Measure per-role, not just the headline number

The headline activation rate is an average, and an average hides exactly the thing this whole approach exists to fix. A blended 37 percent could be 60 percent of admins activating and 15 percent of operators churning, and you would never see it in the top-line number. The discipline that makes role-based first runs compound is to report activation and time-to-value per role, every week, and to treat each role as its own funnel with its own benchmark.

For each role, track two numbers: the share who reach that role's activation event, and the median time they take to get there. Activation rate tells you how many get there; time to value tells you how fast, and the two move independently, so you need both. Median time to value across hundreds of SaaS products sits around a day and a half, but that median swings wildly by category and, inside your product, by role. The role with the worst time-to-value is leaking your best-matched users the slowest and most invisibly, which makes it the most valuable thing on your roadmap. Watching the retention curve split by role tells you whether the activation lift is real or cosmetic, because a genuine role-fit win flattens that curve for that cohort.

The per-role view turns a vague "improve onboarding" goal into a specific, falsifiable one: pick the role furthest below the 40-to-60 percent good-activation band and shorten its path. Because each branch is independent, you can rebuild one role's first run without touching the others and read the result cleanly. This is the loop that takes you from a typical 37 percent median towards the 40-to-60 percent good range and, for the strongest paths, past 60. Role-based first runs are not a one-time build, they are a per-role optimisation engine, and the measurement is what keeps it honest. Feed the activated users straight into the lead-scoring and handoff motion so activation becomes pipeline, not just a product metric, and watch the early-retention lift compound into ARR the way the compound-growth maths predicts.