I A/B test pricing pages for a living (almost). What I have learned building Pricecheck

Data analyst building Pricecheck slowly. Findings from 30 pricing experiments I have run or reviewed:

Annual default beats monthly default (28% more annual uptake, no monthly loss). The default is the decision for 70% of visitors. Three tiers, middle highlighted, decoy enterprise ("Contact us" converts nobody directly and lifts middle 12%). Charm prices (19 vs 20) work on monthly, vanish on annual (1,900 vs 1,912 annual shows no difference; buyers do annual math differently). The guarantee near the price beats the guarantee in the footer (same 30 days, 3x the effect when adjacent to the button). Free trials that require cards convert worse and retain better (card trials: fewer, stickier; no-card: more, flakier; pick your religion knowingly).

My religion: no-card trial, annual default, guarantee at the button. What is yours?

Same religion, one difference: card trials for regulated markets. German accountants will not hand card details to a trial, and forcing it selects for nobody. Know your buyer's card religion before picking yours.

Designer amen: guarantee-at-the-button is layout, not copy. Proximity is persuasion. The same sentence moves numbers by position alone. Design is pricing.

Proximity is persuasion, writing that down. Position beats prose. The button-adjacent guarantee is the highest-ROI pixel on the page.

Cohort wall: test card-vs-no-card by cohort retention, not trial conversion. No-card wins trials, card wins year-one retention. The religion question is really a horizon question. Pick your horizon, pick your trial.