Site Search personalisation

Raptor’s Site Search helps customers quickly find products through your site’s search bar.

While built on top of Typesense, Site Search offers unique Raptor features that make search results more relevant for your customers, which include:

• RaptorScore: highlighting popular or trending products.
• Merchandising campaigns: manually promoting specific product selections.

From general to tailored product recommen-dations

This update introduces Site Search personalisation, tailoring product rankings to each user’s preferences to increase engagement and conversion.

When personalisation is enabled, Site Search results are ranked using the following logic:

1. Text relevance: products matching the search keywords are selected first.
2. Personalisation: a similarity score is calculated between the user profile and each product (vector similarity):

• Products above the personalisation threshold are promoted and ranked by their vector similarity score.
• Remaining products are ordered by RaptorScore, reflecting global popularity.

Let’s look at an example where a user searches for the word shirt, with personal preferences that include ‘Black clothing‘ and ‘Brand A‘:

Search personalisation is available with Raptor’s Typesense-based solution or any search engine that supports vector search and similarity. To learn more, check out this article or get in touch via [email protected].