Consider rules-based personalization for website optimization

1. Decide whether it’s important to be able to show a different version of your website to different visitors.

A rules-based personalization approach makes it possible to serve up tailored experiences to each segment of your visitors. You are only limited by the number of segments you want to manage.

2. Separate your target audience into key segments by industry, company type, location, on-site behavior, or potential value to your business.

It makes most sense to use rules-based personalization when it’s inappropriate or brand-unsafe to show particular content or a specific experience to an audience. For example, if the site visitor is located in the New York area, then your site should show them New York-related content rather than Boston-related content. It would typically be a bad experience to show a New Yorker content about the Boston Red Sox.

3. Figure out how you’ll identify members of each of these segments when they visit your website.

To place each visitor into a segment, you may need third-party data or to infer segments based on on-site behavior.

4. Look at the elements on your website and identify good personalization candidates.

What needs to be customizable to serve up a personal experience to your visitors? Be aware that as you increase the elements you personalize, the rules you need can increase exponentially. This won’t just increase the amount of setup work, but also the maintenance requirements.

5. Consider whether you have the time, traffic, and resources to run A/B testing or multivariate testing to check your personalized variations.

Be aware that you’ll need to test within each audience to determine the right content for the rule. For example, you might test Headline 1 against Headline 2, using an A/B test limited to New York visitors. If Headline 2 wins, then your rule is: If the visitor is from New York, show them Headline 2. This is great because it is data-driven. However, this A/B test will take longer than a site-wide A/B test because it can only run on a subset of traffic - people in New York.

6. Calculate how often you’ll need to change your personalized messaging based on changes to your advertising, targeting, and promotions.

Rules-based personalization will not adjust to these changes in website visitor behavior. You will need to re-run audience-specific A/B tests to learn whether the existing rule still provides the best experience.