In my previous blogpost, you got a brief introduction to the new calculated metric interface and some of the basic capabilities.
As described, the update also came with some more advanced mathematics functions and logic, that analysts can apply to deliver better data.

To illustrate how to use some of those new capabilities, let’s start by looking at something as simple as a bounce rate metric. A metric we use when looking at for example marketing performance and landing page optimization.

Adobe Analytics already have a bounce rate metric. Try run the entry page report, and apply the bounce rate metric. You will properly see that many pages have a bounce rate of 100%, just like I do in the report below.

bounce rate report example

The problem here is, that many pages just have a few entries meaning that only a few entries will have to bounce here, in order to generate a high bounce rate.
To ensure we only looks at the pages that matters for us, we will have to remove the pages with a low number of entries, as they are not our key entry pages and they should not be those we are focusing on in our analysis.
Adobe do not have a filter to remove pages (or other dimensions) with a low metric number. But we could set a fictive bounce rate of 0% for pages with a low number of entries, to push those to the bottom of our list.

In the following example I have chosen to set bounce rate to 0% for pages with less than 500 entries. This number is subject to be changed if you change the time period and depending of the number of visits to your site.

The definition of a calculated metrics that sets the bounce rate to 0% based on low entries (under 500 in this example) look like this:

Filtered revenue per visit

The metric builder canvas will look like this:

Filtered revenue per visit

We simply say that if entries for an entry page is above 499, then we calculate the bounce rate. Otherwise we default to a 0-value. Adding this metric to our report will show us something like this:

Entry pages report example

 

You can of cause expand those calculated metrics to only hold desktop or mobile traffic, to understand how users engage with your site and content based on what device they use.

In my past article, I used a function called COLUMN SUM. This function simply sum all line items of a specific metric together. In my previously post I used it to get a total of new visitors, allowing me to calculate the share of new visitors coming from specific traffic sources.
With the latest release of calculated metrics in Adobe Analytics, Adobe added many new functions such as Mean, Median, Column Minimum/Maximum, Row Count/Min/Max, round as well as a logic functions such as and, or, if and greater/less than operators.

All those new functions allow analysts and marketers to get better insights and metrics directly inside Analytics.

An example could be that you as a marketer need to identify products that generate high revenue per visit, to priorities what products to promotion on your homepage. You could easily just extract visits and revenue for all of your products, and then divide the two figures by each other.
That would basically give you the figures. But that calculation does not take into account products with low traffic and high price points.
Depending of the traffic to your site, you may want to use a different threshold then I do in my example, and you may also want to adjust it as you start working with it.

In the example below, I setup a calculated metric that will calculate the revenue per visit for products with more than 100 visits (in my reporting period). Products having fewer visits than the 100, will just be assigned the value of 0.

The metric definition looks like this:

calculated medtric - revenue per visit

If the statement is true, then the metric will simply calculate revenue / visits and if the statement is false, it will just assign the product (line item) with a 0.

Below a report where the metric is added. Here we see that our Nikon D3300 DX-format DSLR Kit w/ 18-55mm DX VR II generated most revenue per visit and right behind are the Canon EOS Rebel T5 DSLR CMOS Digital SLR Camera.

6Filtered revenue per visit report example

If we would just do a simple calculated metrics taking revenue and divide it by visits, our report would look like this, resulting in that we may have chosen the Nikon D3300 DS-format camera to be one of our promoted products on the homepage, even so it isn’t visited or ordered much.

Revenue per visit report example

When doing this type of reporting, you need to take into account if products was released or promoted within your analysis period, and therefore performs differently that normally.

The latest update to calculated metrics in Adobe Analytics simply gives you as a marketer or analysts many new opportunities for building smarter reporting and performing deeper analysis on your data in an easy way.