Data Infrastructure
When every team defines the KPI differently, you don’t have insight — you have a meeting factory.
Product and growth were shipping, but nobody trusted the readouts.
That is a quiet crisis: not missing data, but missing agreement. The fix was not another dashboard—it was naming the few numbers that mattered and who owned them when they moved.
Problem
Metrics lived in more than one tool. Event names drifted. People reconciled by hand before they could decide.
Experiments were slow to read because the baseline kept moving.
Insight
A tighter event taxonomy and written business definitions removed a lot of interpretation overhead.
Once people trusted the same numbers, execution got faster — not because of dashboards, because of alignment.
What I did
- Standardized the tracking schema on core product surfaces.
- Rebuilt core dashboards around lifecycle, retention, and monetization — not every side metric.
- Documented owners and definitions for the KPIs that actually steer the business.
Impact
- Roughly 60% less time reconciling dashboards each week.
- Experiment readouts people could stand behind — fewer shadow analyses.
- Weekly planning that used leading signals instead of lagging arguments.