FAQs

Frequently Asked Questions

Segments in Recommend

What are the uses for Segmentation?

You can build segment-based targeting rules within campaigns for personalized promotions.

How is the value of segments measured?

You can create a Multivariate Test (MVT) to learn if merchandising rules, strategies, or promotions increase the value of recommendations. For more information about setting up MVT, refer the MVT documentation.

What are best practice implementation tactics for Recommend Segmentation?

Customers that belong to small segments: It is suggested that merchants create segments that include a large demographic. Strategies don’t work well for customers belonging to small segments, because of less behavioral data to adapt recommendations. However, this is not the case for RichPromo. Tailored promotions can be shown to small segments.

No parent/child associations of segments: If a customer belongs to top-level segments and subsegments, each segment must be supplied to Algonomy. For example, if there are two top-level segments, such as medical and government, and within government there are subsegments, such as military and healthcare, and the customer belongs to military (subsegment) and government (top-level segment), the Algonomy system needs to be informed that the customer belongs to both. Customers are not automatically associated with the top-level segment (in this example, government).

Site Configurations

When we set the option “don't recommend previously purchased”, on the Site Configurations page, how many products, orders, or what length of time does it take into consideration? In other words, what is the limit for this option?

There are limits to the size of both cookies and URLs. For cookies, there is a limit of 20 items. On URLs, the limit is about 2000+ characters.

There is also a combined limit on the size of the HTTP request (URL, cookies, other headers, or any request body) of about 6000 characters.