Pricing page: lead message
CONCLUDED · B WONPRICE_LEADHYPOTHESIS
Price-led visitors aren't only asking "how much", they're asking "will this bite me later". If the hero leads with transparency ("no surprises, public pricing") rather than the number itself, more visitors start a trial and survive day one, because the expectation set matches what they find inside.
THE VARIANTS
DATA · trial-start rate by week (illustrative)
A · From $16/moB · No surprisesC · Cancel anytime
STATISTICAL READ
B beats A by +1.2pts
98% probability that B is the true winner · n = 4,912 visits · decision rule: ship at ≥95%. Shipped. Next: B's framing is now the control, and the day-7 retention of B-acquired trials is under test, because a better message that attracts worse trials would be a false win.
WHAT +1.2PTS IS WORTH
On this one page: 4,912 visits/month × +1.2pts = +59 extra trials/month. At ~20% trial→paid and ~€62 new MRR each, that's ≈ +€730 new MRR per monthly cohort (~€8.8k ARR), from changing a headline. Modest on its own, but it stacks. This same test runs on every page in the system and the winning message compounds into ads, sitelinks and retargeting.
GUARDRAILS · did the lift cost us anywhere else?
The uplift is clean: B brings more trials without bringing worse ones. Cancellations, refunds and plan mix all hold. That's the check that separates a real win from a vanity win. A message that inflated starts but degraded cohorts would have been rolled back.
Illustrative data. The mechanism is the pitch, not the numbers.