Onboarding: first-session nudge
RUNNINGONBOARD_NUDGEHYPOTHESIS
The trial-to-paid leak isn't a signup problem, it's a first-session problem. Whatever a trialist achieves in session one decides whether they come back. We test which first task makes the product feel "already working": connect a channel, build the website, or import existing bookings. 48-hour activation is the metric and trial-to-paid is the payoff we're really moving.
THE VARIANTS · first-session prompt
DATA · 48h activation rate (illustrative)
STATISTICAL READ
C leads by +5pts on activation
89% probability C is the true winner · 5,214 trials enrolled · decision rule: read at n = 6,000 or ≥95%, whichever first. Still running. Early signal: an imported calendar makes the product real on day one. But we wait for the number, not the story.
WHAT +5PTS OF ACTIVATION IS WORTH
If the correlation holds: +5pts activation on ~5,200 trials = +260 activated trials/month, and activated trials convert at roughly double the rate. If that carries through to trial→paid, it's worth ≈ +€3–4k new MRR per monthly cohort. That's exactly what the test is for: we don't bank the correlation, we read the conversion number before calling it.
GUARDRAILS · tracked while it runs
So far the push isn't costing anything downstream. Cancellations, support load and onboarding-call uptake all hold across arms. If C wins on activation and the guardrails stay flat, it ships.
Illustrative data. The mechanism is the pitch, not the numbers.