Evergage Recommendations brings advanced, per-user product and content suggestions using complex algorithms and a deep understanding of user behavior. The solution offers configurable "recipes" that can be used to boost content or product discovery on your site. Recipes consist of ingredients, exclusions, boosters, and variations which can be constructed in a variety of combinations to serve up the right content or products based on the individual visitor's behavior and affinities on your site. These recipes are then queried against the proprietary Evergage Recommendations engine and the individual query results are presented to each visitor as his or her personalized recommendations.



This article provides an overview of the process for adding variations at the recipe-level.  Also, please refer to related articles for information on creating recipes, and adding ingredientsexclusions and boosters.





You must have Evergage Recommend integrated with the dataset in which you are configuring the recommendations. Please contact your Customer Success representative to complete your Recommend integration.


Variations

With Variations, you can modify the returned results such that end users will see a larger breadth of items. For example, you can configure a variation force a recipe to show at most two items from the same primary category. This means if a visitor was very affine to the the "hats" category, instead of seeing just hats, they would see only the two most relevant hats and a variety of other items from different categories.

  1. Create a new recipe or edit an existing recipe
  2. On the Variations tab, click Add a variation to select the exclusion from the drop down; scroll to see all options
  3. Adjust configurations for the variation as needed
  4. Click  to delete a variation

 

Variation Descriptions and Configurations

VariationDescriptionConfigurations
DimensionalVaried such that a limit is set on the number of items with the same primary tags or categories which can be returned in recommendations. The primary category or tag is taken to be the first one in the list on products, articles or blogs. This exclusion can only be used with products, articles and blogs.
  • Item Type - Categories or a tag type. Choose which dimension to vary.
  • Item Limit - The maximum number of items that share the primary category or tag. This value can be at most one half the number of maximum recommendations returned in the dataset.
RandomizedVaried such that the results are randomized for each user. All other recipe rules still apply, and boosted recommendations will generally still show up first.
  • Randomization Time Period - The time period in which the same user would get consistent recs. Available options are:
    • One Second
    • One Minute
    • One Hour
    • One Day
    • One Week
    • One Month