How does the Learning Search Multiplier work?

How many days of click history impact the Popularity Score? 

The system default is set for 30 days, but this can be changed in the system parameters.

How many clicks before the Popularity Score begins to be applied?

Each click will have an impact for popularity.

How is the Popularity Score calculated?

• The boost is based on click tracking.
• Each click on a product is saving the absolute position of the item in search results (for instance, if a product is on the second page and it's position is the results was 2, and there are 12 items per page, the absolute position will be 14) and the number of search results from the current search.
• For a Popularity Score boost, only items with the number of results greater than 5 and the absolute position greater than 3 are taken into account.

If other global boost scores are in place, with the Popularity Score still work?

The Learning Search Multiplier is working with other boosts from the Boost/Bury rule. 

During indexing, learning tracking from the last 30 days is being taken with the following actions:

1. For each of the unique product IDs (UniqueID), all of the clicks are gathered and iterated through.
2. For each click, the partial boost is calculated (Min (clickTracking.Aboslute / clickTracking.NofResults, 1) ) and added to the global boost for the product (UniqueId).
3. Step 2 is repeated for all tracked unique product IDs (UniqueId).
4. The values of the global product boost are normalized for all products into a range of 1-10. The values are stored in the index.

During search, these values are applied by factors defined in the Dashboard.