This walkthrough assumes you've already created and designed your alternate template in Shopify. If not, start with Managing your variant templates.
Create the experiment
In Cooee, go to Experiments and create a new experiment.
Choose Template experiment as the type.
Give it a name that describes your hypothesis, e.g. "Product — reviews above the fold".
Cooee sets up two arms automatically: a Control (your normal template) and a Variant.
Choose the page type
Select the page type you're testing — Product, Collection, Homepage, Page, Blog, Cart, or Search. Cooee reads your live theme and lists the alternate templates available for that page type.
Pick the variant template
In the Variant arm, choose the alternate template you built. If the list is empty, the template either isn't on your live theme yet or doesn't exist for this page type — see Managing your variant templates. If a previously-selected template is missing from the theme, Cooee flags it with a warning so you can fix it before launching.
Target the right pages
By default the experiment applies to every page of the chosen type. To narrow it, add one or more target URLs — only pages whose path matches run the experiment.
Enter the page path, e.g.
/products/wavbeach-amalia-daypack.Use a trailing
*as a wildcard prefix, e.g./products/wavbeach-*matches every product whose handle starts withwavbeach-.
Matching is done against the page's canonical path. Shopify serves the same product under more than one URL — for example both /products/wavbeach-amalia-daypack and /collections/amalia-daypack/products/wavbeach-amalia-daypack — so you only need to list the canonical /products/<handle> path. Cooee matches it no matter which route (or query string) the visitor arrived through.
Set the traffic split
Choose how traffic is divided between Control and Variant (50/50 is the default and usually the fastest to reach significance).
Launch
Publish the experiment. From then on:
A visitor who lands on a targeted page is deterministically assigned to Control or Variant.
Variant visitors are sent to
?view=<suffix>and see your alternate template; Control visitors see the normal one.The assignment sticks for that visitor for the life of the experiment, across visits and pages.
Stopping the experiment
When you stop it, Cooee stops adding ?view= and every visitor sees the control template again. There's no cookie or redirect left behind. Your alternate template remains in your theme until you remove it (see cleanup in Managing your variant templates).
