we've walked into operations with a data lake, dozens of dashboards and a busy data team — and growth flat anyway. The data was there. The decision wasn't. Collecting became an end in itself; the question “and what did we change because of this?” never had an owner.
Growing a digital product with purpose isn't instrumenting everything. It's knowing which decision the number should unlock — and having someone with the mandate to make it.
§ 01 / ThesisThe bottleneck is the decision, not the collection.
No company dies from lack of data today. It dies from an excess of data with no criteria. When everyone has access to everything and no one owns a single decision, data becomes wallpaper: pretty, present, ignored. Before instrumenting, ask which decision is waiting for that number. If there's no decision behind the metric, the metric is vanity.
§ 02 / The metric that mattersAn indicator that anticipates, not one that mourns.
Revenue is a lagging indicator. When it drops, the problem happened weeks ago — too late to fix. Activation, feature adoption, first value delivered to the user are leading indicators. They warn while there's still time to act.
- A controllable north star. A metric too high up the chain (revenue) is swayed by a thousand factors outside the squad; too low (uptime) doesn't connect to the business. The sweet spot is what the team directly influences and that clearly ties to the outcome.
- Segment before concluding. “Conversion dropped 10%” isn't a fact — it's a headline. In which segment? Is it significant? Seasonal? Without segmentation, you decide in the dark.
A lagging indicator mourns the past. A leading one changes the future. Products grow on the second.
§ 03 / Data inside the operationNot in a parallel report.
Data living in a report someone opens once a month doesn't change an operation. What changes it is data integrated into the system where the decision happens — the alert that fires in the right channel, the number that shows up on the screen of whoever acts, the reading embedded in the flow. It's the same logic we defend for AI: it works when it's inside the real operation, not when it lives in parallel.
§ 04 / ClosingPurpose before pipeline.
Growth with purpose starts at the business question, not the tracking tool. First the decision, then the metric that informs it, only then the pipeline that collects it. Inverting that order is like building an observatory without knowing which star you want to watch — expensive, impressive, and useless.
Data doesn't grow a product. A decision informed by data does. The difference is who's responsible for acting.
end · field note #50 · noûs / may 26