Merchandise boosting is a way of tuning the recommendation algorithms, where you can boost products with high ratings, a better profit margin or with a high inventory to move products, or low inventory to sell out. The merchandise boosting factor is an optional input parameter in the Control panel in each module.
Merchandise boosting can be applied on most Raptor modules, except “GetCookieHistory” because this module just provides you with the last seen products and no boosting is relevant.
For boosting products with a high profit margin, the profit margin must be included in the product feed, which will be mapped as a custom field in Raptor. Additionally, you need the full data output from the product feed by typing “true” in “full info”.
In the parameter “Select” in the box “Output formatting and filtering”, you type in RecommendedId,Priority,CustomfieldXX, this is to get a more simple output with only the information from the product feed you need.
You decide by which boosting factor you want to boost - between 0 and 20. 0 means no boosting, 10 means medium boosting and relevance of the product recommendation still have a high priority, and 20 means max boosting and the products will have top priority in the recommendations, where relevance become less important.
As you boost with your own business knowledge, be aware that it might have some negative consequences. An example could be that you might include products only because of their high profit margin and consequently exclude the most relevant products in the specific product category. This might have a negative impact on your conversion.
We therefore always recommend that you check the output manually, to make sure that the recommendations are still relevant to the product category.
If you want to include a new default merchandise boosting weight, simply just save as default. And if you want to have a different boosting weight on specific categories, sent it as a query parameter in the API call. This is an option if you want to us different weights for different categories.