10 secs Pages that moved a number.
Short, screenshot-backed write-ups of what real ecommerce teams changed and what it did to conversion, plus longer teardowns: a whole page annotated change by change.
3 short examples, 2 CRO tips, 1 teardown. Every Tuesday.
Two kinds of card
An example is a real client, a published number, and the page it happened on; its card says whether the figure is an engagement outcome or a controlled test, because only one of those isolates a cause. A teardown carries no result number. It is a whole page annotated, and the card counts the changes instead.
Where the losses are
Not on this wall. These are engagement outcomes, and an engagement that went nowhere does not produce one. The losing tests are counted in the case studies, where 39 of 71 tests beat control and the other 32 are part of the same number.
How to use them
Treat each one as a hypothesis worth borrowing, not a rule. A rebuild changes many things at once, so what moved a number on one store is a starting point for your backlog; whether it works on yours is a question for a test. The teardowns say per pin whether a change is safe to ship or needs an experiment.
10 secs
10 secs Denik's rebuild cut CPA by $18 and lifted revenue per visitor 24%
10 secs Erth Wellness: +18% conversion and +32% average order value
10 secs The Facial Works: +29% conversion and +35% ROAS after the rebuild
10 secs A starter-kit product page built to convert: +37% CVR, −25% CAC
10 secs +55% cold-traffic conversion let Multiblue scale ad spend 2.5×
10 secs A trust-first sleep gummy store holding 3.8% sitewide conversion
10 secs Rebuilding one hero SKU page lifted Raan 29%, add-to-cart 31%
Teardown 10 secs