Skip to main content


Optimizely Knowledge Base

Build Experimentation-in-Personalization campaigns

  • A/B test the personalized experiences you deliver to customers
  • Create two versions of an experience in a campaign, and compare how well they perform
  • Test and personalize together

In Optimizely Web, the Experimentation in Personalization feature helps you test and personalize experiences at the same time.

Personalization allows you to deliver tailored experiences to different audiences. But sometimes you're not sure what type of experience will make the biggest impact for a certain group of visitors. Use Experimentation in Personalization to test multiple variations for a single audience in your personalization campaign.

Visitors who meet the audience conditions for a certain experience will be randomly bucketed into the different variations for that experience.

Imagine that you're personalizing the messaging on your homepage for two audiences: "Luxury Travelers" and "5x Shoppers." You're pretty sure what you'd like to show visitors who qualify as "5x Shoppers." But research shows that your "Luxury Travelers" visitors prize both luxury goods and a sense of adventure.

So, you create two different variations for that audience in your personalization campaign and test them.

"5x Shoppers" will see the "5x Shoppers" experience. "Luxury Travelers" will be randomly bucketed between the luxury goods and the adventure variations.

The Results page will show which variation performed better, and how well each audience in the campaign performed overall. 

Running an A/B test in a personalization campaign allows you to deliver value, while honing in on the optimal site experience for important groups of visitors -- at the same time.

This feature is available to accounts using both Optimizely Web Personalization and Optimizely Web Experimentation.

Add a variation to a personalized experience

Start by creating a new Personalization campaign. When you click Create Campaign, you'll enter the Manage Campaign dashboard.


Each audience that is attached to the campaign, such as "Luxury Travelers," will contain a single variation and a holdback. To experiment with multiple variations on a personalized experience for a single audience, add more variations.

  1. Add a new variation to an audience by selecting Add Variation.

  2. Give your new variation a descriptive name. Then, click Save.

    Great! You've now added a new variation for that audience. Next, you'll edit that variation.

  3. To modify a variation, click Edit.


    Use the Optimizely Editor to make changes to that variation.

Once all experiences and variations look and work the way you'd like, hit Publish to start delivering them to visitors.

Visitors who meet the audience conditions will be randomly bucketed into the variations. You'll be able to compare how visitors responded to the different variations and track results for that audience as a whole on the Results page.

Traffic allocation

Optimizely automatically splits traffic evenly between variations.


Visitors who qualify for the audience will be randomly bucketed into one of the variations, or the holdback.

Here's how to change your traffic allocation:

  1. Click the ... icon for the experience card and select Settings.

  2. Click into the traffic distribution to change it. Then, click Save Experience.

Total traffic distribution must equal 100%. If it adds up to more or less, you'll see a warning message that prompts you to adjust your distribution.

Change the holdback

The holdback in a campaign helps you track how well the personalized experiences you deliver to each audience perform, compared to a generic experience.

Optimizely sets one holdback per campaign. So, no matter how many experiences and variations you add to a campaign, the holdback will stay the same. 

Here's how to change your holdback.

How many variations can be added?

There's currently no limit to the number of variations you can add for each audience in a Personalization campaign. You can also prioritize audiences so that if a single visitor qualifies for multiple audiences in Personalization, they'll see only one experience by default.