Note: The first in the Apollo BR series, this post discusses Apollo BR37370, the business requirement for Product SKU Selections, one of over one-thousand Adobe Analytics solutions that Search Discovery has added to Apollo. If you’d like to improve your Adobe Analytics implementation, please reach out to us at firstname.lastname@example.org.
If you are an online retailer, tracking products and SKUs in Adobe Analytics can be tricky at times. Back in 2012, I wrote about this and why you should use a Merchandising eVar for tracking SKUs. But as I have been building out solutions for the Search Discovery Apollo product, I’d like to dig a bit deeper and share some ideas I have been thinking about (along with my partners in crime Anna Salome and Hailey Meekins).
In my previous post I showed how to track Product Views and SKU Views. Now, I want to talk about SKU selections. Let’s imagine that your business requirement is to understand which product SKUs were selected the most for each product (assuming your products have product SKUs). You may want to know this to determine inventory needs or possibly if there are some SKUs that can be dropped altogether. Unfortunately, this isn’t as easy as it sounds. To measure this, the end goal is to set a metric (success event) each time a visitor selects a SKU for a product and to divide this metric by the number of Product Views. For example, if a visitor views Product A and selects SKU A1, A2, A3, there would be 3 SKU Selection events and 1 Product View event. SKU selections would be set in custom links so as to not inflate the number of Product Views. In this scenario, Product A would have an average of 3 SKUs Selected per Product View in the calculated metric:
Here is sample report you can run that shows this calculated metric in action, looking for products that have the most SKUs selected per product:
How to Utilize Default SKUs
This concept gets a bit more complicated if your website/app allows SKUs to be selected by default. In some cases, product pages open with a SKU selected by default, and in other cases no SKU is selected until a visitor manually selects it. Both of these cases have to be taken into account.
In cases where a SKU is selected by default, you don’t want to set the new SKU Selected success event (mentioned above) when the product page loads, since no one actually selected a SKU. If you did, it would dramatically inflate the count of the default Product SKUs Selected. Instead, set a different success event called “Product SKUs Selected (Default),” and only set the “Product SKU Selected” success event when visitors manually select a SKU.
By splitting these out into two different success events, it is much easier to see which Product SKU Selections were done manually by visitors and which were not. Of course, you could just use one success event and flag the different types with an eVar, but that would require a bunch of work to make derived metrics off of that eVar (and wastes an eVar). With this approach, you can easily see Product SKU Selections by each type (or combined) by adding the default and non-default success events in a calculated metric, like this:
Here is a sample report that shows product SKUs and product SKU selections done manually, by default, and the total:
Finally, if your site does have default SKUs, you may want to add a second version of the previously mentioned Average Product SKU Selections per Product View calculated metric. The one we created earlier assumes manual selections, but you can create a second one that also includes default SKU selections:
Dealing with products and product SKUs can be difficult. There are a lot of intricacies and moving parts, but if you want to improve your understanding of your products, you should make the time to do it right. Tracking product SKU selections is a great way to understand which products SKUs your customers are interested in and which they are not. This can help you refine your product catalog, determine which SKUs to highlight, change the way you choose promotions and cross-sells, and inform product inventory.