Cohort Analysis is one of the most important reports in Analytics, yet it is overlooked by most. In a nutshell, it provides actionable data with minute details of your audience which is usually missed while tracking. Overtime, it tells you if your business is doing good or not.
Getting in the details
What is cohort analysis?
To begin with, a cohort is a set of users who share the same characteristics. For example, it could be all users who were acquired in the past week. In Google Analytics, you can segment your cohorts based of their acquisition date i.e. the date of their first visit.
You have a choice of tracking your cohorts based on the day, the week or the month they were acquired. So for instance, my first visit on your site was on 14th November 2017, I belong to the 14th November, 2017 cohort along with the 3rd week of November cohort and the November month cohort. But you won’t find me in the October or December cohorts.
Why is it even important?
Cohort Analysis is observing the sets of users and tracking a change in behaviour. You might find out that your second set of email converted a lot more people than the first one. You can compare the behaviour of both the sets and come to a conclusion as to what triggered the change.
Cohort Analysis helps track retention, engagement and how users react to your marketing efforts. It leads you to find the gaps, look for behavioural differences, appropriate time and way to execute your future campaigns and ask relevant questions.
At the beginning of this post, I mentioned this is the most overlooked report. That is because most of us look at the numbers in the form of averages. We don’t break them down to get into the details and therefore, lose valuable insights. Cohort Analysis brings to your notice the development and decline in your metrics which makes your data more actionable than the usual tracking.
Breaking down the cohort report
Taking a look at the elements of the cohort analysis report, you will find web analytics tools that will help you to create your cohort report, data in a graph and data in tabular form.
The most important things you should be aware of are:
1) Cohort Type – It’s the date of acquisition you want Google Analytics to select while creating a cohort for you. It simply means you wish to see data of users based on their acquisition date. For now, you can segment your users in this report only on the basis of acquisition i.e. their first visit to your website as it is presently in beta.
2) Cohort Size – This is the period you want to select when tracking your users. For instance, if you want to track the acquired users in the second week of October, this will be your cohort size. Similarly, if you wish to see the data of acquired user on 10th October, this particular date is your cohort size.
3) Metric – It is the data that you want to measure. It can be user retention, revenue, duration of a session, etc. These metrics are used to measure the performance of a specific cohort and look for intense changes taking place.
4) Data Range – This is the time range you want to select for your cohort. If it is the last 7 days then Google Analytics will show data from the 7th day till today.
Let’s understand the Cohort Table in Detail
- Box #1
Here you see the cohorts on the basis of their acquisition date. They are sets of cohorts acquired on the same date. The elements chosen are – cohort size in weeks and data range of the past 6 weeks. - Box #2
This shows the time period of their retention over the past 6 weeks. - Box #3
This is where the important stuff is. This is the data you have to analyse and crack. Take any row from the table above and see how the retention rate drops from 100 to 0. Third row on the table which represents the cohort 12th March to 18th March has 3.38% of return visitors in week 1 from the 100% in week 0.
This is basically how you analyse each cohort and then compare the behaviour with different cohorts.
Common cohorts to use –
Even though we are restricted to one cohort type – acquisition date, there are basic cohorts you can use as per your business model. They are as follows:
1. In App Purchases –
It’s great how you can look at the behaviour of users from the time your app is downloaded by a user. Your cohort for everyday viewing can be – User Revenue for the last 7 days, 14 days or a month.
You can also create a cohort that tracks the users who haven’t engaged. For that your cohort can be – Sessions per user for the last 7 days, 14 days or a month.
2. Ecommerce –
For this business, we can create and measure cohorts for user retention, user transaction and revenue in total. You can track customers who purchased things that are expensive or those who did not purchase a lot. Creating a weekly cohort will get more insights in this case than a daily cohort. So your cohort will be User Transaction for the last 14 days.
There’s another theory that says if users have made purchases of less expensive products they might purchase more frequently. Here you can try cohorts that are traceable on a daily basis. It can be daily cohort for transactions per user for a month.
3. Publishing and Lead Generation–
Here, you’d want to track overall engagement so your cohort must be able to show you engagement on high priority. This can be done by tracking the Sessions per users, Page views, User Retention and Sessions duration per user for 7 days, 14 days or a month.
The duration also depends on how often you publish content. If you are a blog that puts content every other day then a 7 day period is great but you post like once or twice a month, a monthly cohort will be of better use.
Similarly, for lead gen the cohorts really depend on how often your products are sold. If you sell expensive products that are bought once in a while, you might want to have a look at cohorts on a monthly basis but you sell less expensive products that are sold frequently, you can benefit more from daily cohorts.
These metrics and data might take a toll on you but it’s only a matter of time until you get used to them. They are a fantastic way to analyse behaviour or users in groups adding an extra layer of detail in breaking down your customers. It’s not easy to crunch these numbers but they give a lot of valuable insights about the way your business is done.
