Recommendation Models Metrics
Last updated
Last updated
You can learn more about how metrics are handled by visiting the Metrics page.
This metrics tab contains data about the performance of your recommendation models. Without selecting a recommendation model, account-wide metrics for all your recommendation models will be displayed. Selecting one recommendation model will restrict the displayed metrics to that model. Selecting more than one will generate a comparison of metrics between the selected models.
This metric shows the number of times the selected products were included in a recommendation in the defined time frame.
Example: In the context of using a recommendation model for a retargeting campaign, this metric will show you how many times a product was included in a recommendation, use this data to understand which of your products are being recommended on your banners. On this graph, the defined time frame was 1 week divided into 30-minute periods. You will also notice that the product recommendation count was nearly 500 most of the time.
This metric shows the number of times the selected products' recommendation was clicked in the defined time frame.
Example: In case you have a campaign using a recommendation model, this metric will show you a count of clicks on your recommendations, this data will help you to measure your recommendation model efficiency, which will allow you to test different marketing strategies and observe how your audience is reacting on it. On this graph, the defined time frame was 1 week, divided into 6h periods, represented as dots. You will notice that on June 22nd at 12 PM there were 4 clicks on the recommended products.
This metric indicates if the recommendations are being fulfilled with products or if it's missing products in the defined time frame.
Example: Suppose you have a campaign using a recommendation model to generate banners. You will need to set the number of spaces to be filled with recommendations. This metric shows you if your spaces are being fulfilled, use this data to follow if your recommendation models are working properly, which means that the recommendations are being 100% fulfilled, and therefore your banner will not display any empty space. This graph's defined time frame was 1 week, divided into 6h periods, represented as dots. You will notice that on June 24th at 12 AM the recommendation fulfillment rate was nearly 80%, meaning that some spaces on the banner were not filled with products.
This metric shows the number of times a recommendation was requested in the defined time frame.
Example: In case you are using a recommendation model to generate your banners, this metric will show you how many recommendation requests were processed by the platform, this data is useful to follow your consumption of the recommendation models feature. This graph's defined time frame was 1 week, divided into 6h periods, represented as dots. You will notice that on June 23th at 6 PM the recommendation request count was nearly 2000 requests.
This metric shows the number of times the selected product recommendation was viewed in the defined time frame.
Example: In the context where you have a campaign that is using a recommendation model to generate banners, this metric will show you a count of views on your recommendations, the data this metric provides is useful to understand if your recommendations are being viewed, furthermore, comparing this metric with the Recomendation Click Count metric will allow you to understand if your banner is interesting to your targeted audiences. On this graph, the defined time frame was 1 week, divided into 6h periods, represented as dots. You will notice that on June 26th at 12 PM there were nearly 4000 views on your recommendations.
These are all the metrics available in the Recommendation Model feature for analyzing the performance of your recommendation models. Additionally, when checking metrics, you can always check our to access additional information about a specific metric.
