You are here:
Boosters in Einstein Recipes
Include boosters in Einstein recipes to boost items matching a customer’s affinity in any recommendations Personalization presents.
Booster Parameters and Affinity Scores
When you include a booster in an Einstein Recipe, Personalization incorporates the customer's affinity score to boost items matching that affinity in the recommendations it presents. The affinity score is based on a customer’s interactions with your site, such as item views, items added to the cart, and items purchased. For example, if a customer has a preference for a particular category of item, Personalization shows items from that category first.
Each booster has three parameters that determine the affinity score.
| Parameter | Description |
|---|---|
| Weight | Affects how much to weight the selected booster in the recommendation. The slider has five positions (1–5) that act as multipliers when factoring the affinity score. |
| Lookback | The number of days to look back at a customer's history for the booster. For example, if a customer bought a product three months ago, and the lookback period is 30 days, Personalization doesn’t include that purchase when factoring the affinity score. |
| Threshold | Determines the point at which the customer has an affinity for the booster. The slider has six positions (0–5) that act as multipliers when factoring the affinity score. If the threshold is zero, any of the items the user has interacted with during the lookback period are available to boost. |
Boosters
| Booster | Description | Available When Recommending |
|---|---|---|
Authors |
Personalization assigns a higher affinity score to the authors the customer shows an interest in. |
Articles, blog posts |
Brand |
Personalization assigns a higher affinity score to the brands the customer shows an interest in. |
Products, articles, blog posts |
Category |
Personalization assigns a higher affinity score to the categories the customer shows an interest in. This booster includes an option to boost a category hierarchy with decreasing magnitude. For example, if a category hierarchy is Women's|Dresses|Halter, Personalization also boosts items in the Women's|Dresses and Women's categories, but less than items in the full category. |
Products, articles, blog posts |
Class |
Personalization assigns a higher affinity score to the classes the customer shows an interest in. |
Products |
Color |
Personalization assigns a higher affinity score to the colors the customer shows an interest in. |
Products |
Combination |
Combination boosters allow more specific items to return as recommendations than when boosting by item type. These boosters determine which items the customer has an affinity for, and only boosts items with the same combination of categories and tags. You can combine up to three item types. If an item doesn’t have a value for a specific item type, Personalization doesn’t boost it. If a recipe has a category booster and a color booster, a customer who buys a red dress has an affinity for the color “red” and the category “dress.” So boosted recommendations are likely to be red, or dresses, but not necessarily red dresses. Because combination boosters only boost items with the specific combination of tags and categories, Personalization doesn’t boost red shoes and black dresses. As a result, those items are less likely to return as recommendations. |
Products, articles, blog posts |
Content class |
Personalization assigns a higher affinity score to the content classes the customer shows an interest in. |
Articles, blog posts |
Department |
Personalization assigns a higher affinity score to the departments the customer shows an interest in. |
Products, articles, blog posts |
Gender |
Personalization assigns a higher affinity score to the genders the customer shows an interest in. |
Products |
Keyword |
Personalization assigns a higher affinity score to the keywords the customer shows an interest in. |
Products, articles, blog posts |
Product list |
Personalization assigns a higher affinity score to product lists the customer shows an interest in. |
Products, articles, blog posts |
Related information |
This booster boosts a specific item type, such as category, based on the selected anchor item. When boosting a category, there’s also an option to boost through a category hierarchy with decreasing magnitude (see the category booster). |
Products, articles, blog posts |
Store |
Personalization assigns a higher affinity score to the stores the customer shows an interest in. |
Products |
Style |
Personalization assigns a higher affinity score to the styles the customer shows an interest in. |
Products |
