Recap
Testing
Testing allows you to compare different versions of page elements (e.g., buttons, headers, copy, etc.) to see what performs best. This is essentially an A/B testing method.
E.g., Determine which heading copy leads to more conversions.

When to use testing:
- When you want to test which of two or more variations results in more conversions
- When you want to directly compare a smaller number of variations (often just two)
- When you do not want to specify an audience or visitor segment for each variation; instead, you want to distribute traffic at random, by percentage or with AI
Personalization
Personalization allows you to specify which audiences, or visitor segments, should receive which variations, allowing you to tailor the experience for those who meet your set criteria.
E.g., Present Version A to those who access my site via organic search and Version B to those who access it via direct navigation.

When to use personalization:
- When you have a hypothesis you want to test about a certain audience; e.g., mobile visitors will prefer a shorter heading
- For a consistent customer experience; e.g., users who see a certain promo in their email will see the same promo when they visit your site
- To customize entire pages for specific audiences; e.g., visitors from a certain industry will see messaging and CTAs most relevant to them (when industry data is available)
AI-optimized
You can enhance testing and personalization with AI-optimization. AI-optimize automatically displays and rapidly tests multiple variations and combinations of variations. It uses machine learning to show visitors the variations most likely to convert them based on how they engage with your site – linger on pages, click certain elements, etc. It continues to improve based on these interactions and data available about them.

When to use AI-optimized:
- When you want to test multiple ideas quickly without managing rules manually
- When you want to test combinations of variations, e.g., multiple headers, images, copy
- When you’re targeting users with diverse needs; e.g., several different versions of the same word or phrase could resonate with different audience segments
- When you want to limit bad ideas by automatically reducing the number of visitors who see less successful variations
Consider this: When might you use each of the optimization options above in your own work? Why?