> ## Knowledge Base Index
> Fetch the complete knowledge base index at: https://help.upsellplus.com/sitemap.xml
> Use this file to discover available pages before exploring further.
> Pure-Markdown content can be obtained by appending a '.md' suffix to the content URLs listed in the sitemap (without the trailing slash).

# Experiments: How to run an A/B test

### Overview

A/B testing your upsell offers has never been easier! **Experiments** let you continuously improve your on-site revenue with our cutting-edge A/B testing technology. AI will generate variants, pick a winner, and start the next test. No code or intervention needed.

# What can you A/B test with Experiments?

You can configure different variants of an offer to control the following:

* **Banner text: **Add a new callout banner
* **Call to Action: **Edit the wording of the Add to Cart button
* **Layout: **Configure a new design of your offer
* **CTA Styles: **Change the look of the Add to Cart button
* **Discount** - Override the discount per product. Each product can have its own type and value.
* **Products** - Upsell different products or variants

### How to run an A/B test?

It's really simple!
* Create one upsell offer 
* Once created, click on THE "A/B Test this offer" button (This would be the **Step 1**)
![](https://storage.crisp.chat/users/helpdesk/website/-/a/b/f/a/abfa0cd47f75e80/screenshot-2026-04-14-at-94649_14p5f3v.png =600x368)
* Edit the internal information (**Experiment Name**, your **Hypothesis, **choose the **Primary Metrics **you'll be testing the offer with).
![](https://storage.crisp.chat/users/helpdesk/website/-/a/b/f/a/abfa0cd47f75e80/screenshot-2026-04-14-at-95236_1jf6b6x.png =600x356)

* (**Step 2**) Under your **Variant B, **choose what you want to A/B test and edit accordingly
* If you want different variants to test under the same offer, click on "+ Add variant". You can add up to 5 variants (including control-the original version you created in the point 1.)
* (**Step 3**) Set the **Traffic & Duration. **The traffic allocation must equal 100%. You can play with how you want to distribute the priority
![](https://storage.crisp.chat/users/helpdesk/website/-/a/b/f/a/abfa0cd47f75e80/screenshot-2026-04-14-at-95537_1usl5ty.png =600x186)
* Schedule or Save and Start immediately your Experiment