How to Run a Cohort Analysis in Google Analytics [Complete Guide]
Reports, stats and insights are a great asset to have, but let’s face it: all that great information, at times, can feel a bit overwhelming and we could end up buried in data.
I happen to be one of those marketers that likes pulling reports (generally by automating them), and checking on my stats quite often, however, I often find it hard to drill down to the reasons behind why some figures are a certain way and uncovering the causes behind growth or decline on some metrics. That is where a cohort analysis can help.
In this blog post, I’m going to explore the basics of cohort analysis and why this is the right path to follow to understand a user’s behavior, engagement and more.
In this article we will cover:
- What is a cohort analysis?
- How to run a cohort analysis report in Google Analytics.
- Elements of the cohort analysis report.
- Use cases: What can you use cohort analysis for?
- How to get started.
What is a cohort analysis?
A cohort is a group of people who share something in common within a defined time-span. Cohort analysis is specifically the analysis of cohorts in regards to big data and business analytics. It’s a subset of behavioral analytics that measures user engagement over time.
Cohort analysis is a useful report to understand seasonality, customer lifecycle and the long term health of your business.
Tech companies and B2B organizations that have long sales cycles and struggle to understand the effectiveness of their marketing campaigns might find cohort analysis a very powerful tool to use.
It’s also beneficial for other industries that rely on long-term engagement. By using cohort analysis to measure retention, you can see the percentage of users that returned to your website over time, to complete transactions or convert.
How to run a cohort analysis report in Google Analytics
Elements of the cohort analysis report
#1- Cohort Type
The cohort type is the dimension that’s the basis of the cohorts. Acquisition date is the only version currently available in Google Analytics. Acquisition date is the first time a user is recognized as interacting with your content, that’s it, when users started their first sessions.
There are of course limitations to Google Analytics cohort analysis reports.
The first one, as previously mentioned, is the possibility to only define cohorts based on acquisition dates.
The second one is that tracking returning sessions with several devices is not accurate in Google Analytics, therefore if a person checks your website from their laptop and the second session is done through a mobile device, then that session could be double counted in that cohort tracked separately. The same applies if that person checks your website from another browser or clear cookies.
#2- Cohort Size
The size of each cohort is determined by the time frame. In Google Analytics you can choose from: by day, week or month.
What does this mean? A cohort size by day will show you all the users that you have acquired during the selected date range (in this case the last 7 days) by day. In the example below you can see the size of the cohort group for each date within their respective rows.
This is where you can select the metric you want to measure for each cohort. There’s three groups of metrics you can analyze:
Metrics per user
All of these metrics are indicators of the cohort, divided by the total cohort size. Metrics include goal completions, page views, revenue, session duration, sessions, and transactions per users.
In the example below, I’m analyzing daily cohorts of users by page views. On March 25, 134 users generated 677 page views, on the following day 106 page views, and on the following day 85 page views.
Why is this data interesting? It shows you users’ interaction with your website after you have acquired them. In order to analyze this data properly, it’s important to have a marketing calendar to track your activities properly and understand which campaigns or initiatives are generating a change in the users’ behavior.
Total metrics show the total indicator for the entire cohort.
User retention is one, if not the most important metric on the cohort analysis report. It shows the number of users in the cohort who returned to your website in a given period of time, divided by the total number of users in the cohort.
PRO TIP: Find which cohorts have a better user retention rate than others, and analyze possible reasons why. Did you launch a retargeting campaign? What is it that you did during that time that impacted retention rate and could be taken as a good lesson learned for the future?
#4- Date range
The date range is the time boundary that determines the data being displayed on the report. These correspond to the columns, and it will show the metrics selected by date range, for each cohort size (rows).
#5- Visualizing data
Now that you know the basics, you can play around with the reports and compare, for instance, two specific cohorts. That can be done by selecting them under Acquisition Date. In the example below, I chose to show cohorts by revenue, for two specific cohort groups, based on acquisition date, by day.
So in short, I am comparing two cohorts of acquired users on April 6 and April 9, 2020 and the revenue they have generated during the last 14 days.
It’s interesting to see a spike and see how conversions occur on day 4, since the time the user was acquired.
#6- Adding segments
Adding different segments to your cohort analysis is key to understanding your users and their behavior.
To get started, click on the Add Segment button on the top right. A list of segment options will appear. Uncheck all users If you want to see specific segments.
For the purpose of this analysis, I’ve selected Mobile Traffic and Tablet and Desktop Traffic.
So, let’s say I want to understand users’ purchasing behavior a few days after visiting the website for the first time, but I want to know what’s the difference between mobile, tablet and desktop.
Why would this be interesting to know? Let’s look at the numbers.
- First of all, mobile traffic is higher than desktop and tablet combined.
- However, revenue figures are similar during week 0, when users were acquired. That could mean that desktop and tablet users have a higher conversion rate.
- Now looking at revenue in the following weeks, that’s where it gets interesting. There are desktop and tablet users that convert during weeks 4, 5 and 6 after being acquired.
These insights can be quite powerful and can help you shape your marketing strategy. Should you focus on acquiring more desktop users and incrementing desktop traffic? Or should you review your mobile conversion path and optimize it? Should you increase your desktop ad spend?
When analyzing these figures, you definitely need to consider other factors influencing these trends, and it’s a good practice to continue to pull these reports over time to understand how marketing campaigns might be affecting user behavior.
What can you use a cohort analysis for?
Understanding sales cycles
Cohort analyses can help you understand sales cycles and behavior. Especially if you are a SaaS company.
If you run a subscription-based model, the cohort analysis report is key to understanding churn and optimizing it over time.
Learn about shopping habits
For eCommerce businesses, purchase spikes and examples mentioned earlier in this blog post can help you have a high-level overview to get started with cohort analysis.
Experimenting and testing a hypothesis
By implementing a cohort analysis, we can observe the outcome of our hypotheses. For example, let’s say our hypothesis is that desktop users convert better. Then you can use the reports mentioned above to test your hypothesis and conduct further research to make sure your hypothesis results are correct. Then, you can optimize your marketing plans accordingly.
How to get started with a cohort analysis
There are a lot of different data visualization tools that can be used to run cohort analysis reports, like Tableau, Looker and others, which have several more advanced options available for cohorts, compared to Google Analytics.
If you are already using Google Analytics, you can test the reports mentioned above, but if you are already using other data visualization tools you can unify all your marketing and sales data into a single source of truth using a tool like Improvado.
Tools like Improvado can help you pull all your data into one place so that you can understand the full path that omnichannel consumers follow — across devices, across media — and so you can see which channels are creating value. Click here to learn more about how Improvado works and integrations with platforms.
While getting started with a cohort analysis might look a bit overwhelming and slightly confusing, following these few simple steps mentioned above in this blog post can help you get started and have a better understanding on how cohorts work. This will help you start optimizing campaigns and your website now.
To summarize, cohort analysis allows organizations to check a very specific segment, analyze only the relevant data, and optimize based on the findings.