Fallbacks generate product recommendations if the primary setting does not have enough data and therefore cannot generate the designated recommendation slot.

The most frequently occurrence is the inability to recommend recently browsed products or any browsing history based recommendation to a user who visited the website for the first time. In the described use-case, Browsing History can’t be displayed because data is inexistent for the given user. An example how this works for example on home page below.

Customer A has viewed five (5) or more products and is clearly a man interested in boots, so Nosto shows the customer his weighted browsing history, giving more weight on those items that have been viewed more recently or multiple times. This is the most powerful recommendation as it’s most personal.

 

Customer B has only viewed less than 4 products, but from a certain category so the recommendation fills up with inspirational items related to browsing history, in our example: Heels. This is a good option to leverage existing data and personalise the experience. This is in our example the first fallback.

Customer C is new to the store and will be shown key product ranges as best sellers illustrating the shops main offering: Clothing for women. This is the third fallback.

 

 

Another example is a product based cross-selling which has been set to display a minimum amount of 4 products. If for a given viewed product Nosto can only find less than 4 relevant matching products to recommend, the recommendation won’t be displayed.

In response of no results and to benefit from product recommendations regardless of the scenario, you can apply a fallback recommendation. This means that if no results as products can be recommended in a given context, the fallback recommendations takes over and displays product recommendations within the same recommendation.

For example, if Best Sellers recommendation type would have been set and applied to a fallback recommendation to a Browsing History, the most trending products of the store would have been displayed instead of hiding the slot, which is the behaviour by default.

This is a powerful way to leverage your traffic and to be able to serve different segments of users.

Learn how to get started here.

Can i fill out a partially filled Onsite Product Recommendation slot with partial fallbacks?

You can use the "Fill with fallbacks" - toggle to fill out partial Onsite Product Recommendations with the next available fallback. 

Example: Primary slot is strictly filtered, and only matches 2 items to fit the range of 1-5 products. If you have set up a fallback recommendation you can use "Fill with fallbacks" to essentially fill out the slot up to 5 items, keeping the 2 strictly filtered items as the first two. 

Did this answer your question?