Create test optimizations

Run experiments to identify high-performing ideas with statistical confidence.

Test optimizations let you run controlled experiments to compare multiple variations of a page or experience. These optimizations are comparable to standard A/B/n tests. You can define how traffic is split between versions, track performance, and identify a winning variation once results reach statistical significance (stat. sig.).

How test optimizations work

These optimizations include a required No Change variation that serves as the control group.

Traffic is split evenly by default, but you can manually adjust the distribution if needed. Once launched, visitors are randomly assigned to one of the variations, including No Change. Test optimizations run for 100% of visitors on the page(s) they’re assigned to — meaning every visitor sees one of the test's variations.

To reliably reach statistical significance, all variables must remain unchanged — which means the test must run from start to finish without edits. This keeps results stable and directly comparable to the control group.

To account for this required stability while still allowing you to make edits, we use intervals. You can stop the test to make changes (e.g., fix a typo, add a variation, update the goal, or apply an audience), then relaunch the test — this starts a new interval and resets the time needed to reach stat. sig.

Once stat. sig. is reached, the top-performing variation is declared the “winner.” You can bake in the winner so all visitors see that variation — or convert the test to AI Optimize to keep multiple strong-performing variations active.

Create a test optimization

Open your Optimize site in Webflow, then:

1. Click Optimizations in the Navigation panel
2. Click New optimization
3. Click Use the visual editor
4. Choose Test, then click Continue
5. Choose a page definition
6. Click Select page & continue
7. Select an element on the page and build the variation
8. Rename the variation in the Variations panel
9. Click Save
10. Click Add variation in the Variations panel to build another variation
11. Rename the optimization by clicking the current name at the top
12. Click Done when you’re finished 

Configure your optimization

The Goals tab and Configuration tab can be accessed at the top of the visual editor, or from the optimization results page after you click Actions > Edit Configuration.

Goals tab

An "AutoGoal" is created by default, which tracks the forward engagement. You can create your own goals that more closely align to the optimization's intent to define what success looks like. Learn more about goals.

Configuration tab > Overview

• Pages — you can add, remove, or edit page definitions to define where the optimization runs
• Audiences — you can apply rules-based or code-based audiences to include or exclude certain visitors
• Traffic allocation — you can edit the traffic allocation from an even split to an uneven split — the values must sum to 100%

Configuration tab > Advanced

If you've added custom code, this page shows code selectors or preconditions you've defined.

Preview and launch

Make sure to preview each variation and confirm everything looks and behaves as expected before launching the optimization.

• Preview — learn how to preview variations
• Launch — click Start Optimization at the top of the experience details page

Manage how your test ends

You can stop the test at any time to make edits or if you just want to end the test to explore other ideas. 

When the test reaches stat. sig., you’ll be notified of the winner. If one or more variations outperform No Change, you can:

• Bake in the winner — lock in the top performer using a personalize optimization
• Convert to AI Optimize — let the system dynamically allocate traffic to multiple strong-performing variations