When a test wins, we make it part of the site
Steadyrack run conversion experiments across three regional stores to find what actually lifts conversion. Plenty of teams stop at the win, and the test just keeps running inside the testing tool month after month, a third-party script deciding the outcome on every page load. It works, but you are renting the result and serving it slower than you need to.
What we did
We do the step after the win. When a test proves itself, we take the winning version and build it into the theme properly, so it becomes permanent site code Steadyrack own and the testing tool can be switched off for that experiment. One was a collection layout that splits a product into its size options, each with its own card, and another was a product-page block that updates to show which size suits which use as the customer chooses. Both were proven in testing first, then hardcoded across all three stores.
The result
The wins are permanent, they load faster because they are native theme code rather than a script fetched on every visit, and they belong to Steadyrack outright. The testing tool goes back to what it is actually for, which is finding the next win rather than hosting the last one forever.
What to take from it
A winning experiment is only worth what you do with it. Leave it live in the testing tool indefinitely and you are paying every month for a result you could own, and serving it slower than the customer deserves. When a test wins, the move is to build it into the actual site.
