Activation metrics that actually predict expansion
Most activation dashboards are autobiographies of the first week. They celebrate the clicks closest to sign-up because those clicks are plentiful, close in time, and easy to instrument. Two quarters later, when someone asks why expansion stalled, the same dashboard has very little to say. The early noise and the later money were never introduced to each other.
Inside Signal to Expansion we treat activation as a hypothesis about later weather, not as a participation trophy. The question is not “did they poke the product?” It is “which completed jobs, inside a named window, show up again when seats or usage expand?” If you cannot hold both ends of that sentence, you do not yet have an activation metric. You have a tour.
Three moments teams keep collapsing
First value, first collaboration, and first invoice are different climates. A founder who exports a CSV has found value. A teammate who uses a shared template has found collaboration. A billing admin who pays has found an invoice. Averaging them into a single “activated” flag is how a 61% slide survives three contradictory definitions in the same company.
We ask operators to write the three moments on paper before they touch SQL. The window matters: fourteen days is a common starting point for B2B workspace tools in our Bangkok studio, not a law. Consumer apps with daily habits may need forty-eight hours. Industrial tools with procurement residue may need a month. The sin is not choosing a window. The sin is refusing to name one.
Correlation without theatre
Once the moments are named, we look backward from accounts that expanded seats or usage in the following two quarters. Which early jobs appear more often there than among accounts that merely stayed? The method is deliberately unfancy. We are not selling a causal engine. We are trying to stop celebrating onboarding steps that never appear in the expansion set.
A recurring finding in our cycles: inviting a teammate looks like activation and often predicts very little. Completing a job that another human depends on — a shared template, an approval, a scheduled export — shows up far more often in later expansion. Your product may invert that pattern. The atelier’s job is to make the inversion visible, not to import our favourite example.
What this does not solve
Packaging still sits outside the metric. If two price tiers cannibalise each other, an elegant activation sequence will not pick a winner. Sample size is the other honest wall: below a few hundred weekly active accounts, the expansion look-back will not resolve. We would rather send you to the Activation Desk Intensive than invent statistical comfort.
If you want the full movement — instrumentation map, activation sequence, PQL vetoes, expansion scorecard — that is the flagship. If you only need to stop reporting first-week clicks as if they were a business, start with the written moments. Bring them to the desk. We will tell you whether they are ready to meet SQL.