Seamless experimentation: Test and personalize together

  • Updated
This topic describes how to:
  • Decide when to integrate testing and personalization
  • Validate your optimization ideas by A/B testing them
  • Launch optimized ideas as targeted experiences

As you know from A/B testing, there are always multiple ways to execute on a given idea. Testing helps you optimize designs for your visitors and measure the impact, so you can learn from the results and iterate on your optimization strategy.

In some cases, you will want to integrate your testing and personalization processes to validate and optimize ideas before publishing them as campaigns. In other cases, you will test site-wide changes separately from your personalization efforts.

Optimizely Web Experimentation also lets you A/B test variations within your personalization campaigns. Learn more about Experimentation in Personalization.

Although it might be wise to test every campaign before launch, Optimizely Personalization still makes it easy to measure results with statistical rigor. Personalization uses a holdback to measure the difference between the original experience and your personalized experiences, so you can track your campaign’s progress.

Integrate testing and personalization

The main principle behind integrating testing and personalization is: test multiple variations and launch the winners as personalization campaigns.

Every experience in your messaging matrix is a hypothesis: a testable statement proposing a cause-effect relationship between your site experience and visitors’ behaviors. Testing these hypotheses can help you optimize the experience. Any improvement you measure also validates your proposal and strengthens the business rationale behind your personalization strategy.

The difference between the original and your variations provides insight on how impactful your personalization campaign can be. It can also give you a sense of the potential lift your changes can achieve. If one variation is outperforming the others, you might ask—should I optimize all other designs to make them more compelling? Or is this type of experience particularly compelling for one type of audience? Use these insights to iterate in testing and to inform future campaign designs.

So, once you have mapped out whom you will target messaging and where you will deliver those experiences on your site, you have an opportunity to test that experience. In Optimizely Experimentation, create each experience as a variation and use audience targeting conditions to deliver it to a segment of visitors.

If you do not have the traffic or time to test each experience, consider testing to validate the general principle of your campaign. Create a single variation—or a handful of variations on the theme—to refine your concept before delivering it in Personalization.

Prioritize your campaign ideas in your Optimization Roadmap. Once you have tested and validated your ideas, you are ready to create and launch them as personalization campaigns.

Test and personalize in parallel

Programs also benefit by testing in parallel with personalization. Some changes to your site experience—such as fonts, color palette, or a single versus multi-page checkout process—are served to all visitors. For most companies, redesigns of the underlying site structure are not likely to have specific applications in personalization. Optimize these site-wide changes in Testing, then build targeted experiences on top of them in Personalization.

Learn more about personalization strategy in this article.