2 min read
In brief
- A CVR drop after a theme update is usually speed, above-the-fold layout, or a broken trust element — not the design being worse.
- The drop appears after the cache turns over, so compare the date against the deploy, not against the day you first noticed.
- Measure mobile separately: theme regressions nearly always hit mobile hardest.
The six things a theme changes
- Speed — new fonts, larger images, an extra app script, a carousel that loads three screens of product before the first paint.
- Above the fold — price, delivery promise and add-to-cart pushed below where they were.
- Trust elements — reviews, payment icons and the returns line removed during a redesign and never restored.
- Navigation — a menu that hides collections two clicks deeper, reducing product discovery.
- Product page structure — variant selection moved, size guide removed, stock messaging changed.
- Checkout entry — a cart drawer replacing a cart page, or the reverse, changing how many steps a buyer takes.
Measuring it without guessing
- 1
Split by device
Mobile and desktop separately, before and after. A blend hides the regression.
- 2
Find the date
Daily CVR across six weeks. The drop has a date; that date is either the deploy or the cache turnover.
- 3
Compare the funnel, not the ratio
Sessions to product views to add-to-cart to checkout. The step that broke tells you which of the six it is.
- 4
Test one fix at a time
Restoring trust elements and speed simultaneously leaves you unable to attribute the recovery.
Recovering without reverting
A full rollback restores performance and loses the design work, and a second large change within weeks makes any later measurement ambiguous. Restore the specific elements — reviews block, payment icons, delivery line, speed budget — inside the new design, and give each a week.
The theme is not the cause; the six things it changed are. Split by device, find the date, compare the funnel, and restore the elements that were doing the selling.
