What is the difference between a search-based split and a cookie-based split in A/B testing?

• A. A search-based split uses cookies to track users; a cookie-based split does not
• B. Both splits allocate users randomly without considering campaigns
• C. A search-based split sends all traffic to the control, while cookie-based split sends traffic to the experiment
• D. A search-based split evenly distributes users between existing and experimental campaigns, while a cookie-based split directs returning users back to the same campaign for more accurate results

Answer: (D)

In A/B testing, how you assign users to variants shapes the reliability of the results. A cookie-based split uses a small piece of data stored in the user’s browser to remember which variant they saw, so returning visitors are shown the same version again and again. That persistence reduces cross-session carryover and makes it easier to attribute differences to the change being tested.

A search-based split, by contrast, makes the assignment based on the current request or search context and tends to divide traffic evenly between the existing and experimental campaigns without relying on a persistent identifier. Because the same user might be shown different variants on different visits, the data can be more variable, but traffic is balanced across variants in the moment.

So the described approach—evenly distributing users across variants for the search-based method, while steering returning users back to the same campaign with a cookie-based method—best captures the distinction between the two.