Personalized recommendations based on reading history

The Personalized engine in Marfeel Recommender uses each reader’s browsing history to surface articles they are most likely to engage with, creating unique navigation experiences on your site.

Scenario description

Printed newspapers offered a clear and structured presentation of their daily content. Readers could easily gauge the full extent of an edition, deciding how much to read with the certainty of a finite number of pages.

Digital editions, by contrast, often rely on static homepages and recirculation modules, which rarely capture the full scope of what is published. Frequent readers may feel they have seen everything within minutes, while much of the content remains hidden beneath the surface.

Personalized recommendations bridge this gap effectively, increasing click-through rates by up to 100%. While ranking systems weigh multiple signals to determine relevance, the Personalized engine prioritizes a reader’s historical interests, leveraging past site navigation to surface the most engaging content.

Previously read articles are excluded from recommendations automatically, ensuring a dynamic and ever-evolving feed.

How to set it up

As a replacement of an existing module

1. Navigate to Experience Manager and click on New experience, filter by Recommender family and Inline format and select any. The experience edition screen will open up.
2. Use the dropdown for the preview url to navigate to the page where the module to be replaced is shown.
3. Click on the target icon to activate highlighting mode and select the module to be replaced. Use the breadcrumbs above to refine the selection if needed.
4. Once the proper element is selected, use the dropdown options at the bottom and click on “Generate both” (CSS selector and layout).
5. Once both have been generated, confirm they look good or tweak what is needed, and click on Confirm. Check the preview to make sure that everything looks as it should.
6. On the Content tab, click on the configured Recommender feed’s three-dot options, then Edit, and select “Personalized” as Engine.
7. Configure the rest of parameters (Time Window, number of articles) as needed.
8. On the Targeting tab, add the filter of “Visitor loyalty” as equal to “Loyal” or “Lover”. Less engaged users do not hold enough data to generate meaningful personalized experiences for them.
9. Add any other targeting required and publish. Before publishing, you can create the experience and test it on multiple devices by sharing the url generated when clicking “Open in browser” from the preview’s three-dot options. Different devices should show different content.

Why should personalized recommendations target only loyal visitors?

Less engaged users do not hold enough browsing data to generate meaningful personalized experiences. Targeting visitors with a loyalty level of “Loyal” or “Lover” ensures the engine has sufficient navigation history to surface relevant content.

How much can personalized recommendations increase click-through rates?

Personalized recommendations have proven highly effective, increasing click-through rates by up to 100% compared to static recirculation modules.

Are previously read articles shown again in personalized recommendations?

No. Previously read articles are automatically excluded from recommendations, ensuring a dynamic and ever-evolving feed for each reader.