Uncover the Hidden Riches from Your Data with the Google Cohort Analysis Report

Buried within your data are treasures waiting to be revealed. One way to uncover these insightful bits of information, aka ‘gems,’ from your data is by effectively and strategically utilizing cohorts and cohort analysis (aka ‘your treasure map’) within Google Analytics. These tokens of insight can help in addressing business questions, but many still don’t effectively understand how cohort analysis works and what information it provides. In this post, I am going to outline what cohorts are and what cohort analysis is; along with explaining why businesses need them and how to set up the report in Google Analytics. 
What are Cohorts?
The definition of a cohort is a group of people who share a common characteristic over a period of time. For example, if you were born in 1983, then you are part of the 1983 generation cohort. In the digital analytics realm, a cohort stands for the same principle. The most common characteristic that defines a digital cohort is usually the user’s first engagement with the digital product or service, and the time period from the date when the first engagement occurred. For example, if you visited website A by clicking on the hyperlink shown in an email campaign on March 9th, 2015, then for website A you are in the March 9th, 2015 email campaign cohort. Similarly, if you downloaded and installed a mobile app on March 15th, 2015, then you will be placed in the March 2015 user cohort for the mobile app. 
Successfully grouping these cohorts can help to measure user engagement over time. By comparing the similarly characterized cohorts that start from different dates over a period of time, the business will be able to find out how the customer behavior differs from one cohort to another and the causes behind it.
Why businesses need Cohort Analysis?
Cohort analysis is not an arbitrary reporting effort. When done correctly, it can help to identify meaning within your data that allows for strategic and improved business decisions.
Understand User Behavior
One reason people choose to use cohort analysis is to understand the user behavior. For example, for an ecommerce website, you might want to know how soon a visitor converts after his or her first visit, how often the visitor converts in the following weeks, and when the visitor choose to disengage with the website. With cohort analysis, you could create a monthly cohort of the users and evaluate their engagement level over time. Using such a technique, we will be able to learn how a user engages with our product during its entire life cycle.
Evaluate Marketing and Advertising Campaign Performance
Another reason cohort analysis is valuable is because it helps us better evaluate our marketing and advertising campaigns. It’s common for companies to continuously launch various marketing campaigns all year round for their digital products and services. However, if marketers and analysts merely rely on the general acquisition reports to answer the questions of campaign performance, they are missing a large part of the story (learn more about successful storytelling). Under the stimulation of campaigns, a large amount of new users could be attracted to the website or mobile app and engage actively with the service. In this case, the acquisition reports will show spikes in the session numbers from these specific marketing channels and indicate that the campaigns are working. However, if these newly acquired users terminate their engagement after a couple of weeks, either because there is a problem with the service or they find the service no longer useful, you will not be able to identify this retention issue from the acquisition reports. The diminishing user engagement number would be properly offset by the growth number of new users brought in by other campaigns and your services issue would go undetected, causing a larger issue over time. 
This could also be problematic in when the evaluation of a certain campaign shows success based on the initial reports when in reality, it failed in retaining users. Your business would be wasting significant investments if launching the same kind of campaign under the false notion that the campaign was a success. By creating cohorts to group the new users coming from different campaigns and tracking their behavior over time, you would be able to circumvent this issue as you and assess that “Campaign A cohort brings more qualified users than campaign B”. 
Measure Functionality and Design Consistency 
When undergoing significant functionality or design updates, it is important for you to find out if the new features are actually improving the user experience or not. This is another good place time to conduct a cohort analysis. For example, if your company decides to publish a completely new website, it will first offer the new experience to a group of users and wait to hear their feedback. During this period, the business could also create a cohort around these pilot users and record their engagements over time. After several weeks, when the users become accustomed to the new experience and their behaviors start to reflect their genuine preference, you could then compare this engagement trend with the one on the old website to see if the new experience works effectively in retaining the users. 
If the comparison shows that over a specific time period, users continue to stay active on the new website more than the users of the old website, you can be more confident in your new website and the migration test can move on to the next stage. However, if you see a significant drop in user engagement after several weeks, you may need to connect with your IT and UX departments to find out why users are abandoning the new experience.  For small design and content updates, you could simply use the existing cohort groups to perform the analysis. In comparing the user retention and engagement patterns over time you will be able to evaluate if changes made during each period generate a positive impact on the product or service.


Direct Business Activities
Cohort analysis could also help forecast expected returns and allow you to make informed business plans accordingly. Take an ecommerce website for example; based on the historical cohort analysis results, the business should already know how their users react to a certain kind of campaign during a certain period of time. Let’s say we know the daily average amount of paper towels each user buys during a household essentials sale. The next time a similar campaign is launched on the website; we could use this information to adjust our inventory during and after the campaign so that we could reduce operation cost by effectively managing the anticipated logistics.  In a similar manner, we could use cohort analysis to help remarket plans and launch new product features at the right time, when user engagement cools down. Overall, cohort analysis is a great tool that allows businesses to make more robust decisions and plan ahead. 
Setting up Cohort Analysis in Google Analytics
Before Google Analyst encapsulates cohort analysis into an independent option in its reporting interface, you will have to create a cohort segment and modify the tracking code to incorporate the first session date in order to perform the cohort analysis. This complex process has been simplified into four click steps. 
  1. Cohort Type
    The Cohort Type is the characteristic you specify to define the cohort. It is based on a date of a contained event. So far, Google Analytics only provides one option here: Acquisition Date. The Acquisition Date is the first time a user interacts with your content. When selected in the Cohort Analysis report, the cohorts are grouped based on when users started their first sessions.
  2. Cohort Size
    The Cohort Size defines the time frame based on which the cohort is created. There are currently three options for this parameter - by day, by week, and by month. If you choose a day as the cohort size, it means you want to see the engagement history of the users who were acquired on the same date. Similarly, if you choose a week as the cohort size, you are telling Google Analytics to show you all the users who are acquired within a week.
  3. Metrics
    Metrics specifies the way the evaluation should be performed. The most frequently used metrics would be ‘User Retention,’ which represents the number of users in the cohort who returned in the ‘Nth’ time period (day, week, month) divided by the total number of users in the cohort. You could also choose average, such as ‘Goal Completion per User’ and ‘Aggregate.’ Metrics should be selected based on your analysis need.
  4. Date Range
    The Date Range parameter is used to define the time range in which the cohort analysis will be applied. The range varies as you choose different Cohort Sizes. For example, if you choose day as the Cohort Size, then you can only specify the Date Range to be last 7, 14, 21, or 30 days. If week is picked as the Cohort Size, then you will be able to run the cohort analysis for the last 1, 3, 6, 9, or 12 weeks. 
After this four-step process, you will be able to view the time trend chart and data table on the same page. The chart by default shows the cumulative metric values for cohorts. You can use the “N selected” menu to select an independent cohort. A maximum of four cohorts can be shown on the chart at once. In the table, you can see that the dimension column lists all the cohorts defined in this report. Each row indicates how the selected metric evolves over time for a cohort. Segments could be further applied to this report to select the exact target users you are interested in.
Even though cohort analysis in Google Analytics is still in its preliminary stage, this report will continue to be enhanced as the popularity of cohort analysis grows. In general, cohort analysis provides us another magnifier to examine data and make better informed business decisions. It’s a great technique to include in your data analyst’s toolboxes and start using in your daily analysis.  
Questions, comments? Reach out to info@stratigent.com or comment below!
By Yang Zhang
About the Author:

Yang Zhang is a Senior Analyst at Stratigent.

Contact Us Now