> For the complete documentation index, see [llms.txt](https://docs.elevateab.com/elevate-helpcenter/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.elevateab.com/elevate-helpcenter/product-testing/create-a-product-experiment.md). # Create a Product Experiment Step by step walkthrough for creating a product experiment in Elevate A/B Testing, including product groups, duplication, and the pre-launch review. *** Seven steps, plus a review before launch. **Before you start:** the theme extension enabled, and a decision about whether you are testing against a product that already exists or one you will create by duplicating. Open **Experiments**, select **New Experiment**, then **Product Experiment**. ### Step 1: General Info & Goal Enter an **Experiment name**. **Description**, **Hypothesis**, and **Assets** are optional. Under **What do you want to measure?**, choose your goal. **Revenue per visitor** is recommended. Select **Browse all metrics** for the full list, or **Add Custom Metric** to track something else. See [Custom Events](https://claude.ai/elevate-helpcenter/experiment-setup/custom-events). ### Step 2: Select Products Choose the products to test, up to **500**. **Search** by product title or ID. **Filter** by product type, status, collection, or tag. Status defaults to **Active**. **Select All** takes everything matching your current filters. **Upload CSV** imports a list. The counter shows how many you have selected against the cap. ### Step 3: Set Variations Choose the **Number of Variations**, then pair each product with something to test against. Each product you selected becomes its own **group**, labelled Group A, Group B, and so on. The control column lists your original products. The variation column asks you to **Select product to test against** for each group. | Option | What it does | Use it when | | --------------------- | ----------------------------------------------------- | --------------------------------------------------------------------- | | **Select Product** | Points the variation at a product that already exists | You have already built the alternative | | **Duplicate Product** | Creates a copy of the original that you then edit | You are building the alternative from the original. The usual choice. | | **Duplicate All** | Duplicates every group's product in one action | You are testing many products the same way | Duplicated products are created **Unlisted**, so they are reachable through the experiment and not through your storefront navigation, search, or collections. Use the edit control on any product card to change it, and **Add products to the group** to widen a group. After duplicating, edit each copy in Shopify to make the change you are testing. A duplicate that is identical to the original tests nothing. ### Step 4: Traffic Allocation Set the split with the slider. Even is the default. **Isolate traffic for this experiment** is optional and off by default. See [Traffic Isolation](https://claude.ai/elevate-helpcenter/experiment-setup/traffic-isolation). ### Step 5: Audiences Optionally narrow who enters, by device type, visitor type, UTM parameters, traffic source, or country. See [Audience Targeting](https://claude.ai/elevate-helpcenter/experiment-setup/audience-targeting). ### Step 6: Enhance Your Variations (Optional) Skippable. These layer extra changes on top of variations you have already set. | Option | Use it when | | ------------------------------- | ----------------------------------------------------------- | | **Content Editing Changes** | You want to edit text on the variation product page as well | | **Custom Code** | You want JavaScript or CSS applied per variation | | **Click-Through Rate Tracking** | You want clicks on specific buttons or links tracked | Adding these makes the experiment multivariate. A win then tells you the combination worked, not which part did. ### Step 7: Review Experiment Check the setup, then select **Create Preview Experiment**. ### The pre-launch review Product experiments have a review before launch with five stages, and the middle three exist because duplicated products need work that Shopify does not do for you. | Stage | What it covers | | ----------------------- | ----------------------------------------------------- | | **Overview** | The experiment setup | | **Theme Extension** | Confirming the extension is enabled | | **Product Reviews** | Copying reviews from your originals to the duplicates | | **Sales Configuration** | Inventory, SKUs, apps, and sales channels | | **Elevate Review** | Final confirmation, then **Launch Experiment** | #### Product Reviews Duplicated products start with no reviews. A product with none against one with hundreds is not a fair test, and the gap will decide your experiment for you. Elevate A/B Testing generates a message listing every original and its duplicate by product ID. **Copy to clipboard** and send it to your review app's support team, who will transfer the reviews for you. Do this before launching, not after. Reviews arriving mid experiment change the thing being tested partway through. #### Sales Configuration Two checks. **Other Apps & Inventory.** Make sure inventory and fulfilment are consistent across variations. If you use apps that modify product listings, connect them to your test variations. Update SKUs or other identifiers so every variation is recognised as the same product by your other systems. **Sales Channels.** Limit duplicated test products to the online store. On the product page in Shopify, uncheck other channels such as Facebook Shopping and Instagram Shopping, so your test duplicates do not appear as separate listings in those catalogues. ### Next steps * [Preview and QA a Product Experiment](https://claude.ai/elevate-helpcenter/test/product/preview-and-qa) * [Troubleshooting Product Experiments](https://claude.ai/elevate-helpcenter/test/product/troubleshooting)