In a PPC campaign, dimensions (e.g. keyword, device, geography, etc.) can be combined with metrics (e.g. click, cost, conversion, etc.) to describe a user interaction.

Suppose, the keyword=‘PPC analytics’ has 1,000 clicks on device=mobile in geography=Houston. In this example, PPC analytics, mobile, and Houston are attributes of dimensions, and clicks is a metric.

As the number of dimensional attributes (i.e. number of keywords, geographies, etc.) and metrics grow in a campaign, the number of combinations will grow proportionally. This growth creates complexity, and makes it difficult for you to detect a change.

Let’s consider another example, suppose we want to know:

• Which device is best for conversion, the choice is going to be 1 out of 4 (i.e. desktop, mobile, tablet, smart TV) e.g. desktop

• Which day of the week is best for conversion, the choice is going to be 1 out of 7 (i.e. 7 days of the week) e.g. Friday

• Which hour of the day is best for conversion, the choice is going to be 1 out of 24 (i.e. 24 hours of the day) e.g. 7am

• Which device + day of the week + hour of the day is best for conversion, the choice is going to be 1 out of 672 (i.e. 4 * 7 * 24), e.g. mobile + 9am + Tuesday

The complexity of combinations can negatively impact your ROI because, when facing a large number of combinations, it becomes difficult to detect where important changes are occurring.