Prioritize your recommended scenarios from highest to lowest using waterfall predictions in Einstein Recommendations. In a waterfall prediction, Einstein goes down the list of scenarios and returns all available matching items until it reaches the total number of recommendations that you set. Set up a default site data scenario last in your hierarchy to ensure that Einstein fills the total number of recommendations.
If you choose not to enable waterfall predictions, Einstein returns recommendations only from the first algorithm that provides the total number of recommendations that you set. In this case, you could miss top-priority recommendations if the algorithm doesn’t meet the total number set.
You set the recommendation number to four items. Because your top priority scenario returns only one recommendation, Einstein moves to the next scenario on the list. It continues down the list until there’s a total of four recommended items.