Custom affinity attributes allow you to define affinities for a set of values not matching our pre-defined affinities. For example;
A fashion brand may be interested if a visitor has an affinity to a specific colour
A jewellery brand may be interested if a visitor has an affinity to a specific metal
A holiday brand be interested if a visitor has an affinity to a specific star rating of hotel
Once these are set-up, they can then be utilised within our Custom Segment Builder.
Custom affinity attributes are denoted by their 'slot' number: attribute_1 to attribute_9. You're assigning a value to a slot as a dimensional attribute.
Set-Up
There are two ways to set up custom affinities;
They can be mapped to a GTM variable directly in your Made with Intent Pageview in the Custom Affinities section as per the below:
They can be sent directly from your codebase, or via a custom GTM tag. More information on this can be found here
Using Custom Affinities
Once set-up, custom affinity attributes can then be used within the Segment Builder and added as a condition.
You are able to rename your custom affinities via the Custom Attributes section within the platform, that can found in the Navigation under 'Segments'.
Simply click the attribute number according to your set-up, enter your desired name and select 'Save'.
These will now pull through into the segment builder with your specified name.
Within here you are also able to see:
A sample of the values populating the attribute
Please note that sample values are taken from sampled data every 24 hours and so won't be visible in this view immediatelyA history of changes (for example if you have renamed the attribute from a previous name
Agentic Context
You are also able to provide each affinity with Agentic Context, helping the agent understand how to interpret and use them when building Agentic Campaigns.
Use it to:
Tell the agent when an affinity is relevant to personalisation (e.g. when to act on a particular affinity)
Define how values should be grouped or interpreted (e.g. bucketing visit frequency into new, returning, and loyal segments)
Give the agent a head start on customers with niche or domain-specific data points
Example: Clothing Size
You might specify that visitors with less common sizes often encounter limited stock and respond well to early access or low stock urgency messaging, while those in popular sizes benefit more from new arrival or trend-led content. This lets the agent personalise around something that directly affects what a visitor can actually buy, rather than just their browsing behaviour.