Behavioral Data
Some recommendation types use behavioral data from your shoppers to train machine learning models to build personalized recommendations. Other recommendation types use catalog data only and do not use any behavioral data. If you want to start quickly, you can use the following, catalog-only recommendation types:
More like this
Visual similarity
So when can you start using recommendation types that use behavioral data? It depends. This is referred to as the Cold Start problem.
The Cold Start problem is a measure of how much time that a model needs to train before it can be considered high quality. In product recommendations, it translates to waiting for Adobe Sensei to train its machine learning models before deploying recommendation units on your site. The more data that these models have, the more accurate and useful the recommendations are. Collecting this data takes time and varies based on traffic volume. Because this data can be collected only on a production site, it is in your best interest to deploy data collection there as early as possible. You can do this by installing and configuring the magento/production-recommendations
module.
The following table provides some general guidance for the amount of time that it takes to collect enough data for each recommendation type:
Most viewed
, Most purchased
, Most added to cart
)Viewed this, viewed that
Viewed this, bought that
, Bought this, bought that
Trending
Other variables that can impact the time needed to train:
- Higher traffic volume contributes to faster learning
- Some recommendation types train faster than others
- Adobe Commerce recomputes behavioral data every four hours. Recommendations become more accurate the longer they are used on your site.
To help you visualize the training progress of each recommendation type, the create recommendation page displays readiness indicators.
While data is collected on production and machine learning models are trained, you can implement the remaining tasks necessary to deploy recommendations to your storefront. By the time you have finished testing and configuring recommendations, the machine learning models have collected and computed enough data to build relevant recommendations thus allowing you to deploy the recommendations to your storefront.
If there is insufficient traffic (views, products bought, trending) for the majority of SKUs, there may not be enough data to complete the learning process. This may cause the readiness indicator in the Admin to look as if it were stuck.
The readiness indicators are meant to provide merchants with another data point in choosing what recommendations type is better for their store. The numbers are a guide and may never reach 100%.
Backup recommendations backuprecs
If there is not sufficient input data to provide all requested recommendation items in a unit, Adobe Commerce provides backup recommendations to populate recommendation units. For example, if you deploy the Recommended for you
recommendation type to your homepage, a first-time shopper on your site has not generated enough behavioral data to accurately recommended personalized products. In this case, Adobe Commerce surfaces items based on the Most viewed
recommendation type to this shopper.
The following recommendation types fallback to Most viewed
recommendation type if there is not sufficient input data collected:
Recommended for you
Viewed this, viewed that
Viewed this, bought that
Bought this, bought that
Trending
Conversion (view to purchase)
Conversion (view to cart)