What are recommendation strategies?
When setting up a touchpoint, you are asked to select a recommendation strategy. In order for you to choose the right one, it is very important you know what every single one does at a basic level. There are a lot of them so you can also use this help section to get some inspiration for creating certain campaigns.
In Datatrics, you can find and edit the recommendation strategies in touchpoints in the content section. The default selected strategy used here is ‘Let the algorithm decide’.
Using multiple strategies
You might have noticed that there is an option to select multiple strategies. In some cases, you may not always want to narrow down your touchpoint to just a single strategy. At the same time, there might be some strategies you definitely don't want to use. You can manually select which (and how many) strategies you want to use in your touchpoint. The selector looks like this:
In this example, we have chosen to use 4 different strategies. The algorithm will attempt to use them starting at the top and working to the bottom. This means that it will first try to base its recommendations on the abandoned cart. If this proves unsuccessful it will try the most relevant cart based. It will then try the most relevant based on last purchase and finally most relevant. The system will work its way down until it has found one that is successful or until it runs out of options. You can change the order by dragging and dropping.
If you set up a touchpoint with 4 slots and there is just 1 item in the cart of the profile, it will continue with the next recommendation strategy and keep adding content items. So it could be the case that the touchpoint recommendation will contain 1 item of the profile's shopping cart, 1 relevant item based on this cart, 1 item recommended based on their last purchase, and 1 other relevant item for the profile.
Note: It is best to end with a strategy that is always applicable. This makes sure the touchpoint always shows something. On the other hand, you should not begin with a strategy that will always be applicable, as the others won’t be used that way.
Overview of the recommendation strategies
We chose to bundle some of the strategies due to the large number of them. To maintain an overview we split the bundles into separate articles. Please see the list below: