User profiles

A user profile is a representation of a user based on their behavior on your website or app. It’s used in conjunction with Algolia’s NeuralSearch and keyword search to personalize results.

How user profiles are built

Sending accurate events is the first step in building relevant user profiles. Algolia requires events that indicate how users are interacting with your app. For example, an online bookstore could track events such as a user clicking on a specific book in a listing, viewing genre and author on the book’s details page, adding the book to their cart, and buying the book.

Once events have been implemented and sent, Advanced Personalization applies various algorithms to your data (index and events). It identifies patterns and relationships within your data to predict user affinities. These affinities are used to create a user profile, which is continually updated with new information.

Schema

A user profile is represented in JSON format and includes a unique identifier for the user and their preferences (affinities).

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
{
  "userID": "user-1",
  "lastUpdatedAt": "2024-05-21T01:15:35Z",
  "type": "basic",
  "affinities": [
    {
      "name": "categories",
      "value": "Software",
      "indices": ["storefront"],
      "score": 18,
      "lastUpdatedAt": "2024-05-20T23:14:26.078071Z"
    },
    {
      "name": "color",
      "value": "Blue",
      "indices": ["storefront"],
      "score": 14,
      "lastUpdatedAt": "2024-05-21T01:15:08.590763Z"
    },
    {
      "name": "brand",
      "value": "Microsoft",
      "indices": ["storefront"],
      "score": 6,
      "lastUpdatedAt": "2024-05-21T01:15:35.465487Z"
    }
  ]
}

Properties

Affinity

Further reading