Lookalikes is a predictive recommendation type. It uses a type of machine learning called content-based filtering to find types of product a customer has shown an affinity for.
How lookalikes work
When a lookalikes recommendation is created, your catalog is analysed. The process groups similar products together. If you were a winter sports retailer, products could be grouped by tags such as 'snowboard', 'female', 'burton', 'red', and so on. Each attribute is scored; the more confident the machine learning is that the tag is right, then higher the score.
Next, a contact’s purchase and browsing history are analysed. Products with a suitable similarity score are chosen. Each contact is then given a ranked product set unique to them.
Our predictive recommendations automatically generate a fallback. For contacts that we cannot predict products for, they will see the fallback. The fallback is another product recommendation, such as a best seller, which you can edit as needed.
Accessing predictive recommendations
If your account does not have predictive recommendations enabled, go to Campaign > Product recommendations, click Create recommendation, mouse over the recommendation type you're after under the 'Predictive' section and click on Learn more.
The sidebar will allow you to request the feature is enabled on your account (for an additional monthly charge). We'll first review your data and then confirm with you that the feature is ready to use, or if there are data dependencies to be resolved first.
Predictive recommendations are available on a 14-day free trial.
Getting the best results
Before getting started, please review the data dependencies for this product recommendation type.
This model is most effective when there's good quality and depth of product data available. Product detail can be improved by using product data enrichment. Catalogs with a high number of products but with low orders will not give optimal results.
Order data is one indicator of customer preference. By using Web Behavior Tracking, the machine learning model's accuracy will be improved.