the part where you make a falsifiable claim.
What to do instead
Before the work starts, write one sentence. Before the launch, sharpen it. After the launch, check it. If you were wrong, that’s information — not failure. The thing you wanted to avoid was being wrong silently.