Excluding OOS variants from self-learning score
By default in Smart Category Merchandising, we apply a self-learning score, based on a combination of product clicks & checkouts within a 30-day window, to configurable products based on the overall self-learning score of its associated variants, including out-of-stock variants.
However, Klevu can support the exclusion of out-of-stock variants from influencing a configurable product self-learning score. This will automatically deboost configurable products if one of its variants that has a high self-learning score goes out of stock (such as popular size variants .i.e Large, Medium, Small).
Let’s say:
- Configurable Product P1 has 2 variants V1 and V2
- Configurable Product P2 has 2 variants V3 and V4
- V1 and V2 have received a self-learning score of 9 and 10, respectively
- V3 and V4 have received a self-learning score of 5 and 10, respectively
Klevu calculates scores for each variant based on its individual performance on the website and aggregates a configurable product score.
With this, in the example shared above:
- Configurable Product P1 would have a total self-learning score of 19
- Configurable Product P2 would have a total self-learning score of 15
So how does excluding OOS variants from learning work?
Now let us assume that V2 is out of stock and you chose not to consider out-of-stock variants in self-learning scores:
- Configurable Product P1 and its in-stock variants (V1) has a total self-learning score of 9
- Configurable Product P2 and its in-stock variants (i.e. V3 and V4) has a total self-learning score of 15
- Since P1 now has a lower self-learning score than P2, due to V2 going out of stock, P2 would be positioned higher in the category page than P1.
To enable this feature, please send an email to Support@Klevu.com and include:
- Your storefront URL
- This support documentation URL and a message saying ‘I request to enable this feature’.