Every user interaction within a site or app signals the user’s interests, preferences, and intent, including adding to a cart, filtering a list, viewing an item, and making a purchase. Products users interact with, particularly their attributes (color, brand, style) and attribute values (red, Nike, t-shirt) teach us about their preferences.
Combining product attributes with interaction types, while taking into account the recency of the interaction, helps you identify user preferences overall and in real time.
Our affinity-based audience condition provides you with the tools to build on the user’s affinity and create highly-personalized experiences based on their personal preferences.
After these audiences are created, they can be used for targeting and exporting the audiences to third-party platforms.
Building a user affinity profile
At the user level, for each type of interaction, we collect and aggregate all the product attribute values that the user interacted with. We then calculate an affinity score for each attribute value. The score sums up the total interactions for each value, favoring recent interactions and high-intent interaction types by giving more weight to real-time activities, purchases, and adding to the cart.
Individual user scores are normalized, using the entire site’s traffic as a baseline. That is, if a particular attribute value usually receives high scores (for example, bread is a popular item at the grocery store), it will not skew results.
Creating audiences using the affinity condition
When defining an affinity-based audience, begin by selecting the desired attribute (color, category, brand, gender) and its respective value (red, dresses, Donna Karan, women). You can define the level of user affinity from highest to within the top 10 to determine how inclusive or exclusive you want your segment to be.
For example, when you want to promote content based on top-level categories, such as a homepage banner for either women, men, or kids, set the affinity level to highest and create an experience for each. Most often, users will have a distinguished preference for only one of these categories.
However, when you have dozens of values (such as is the case with brands) but only a handful of available promotions, widen your affinity level beyond highest. You might not "hit" every user’s favorite brand, but you will still be within their preferences and the segment size will be significant.
Learn more about audience creation.
Notes and limitations
- Affinity is based on your product feed. Only attributes with discrete values (closed list) can be used for the affinity condition.
- By default, we provide the category and its values as an out-of-the-box attribute. Additional attributes must be manually added in Experience OS for each individual site.
- Affinity data is generated after the user has viewed at least one product or category page, or after any other product interaction, like Add to Cart or Purchase.
- Affinity is available for e-commerce sites. If you are interested in enabling this feature, contact your Customer Success Manager.