Tracking & Analytics

Analytics gets framed as a dashboard problem: collect the right metrics, look at them often, optimize. That framing is half-right and half-wrong. It's right that you can't refine what you don't measure. It's wrong that more measurement is better. Most measurement is theater — numbers tracked because they're easy to track, reviewed because there's a habit of reviewing them, never tied to a decision that would change. The metric that matters is the one you'd act on this week if it moved.

Working on Analytics means picking three to five numbers that are actually leading indicators of the outcomes you care about — and looking at them on a cadence that's short enough to react but not so short you confuse noise for signal. For most people, that's weekly or biweekly, not daily. The other move is reviewing in company: a mastermind group looking at five members' five-number dashboards together produces better calibration than any individual reviewing alone. The numbers don't refine you; the conversation about the numbers does.

The failure mode is the dashboard everyone has, no one reads, and no one acts on. The graph is colorful. The alerts fire. The Slack channel pings. And the same wrong things continue to happen because nothing in the workflow says: this number moved, therefore I'm doing this differently. A metric not tied to a decision is overhead. The fix isn't more metrics. The fix is fewer metrics with one decision attached to each.