Ending an Experiment
This guide covers how to conclude any experiment in Elevate — marking it as complete, applying the winning variation, and handling post-experiment cleanup.
When to End Your Experiment
Before ending your experiment, check the statistical significance status on your Results page:
Significant — The results are reliable. You can confidently act on the winning variation.
Near Significance — Getting close, but not yet confirmed. Consider letting it run longer.
Collecting Data / Trending — Too early to make decisions. Ending now means the results are directional at best.
For a full explanation of what each status means, see Statistical Significance.
You don't have to wait for significance to end a test — but if you end early, the results may not be reliable enough to act on.
Step 1: Mark the Experiment as Complete
Open the experiment from your Elevate dashboard
Click the Settings dropdown in the top right
Select Mark as Complete
This ends the experiment immediately. Traffic will stop being split, and all visitors will see the original experience.
Step 2: Apply the Winner (Optional)
If a variation won and you want to keep the changes, click Apply Winner from the experiment page. What this does depends on the experiment type:
Price Testing (Duplicate Products)
Updates the original product's price to match the winning duplicate. Duplicates remain archived.
Price Testing (Native / Shopify Plus)
Sets the winning price directly on the product.
Product Image Testing
Applies the winning images to the product.
Page Testing
No automatic application — publish the winning template as the default in your Shopify theme editor.
Theme Testing
No automatic application — publish the winning theme from Online Store → Themes in Shopify.
Visual Editor Testing
No automatic application — create a Personalization with the same changes targeted to all visitors to keep the winning experience live, or implement the changes permanently in your theme.
Custom Code Testing
No automatic application — create a Personalization with the same code targeted to all visitors, or add the winning JS/CSS permanently to your theme.
Split URL Testing
No automatic application — update your store to use the winning URL as the permanent page.
Shipping Testing
Update your Shopify shipping rates to match the winning configuration.
Checkout Testing
No automatic application — create a Personalization with the same checkout configuration targeted to all visitors to keep the winning experience live.
Keep the Original
If the control won — or if you're not ready to make a change — no action is needed. Your store continues as-is. This is still a useful result — you've confirmed that your current setup is strong.
Step 3: Post-Experiment Cleanup
Update Marketing Materials
If you applied changes that affect what customers see (new prices, new page, different shipping rates), make sure it's reflected across all channels:
Email campaigns and automations
Social media posts and ads
Google Shopping / product feeds
Any landing pages or content that references the changed elements
Duplicate Products (Price Tests Only)
If your experiment used duplicate products, the duplicates are archived automatically when you complete the test. You can:
Leave them archived — they won't appear in your store but remain accessible
Delete them — if you want to clean up your product catalog
What Happens to Your Data?
Completing an experiment doesn't delete any data. Your results, reports, and analytics remain fully accessible on the experiment page. You can revisit them at any time to reference the findings or inform future tests.
You can also resume a completed experiment if you want to collect more data — see Experiment Lifecycle for details.
Last updated