10. User Data Import¶
By default, the recommendation engine has very little information about the user visiting the website, especially, about the first-time visitors. The only information provided in the event tracking and recommendation request is the customer ID of the website owner and an external user ID (usually an anonymous session ID). This external user ID is again anonymized in the recommendation engine by a hashing algorithm.
Sometimes, it is helpful to store additional attributes for every user, for example age or home city. It is possible to retrieve the stored information based on the external user ID, but due to the hashing step mentioned above, it is not possible to get the external user ID within the user's attribute set (unless the external ID is not explicitly stored as an attribute of course).
For more information about importing and utilizing user attributes, and about the import format, see User API in the developer documentation.
For a detailed explanation of how user attributes are used in the recommendation process, see Recommendation API in the developer documentation.
For an example of how user attributes can be used in the recommendation engine, see Boost-Filters.