An important element of recommendations is often to set filters. Various filters limit the total pool of products, content, or users that are being drawn upon by the recommendation algorithm, thereby creating more relevant recommendations, when it makes sense to use them.

Adding Filters to Recommendations

Filters are to be found at the bottom of the algorithm portion of any given recommendation wizard.

Here, you’ll find various commonly used recommendation filters presented as switches to flip on or off, with the option to add more custom recommendation filters. The first drop-down will show you which product variable to filter by, for instance, brand:

In the second drop-down menu, you’ll see the variables themselves to select. For instance, iphone:

These can be based on the user data, which is relatively the same over time, such as account type, or based on which product is being viewed.

Once these filters are added, you will start to see the effects in your recommendation algorithm’s output.

Content Recommendation Filters

Filters in content recommendations will look a little bit different, but work fundamentally the same way.