Calculated metrics for Universal Analytics

Google have just announced that they have released calculated metrics in BETA for Universal Analytics. Calculated metrics fill a sizeable chink in the GA armour as it was one of the reasons why some analysts preferred the highly customisable Adobe Analytics, which has had this feature for years. Available for Universal Analytics accounts only, you can now create calculated metrics using all of the available metrics - standard and custom - as well any numerical strings. While you won't be able to apply these metrics in standard reporting, you can access them through:

  • Custom Reports
  • Custom Dashboards and Widgets
  • Unsampled Reports

Creating calculated metrics

Once there, you'll need to name it (the external name used for the API will be automatically populated), give it a formatting type, then the formula. It's a pretty stripped down widget, but it's pretty intuitive and once you start typing the autofill feature kicks in so there's relatively little to do other than knowing the names of what's available in your setup.

There are five available formats:

  • Floats - number with decimal places/fraction
  • Integer - whole numbers
  • Currency (Decimal)
  • Time
  • Percent

Use Cases

Calculated metrics are especially useful for organisations that have models different to the product based ecommerce site, for example: the subscription sites, content sites and sites that promote offline transactions through call centres. Also, for doing deeper analysis with metrics that are unique to your organisation, calculated metrics within GA will stop you having to export your data into Excel and run calculations there.

Customer life time value

If you have a website that sells a product based on a recurring fee - insurance and publications for example, the initial value measured in GA will rarely reflect the actual value of the customer as they will be paying on a monthly basis. This often makes the revenue figure redundant in your reports. With calculated metrics you can start to make these more relevant by woking out the average amount of months a customer subscription lasts for your product, then using a calculated metric to multiply the initial price by the average customer lifetime.

Average scroll depth per page

For content sites, you can measure average scroll depth per page to get an idea of how engaging the content is and at what point users stop reading. This can affect not only how you structure copy, but also where linked stories and promotional material should be placed on the page. For this to work you will have to be capturing scroll depth in a custom metric. We'll be looking at 4 scroll points 25%, 50%, 75% and 100%. As custom metrics can only be captured at the hit or sessions level, these are captured as separate custom metrics - '25% Scrolled', '50% Scrolled', '75% Scrolled' and '100% Scrolled' - so when a user scrolls, each of these metrics will accumulate a value.