Early products need a single number because the room is small and the story is simple. After the first million sessions the story is not simple. Guest traffic, seasonal campaigns, and a paywall that moved three times have all poured into the same tile. The north star still glows. It no longer points.
We see this often in United Kingdom consumer apps that grew through 2024–2025 performance spend. The original star mixed logged-out browse with a completed value action because, at the time, almost everyone who opened the app was a hopeful member. That is no longer true. Keeping the star becomes a loyalty to a slide template.
Retirement should be written like a metric letter. Name what the star used to mean. Name the population it now accidentally includes. Propose a successor that a sceptical founder can repeat. Run both numbers in parallel for two reporting cycles. Then stop printing the old one in the opening pack. Hiding it in an appendix is allowed; leading with it is not.
Dashboard Rhetoric, our short evenings for people who sign board packs, exists because founders often fear that changing the star looks like failure. It looks like literacy. App analytics is allowed to grow up. The studio’s job is to make the successor harder to game than the original, which usually means excluding SDK ghosts and requiring a human action that costs attention or money.
If you are mid-argument, bring the two definitions to a producer. We will not pick your star for you on a contact form. We will tell you whether the successor still mixes guest and member water.