Dashboard
Last updated
Last updated
The Product Recommendations dashboard provides a high-level view of your recommendations performance. It can be accessed via the left sidebar in the Merchandising section of the portal.
The dashboard shows tables and charts which help visualize the absolute metrics and falloff rates between each of the four key product events: Recommendations ⇢ Products Clicked ⇢ Products Added to Cart ⇢ Products Purchased. It also includes crucial revenue information that can be used to see how Fanplayr's recommendation service adds to your revenue.
If your account has Personal Shopper enabled, the dashboard will also contain a few general shopper-related KPIs as well as an additional filter to limit the entire dashboard's data to just events originating from a Personal Shopper.
Data on the page, other than User Events and All Time values, can be filtered based on the date and model filters at the top right of the dashboard.
At the very top of the dashboard, there is a sentence that provides a few real-time "User Event" numbers. These numbers encompass all product views, add to carts, and purchases that Fanplayr tracks, regardless of whether the action is linked to Merchandising. These numbers exist simply to give you a relative idea of the role Merchandising plays in your overall metrics. They also reveal some of the underlying statistics that Fanplyar uses in order to build and train your AI recommendations models, like Frequently Bought Together and Recommended For You.
In the "Overview" section, you will see "All Time" numbers below each statistic which provides the value of that statistic for all time with no filters applied. The percentage next to each statistic indicates how each value has risen or fallen relative to the previous date period. For example if the filter is set for the last 30 days, the percentage will compare the current data to the prior 30 day period.
The remainder of the dashboard consists of various charts and tables that highlight certain aspects of your recommendation performance, including trends of user behavior and revenue statistics.
At the bottom of the dashboard is a table of raw recommendation statistics. Above the table is a small aggregation editor where you can select the aggregation dimensions for the table. If you choose not to aggregate by any dimension the resulting table will be only one row and contain aggregations of all data (essentially the same as the "Overview" section). If you aggregate by date and model you will see a breakdown of performance by date and recommendation model. This is a powerful tool that can be used to analyze your merchandising performance at a granular level, and it will help you to find what strategies work best for your site.
Property
Description
Recommendations
The number of times the a model was presented to visitors in a Fanplayr widget.
Products Clicked
The number of times a recommended product was clicked. This action directs users to the detailed product page.
Products Added to cart
The number of times a recommended product was clicked and then subsequently added to cart.
Products Purchased
The number of times a recommended product was clicked and then subsequently added to cart and purchased.
Sessions with Orders
The total number of sessions where at least one order was placed containing at least one recommended product.
Revenue (Recommended Products)
The total revenue generated from the purchases of recommended products.
Revenue (Orders with Recommendations)
The total revenue from orders that contain at least one recommended product.
Personal Shopper Views
(If enabled) The total number of times your Personal Shopper(s) was viewed during the filter period.
As mentioned above, this dashboard will contain additional data points if your account has Personal Shopper enabled. If enabled, there are a few things to keep in mind:
There is a new filter called "Only Personal Shopper events" which will filter all data to just those that originated from a Personal Shopper. This includes recommendation views, revenue, session counts and so on.
Personal Shopper views are not tied to a single algorithm, so breaking down the data by algorithm (either by filter or aggregate) will not limit the view count for Personal Shoppers.
Recommendation views are still counted when viewed inside a Personal Shopper. For example, if there are five recommendation components in one Personal Shopper and the shopper is viewed once, that will count as five recommendation views towards the total. Of course, the shopper view count will only be incremented by one.
Outside of filtering data, there are two other actions you can take on this dashboard:
Download order data — At the top of the dashboard, under the "Dashboard" title, you will see this button. Clicking this will download a CSV of all orders + products purchased linked to recommendations. The output will be one row per product purchased.
Download aggregate data — At the bottom of the dashboard is a table that contains merchandising statistics aggregated by customizable dimensions.. Clicking the "Download aggregate data" button will download the entire table as a CSV. This table powers much of the data throughout the dashboard.