Usage Is the Receipt

The one-sentence version

If nobody uses it, it is worth nothing — no matter how sound the methodology, how elegant the model, or how impressive it looked in the room where you built it. Usage is not a nice-to-have metric alongside "real" success measures. In an analytics org, usage is the only success measure that cannot be argued with, because it is the one thing a stakeholder cannot fake enthusiasm for in a meeting and then quietly never open again.

Why "we shipped it" is not the finish line

A board meeting celebrates a million registered users. Champagne, applause, a great slide. Three months later the real number surfaces: a sliver of those users ever came back after the first thirty seconds. The team was measuring signups when the business needed engagement, retention, and revenue — the things that actually keep the lights on. It is the same failure at the analytics team level: forty-seven dashboards built this year, proudly announced, until someone asks how many have five or more weekly users and the room goes quiet.

The manufacturing framing makes this concrete: what percentage of raw material — leads, signups, data, requests — actually becomes a finished product someone repeatedly chooses to use? If that yield is in the single digits, the team is not underperforming at building. It is over-performing at building the wrong things, quickly.

Halo metrics beat metric spam

The instinct after getting burned by a vanity metric is to start tracking everything. That is its own trap — twenty metrics on a dashboard is the same as zero, because nobody can hold twenty numbers in their head during a decision. The better move is picking three to five halo metrics: measures with enough gravity that improving them forces a dozen other things to get better on their own, and that cannot be gamed by cutting a corner.

"Percentage of solutions used weekly by five or more people" is a halo metric — you cannot improve it by shipping more, only by shipping things people actually want. "Average time from idea to production" is another — you cannot improve it without killing the waste in your own process. Pick metrics that make the uncomfortable truth visible, review them weekly as a team, and treat any metric that stays comfortably green with zero effort as a sign it has stopped measuring anything real.

What to check this week

Pull usage data for everything your team has shipped in the last year. Flag anything with no meaningful activity in ninety days. You do not have to delete it today — but you do have to know the number, because most teams that skip this step are astonished by what it says about where their time actually went.