What has been their challenge?
Lists on news platforms are frequently generic and often show the same articles, even if the visitor has already read them. Mediahuis already included a Most Read list that recommended popular articles, excluding ones that the visitor had already seen. But they wondered if it was possible to use personalization to further increase visitor engagement.
How did we approach this case?
We tested more complex personalization using different approaches to discover what worked best for Mediahuis. The approaches included: · Random selection · Most read articles · 1:1 personalization that considers account user behavior, filtering capabilities, and diversification · Full-suite personalization which uses impression discounting and several algorithms with different weights to discover the most effective.