What this report is about
Every visitor who landed on the product page during the 10‑week test was randomly shown one of eight versions (A = today's live page, B–H = candidate redesigns) and stayed on it for the whole test. We tracked the full funnel: landed → clicked the primary CTA → added to cart → purchased. Below, "conversion rate" always means purchases ÷ all visitors — the metric that actually reflects revenue impact.
Bottom line
1. Overall conversion rate by variant
Purchases as a share of all visitors — the single number that matters most. Bars show the observed rate; the dashed line marks variant A (today's page) for reference. Variants are colored green if statistically significantly better than A, red if significantly worse, grey if not distinguishable from A given the sample size.
2. Where in the funnel do variants actually win or lose?
Conversion rate is the product of three stage rates: click‑through rate (CTR), add‑to‑cart rate (of clickers), and purchase rate (of adders). A variant can look great on one stage and be mediocre overall — this chart shows why.
3. Was this consistent over the 10 weeks?
Weekly conversion rate per variant. Click a legend item to isolate a variant. If the ranking had been noisy or driven by a single odd week, we'd see lines crossing constantly — instead the winner and loser are stable from week to week.
4. Sanity check: was traffic split evenly?
Confirms the randomization worked as intended — every variant got essentially the same number of visitors, so differences above aren't an artifact of unequal sample sizes.