Why do A/B testing?

Use A/B testing to validate changes and improve relevance.

Relevance is the difference between having a search, and having a great search. Most Algolia API methods and settings, as well as Algolia dashboard features, are devoted to giving you direct control over the relevance of your results.

Once your search index is live and returning results, your next job is to make those results useful. That means evaluating relevance: whether what shows up first is what users expect. To get better results, you can:

• Tweak Algolia’s default settings (such as typo tolerance and language-handling).
• Set up your data and attributes differently.
• Add synonyms, rules and filters.
• Re-order your results with custom ranking and alternative sorting strategies.

Relevance tuning can get quite advanced, but it’s necessary if you want a great search experience.

But how do you know what to change? A/B testing is one way to compare two configurations under the same conditions. These structured, controlled tests help you validate changes and draw conclusions based on data.