In his post on retailers driving timely and contextual engagements, Paul Mandeville noted how robust technology capitalizes on modern shopping behaviors and seasonal trends.

Such technology probably includes cohort analysis among its features. Cohort analysis can help retailers better understand observed online customer behavior, a boon as more people research details regarding their purchases.

The Basics Behind Cohorts

Let’s first define what a cohort is and explain its significance to retail. 

A cohort is a group with a common defining characteristic within a specific time. An age group is a typical straightforward example. In another example, cohorts are defined by an activity, such as new memberships during a certain season.

The purpose behind a cohort is to highlight a natural customer behavior cycle for a retailer. This helps a retailer in a few ways:

  • Useful to match marketing automation, such as remarketing against the number of days that a cohort is exposed to a site
  • Highlight emerging pattern among the examined cohort that can influence purchase behavior
  • Highlights how the customer behaviors within the cohort differ than that of other segments

Historically, business analysts have conduced cohort analysis, but most have had to use custom solutions or a spreadsheet to develop their analysis.   

Google Analytics Cohort Reports

Google Analytics Cohort Analysis.jpeg

To make an analysis more integral to other Google Analytics reports, Google issued a dedicated cohort report. The report, in beta mode, appears under the Audience menu.

The cohort report has four main adjustments: Type, Size, Metrics and Date Ranges. Each adjustment offers a degree of selection, but because the report is in a beta mode, choices on each adjustment are limited. Cohort type, for example, is limited to one selection, acquisition date. Acquisition date is the date specified on which the cohort behavior was started.

The date ranges indicate the number of days from an acquisition date. The date ranges are adjustable into 7-, 14-, 21- and 30-day settings. Cohort size can be selected by day, week or month. So the range of data, cohort size and acquisition date are within pre-specified parameters.

Potential Uses

Metrics is simply the measured output that is expected from a cohort examination. The metrics are based on what's currently available for the standard reports. User retention is the most useful metric among the selection, given that so many cohort reports are typically used to highlight sustained behavior over time. But other available metrics include goal completion, page view, sessions, session duration as well as metrics per user selections such as page views per user or session duration per user. 

Analysts can introduce comparisons between segmentation features in the same manner as other Google Analytics reports. This includes default segments such as search and referral traffic, as well as custom segments set by the user. The Google Analytics solution gallery can also be used or imported into the cohort analysis. This lets users take advantage of solutions developed by other analytic practitioners.

The report results appear as a triangular table of metrics – expected in a cohort analysis with cohort periods along the top leg and date ranges along the other. Metrics appear in the table. When taken in, the metrics should reveal the degree of sustaining behavior that is occurring. The example image notes a 6.67 percent average retention of organic traffic 5 days after the acquisition date.

A timeline graph also appears, as it does on most Google Analytics reports, but more than likely, interest among most analysts will lie in the table.

Overall analysts will find ways to make the most of the cohort reports, and beginning to qualify data from sustainability perspective – if there is a rise in session volume, for example, analysts can see if a certain cohort is really responsible and if the increase is meaningfully sustained. Cohort analysis can meaningfully address retailers’ desire to understand their customers better.