What Is Personalization?

Personalization strengthens interactive search: it adds a personal layer to the relevance strategy. Adding personalized preferences to the search experience makes results more engaging for individual users.

Advantages of Personalization

Algolia’s out-of-the-box relevance strategy treats every user the same way, while Personalization takes into account a user’s individual tastes. Textual relevance and business metrics are the first steps to great relevance, but Personalization allows you to engage your users on the next level.

Queries mean different things for different people. For example, a user who searches for “harry”, and has a preference for children’s literature, likely wants to see “Harry Potter” results appear on the first page. Politically minded users may be more interested in seeing “Harry Truman” in their results. Personalization incorporates a user’s past behavior when determining which results are most relevant.

Better relevance minimizes the user’s effort to find what they’re actually looking for. It also encourages users to stay on your site or app longer by exposing them to more options they’re likely to find appealing. Thanks to giants like Google and Amazon, most users expect a digital experience where their individual preferences are used to give them with tailored results. Algolia’s Personalization feature brings this capability to your business.

Personalization is available on Premium plans. Please reach out to the support team if you have any problems accessing Personalization on your dashboard.

How does Personalization fit into Algolia’s relevance strategy?

Effective relevance has two main goals:

• Enabling your end users to find results that match their expectations.
• Providing results that align with your business needs.

You can address these twin goals by fine-tuning the first three layers of Algolia’s relevance strategy:

1. Textual relevance - intelligent and robust textual matching that includes typo tolerance, synonyms, natural language processing, and a host of other capabilities and settings.
2. Business relevance - business-centric relevance that ranks results according to business metrics via custom ranking.
3. Merchandising - boosting and burying specific results or categories using Rules.

You can also enable AI Re-Ranking to boost trending results and categories based on the collective search data of all your users.

All these contribute, in different ways, to producing relevant results that apply equally to all your users. Personalization acts in concert with these to individualize search results. It doesn’t replace Algolia’s ranking algorithm—it refines it by changing the order of pre-sorted results to promote more relevant results to the individual user. This, in turn, can increase user engagement and drive more conversions.

Personalization kicks in after the engine computes your results’ textual relevance and applies your business relevance. If you use Rules, the engine applies them after the Personalization.

If you’ve enabled both Personalization and AI Re-Ranking, the engine applies Personalization for any users with enough data to personalize the search. The engine applies Re-Ranking for users it doesn’t have enough data to personalize a search for, such as first-time users. The engine doesn’t apply both Personalization and AI Re-Ranking concurrently.

In summary, the engine applies the relevance strategies in this order:

• Textual relevance (through the textual ranking criteria)
• Business relevance (through custom ranking)
• User-based preferences (through either Personalization or AI Re-Ranking)
• Merchandising (through Rules)

The only exception is when you’ve set your Personalization impact to 100. In that case, the engine prioritizes Personalization over business relevance. Then the engine applies the relevance strategies in this order:

• Textual relevance (through the textual ranking criteria)
• User-based preferences (through Personalization)
• Business relevance (through custom ranking)
• User-based preferences (through AI Re-Ranking, only if there isn’t enough data to personalize results for a particular user)
• Merchandising (through Rules)

Before you enable Personalization, you should simulate your Personalization strategy and A/B test the effects of Personalization first.