Identity Clusters

Output format describing a computed identity cluster and its associated data

What are Identity Clusters?

An identity cluster is a JSON output format describing a cluster of IDs and their associated data. Clusters are automatically computed by the Optable DCN based on ID profile data loaded via the configured sources. Clusters are a supported data format when exporting an audience to a destination.

Clusters may contain traits which are resolved from the ID profiles that make up the cluster. An identity cluster represents an individual person and their associated data. If you've loaded household scoped ID profiles, household data will be resolved separately and then associated with one or more identity clusters.

Identity Cluster Structure

An identity cluster is a JSON object containing the following attributes:

• The "ids" attribute value is an array of type-prefixed ID strings that make up the cluster. These are identifiers associated with the person that the identity cluster represents.

• If there are computed person-scoped traits associated with the cluster, the "traits" attribute value will be an array of objects, each containing "key" and "value" attributes. For each trait:

  • The "key" attribute is a string equal to the name of the trait.

  • The "value" attribute is a string equal to the value of the trait.

• If there are household identifiers associated with the identity cluster, they will appear as "ids" sub-objects within a top-level "household" attribute.

• If household identifiers associated with the identity cluster also contain traits (e.g. "Postal Code" , the trait will be resolved at the Person level.

• The "sources" attribute value is an array of source IDs. The sources associated with a cluster are those that have loaded ID profiles that are part of the cluster.

Example Cluster JSON

{
  "ids": [
    "v:2zsD56opk8G6vL9aMxyUzN",
    "v:3qBLvZUcJjC6nXTduW5r8A",
    "c:bd98c45acbb724082",
    "a:114a9d03-2f95-465b-8b04-cd3883396599",
    "v:dRZJxwtcVuz63nEhS8ygm2",
    "c1:84e541",
    "e:2a2067ebbd99c8a07b85dd3b920549d4fe94ee15e00739d528514054cd78a6d3"
  ],
  "traits": [
    {
      "key": "profession",
      "value": "teacher"
    },
    {
      "key": "gender",
      "value": "F"
    },
    {
      "key": "ageGroup",
      "value": "25-34"
    }
  ],
  "sources": [
    5,
    3,
    36
  ]
}

Here is another example of an identity cluster containing associated household data:

{
  "ids": [
    "a:114a9d03-2f95-465b-8b04-cd3883396599",
    "v:1zkRnRv7fAgAmg9XDBPNit",
    "c1:84e541",
    "e:2a2067ebbd99c8a07b85dd3b920549d4fe94ee15e00739d528514054cd78a6d3"
  ],
  "traits": [
    {
      "key": "gender",
      "value": "F"
    },
    {
      "key": "ageGroup",
      "value": "25-34"
    },
    {
      "key": "Postal Code",
      "value": "H0H 0H0"
    }  
  ],
  "household": {
    "ids": [
      "r:0b179df0-6cd5-49f1-be21-425d002e0d22",
      "s:e0ef86a8-6ebf-4c9d-9127-e69407fe748d",
      "i4:192.168.3.12"
    ]
  },
  "sources": [
    5,
    3,
    36
  ]
}

Note that person and household scoped ID profiles are resolved into identity clusters separately as documented here.

The same household data may be associated with more than one identity cluster.

It's also possible to export identity clusters that contain only household data and no person-level ids and traits. This will be the case when importing household scoped ID profiles and not connecting them to any person scoped data, then subsequently selecting such household clusters in a query audience.

Cluster Computation

Identity clusters are automatically computed by your Optable DCN based on all unexpired ID profile data loaded from all of your sources. ID profiles that specify neighor IDs will end up in the same cluster as their neighbors.

An ID profile is a member of a single identity cluster, along with all of its neighbors.

Traits Resolution

The traits that are included in a computed cluster are the traits of all of the ID profiles that make up the cluster. When there are multiple values for the same trait key the last updated value is selected for the cluster.

Household Data

Household data is resolved separately to person data. The details are explained in the identity resolution concept page.

Exporting Clusters

To export clusters computed by your DCN, first create an audience and a destination, and then export your audience.