Decisive day      3 · hazard 10.5% · 31 conversions on that day alone
Cumulative        81.3% of this flow's conversions land by day 4
Current schedule  welcome d0 · activation nudge d7 · expiry sequence d12
Recommended       welcome d0 · activation nudge d2 · stalled check d5 · expiry d12
Reason            the d7 nudge arrives after the decision has already been made
                  for 92.3% of the users who will convert

The nudge is not wrong, it is late. On this flow the conversion decision resolves on day 3, which means the day-7 email is being sent almost entirely to users who already converted or already stopped opening. Moving it to day 2 puts it in front of the population that is still deciding. The day-5 stalled check is new: it is the first day where a non-converter is genuinely informative rather than just early.

Decisive day      13 · hazard 8.7% · 86 conversions on that day alone
Cumulative        9.4% of conversions by day 4, 21.1% by day 9
Shape             flat hazard through day 9, then a spike at the expiry sequence
Reading           the trial deadline is making the decision, not the product
Recommended       treat d13 conversions as at-risk revenue and instrument
                  first-value time separately before touching the email schedule

A hazard curve that is flat for nine days and then spikes at the deadline is the signature of users who never reached the thing the product is for and bought anyway because the clock ran out. That converts, and it churns. This one is not a schedule fix. Sending the expiry email earlier would move the spike, not the outcome.

Completed only    28.3% on 1180 vs 36.0% on 100
                  z = -1.63, p = 0.10 · cannot conclude
Log-rank, all     chi-square = 44.8, p = 2.2e-11
Observed/expected Control onboarding: 351 vs 394 (0.89x)
                  Guided setup: 91 vs 48 (1.88x)
14-day total      28.2% [25.7% to 30.7%] vs 32.0% [25.8% to 38.1%]
Status            timing separated · totals not separated

Two findings, and only one of them is settled. The conversion hazard of the two flows is different beyond any doubt, which is enough to change the email schedule today. The 14-day totals are not separated: the intervals overlap, and guided setup has only reached day 14 for a hundred trials. Do not call the variant a winner on conversion rate yet. Do move the nudge, because that decision rests on the finding that is resolved.