Saturday, July 4, 2015

What is Cohort Analysis? - Overview

What is Cohort Analysis?

Cohort analysis is a subset of behavioral analytics that takes the data from a given eCommerce platform, web application, or online game and rather than looking at all users as one unit, it breaks them into related groups for analysis. These related groups, or cohorts, usually share common characteristics or experiences within a defined timespan. Cohort analysis allows a company to “see patterns clearly across the lifecycle of a customer (or user), rather than slicing across all customers blindly without accounting for the natural cycle that a customer undergoes.”

A cohort is a group of people who share a common characteristic over a certain period of time.

A cohort is any group of people sharing a characteristic. 


Example from

Perhaps the most popular cohort analysis is one that groups customers based on their "join date," or the date when they made their first purchase. Studying the spending trends of cohorts from different periods in time can indicate if the quality of the average customer being acquired is increasing or decreasing in over time.

Cohort Analysis in Google Analytics by Yoast

So a cohort analysis is basically the analysis of a group of people, in this case people who interacted with your website at the same date or date range. When clicking Cohort Analysis in Google Analytics, it’ll look something like this:

I don’t know about you, but this isn’t really immediately clear to me, so let me walk you through how to look at it. The chart at the top is a visualization of the average user retention (percentage of returning visitors) for the date range, which is 7 days by default.

The most interesting, however, is the table below the chart. This actually gives us insight in what percentage of people returned to your site within 7 days of visiting it for the first time. Day 0 corresponds with the date in the first column. Day 1 is the first day after someone visited your website for the first time. So the 4.32% at Day 1 in the March 10th row means that 4.32% of the people who visited for the first time on March 10th, visited again on the next day (March 11th). Day 2 is the second day (March 12th) and so on.

Note: this is a breakdown of New Users, so although it says “All Sessions”, this only includes people having visited your site for the first time.

#What can I do with this?

This is a question that I immediately asked myself. It wasn’t completely clear to me right away, so I might be a bit slow, or it’s just not that obvious. I’ll let you be the judge of that ;)

Let me give you an example (not, by the way):

So what happened on March 14th or 15th that made people who visited this website for the first time on March 14th visit again the next day? The retention rate is about 2% higher there, and even on day 2 the retention rate is higher. Maybe they wrote a nice post? This can be a great way of figuring out whether what you’re trying (new content, new campaigns, etc.) is actually working.

Breaking down the cohort

If you need a more specific look on what’s happening, either because you don’t know why the retention rate was lower/higher, or because you’re just a data geek, you’re in luck. You can actually ‘break down’ your cohort analysis by using segments. For instance, if I were to use the Mobile and Tablet Traffic segment on the data above:

Google Analytics will give me this cohort report:

This shows the data for people who not only visited your website for the first time in the set timeframe, but were also on a smartphone or tablet when viewing the site. You can have up to 4 of such segments active at the same time. This way you can see whether the (expected) effect happened for all sorts of people, such as people on mobile phones, people from search engines or direct visitors, etc.

Other metrics

You can actually select quite a few metrics that will make the cohort analysis useful for a lot more than returning visitors:

Although the Cohort Type has a dropdown, it actually just has the one option. The Cohort Size can be set to ‘by day’, ‘by week’ or ‘by month’ and the Date Range will change accordingly. The most interesting though, is the Metric dropdown. You can select a lot of per user metrics (revenue, pageviews, transactions, etc.) or total metrics (again revenue, pageviews, etc.) apart from the Retention metric I used in the examples above.

This means you can actually see a lot of effects, such as whether your overall revenue or revenue per user has increased after a post or campaign. Of course, you can normally see your sales or revenue increase if you have a successful campaign, but this data is different.
You can now see how much revenue you got from people that visited your website for the first time on a specific date and see if these new visitors bought something on that date or in the days to follow. And since you can see this for an entire date range, you’ll also be able to see if that’s a higher or lower revenue than was to be expected.

Let me give you an example. Say you changed your landing page recently, which is tailored to just convincing new visitors of your site to buy a product. You could just be looking at the revenue from new visitors and see if it increases. However, if a visitor were to visit your website for the first time, only to return the next day to buy your product, Google Analytics wouldn’t show it as a new visitor anymore. And that’s why these cohorts actually work: the visitor was new at the set date, so even if they buy the product a day (or 2, or more) later, they’ll still show up in the cohort analysis. So you’re not just measuring direct effect anymore, you’re measuring delayed effects as well!

By the way, to be sure you have just the visitors that visited that specific landing page, you should create a segment for visitors who visited that page.

#The downsides

While looking at the cohort analysis for, I noticed that the Retention metric is quite difficult for our domain. Our traffic, even from the new visitors, is just too stable. The pattern was just the same all the time, no matter what date range I selected. This is probably because we have such a steady flow of new visitors, mainly from Google, that any lift here would only be a small change in percentage.

So, the changes in the percentages are too small; if everything between 3.5% and 4.5% is the same color, it’s pretty hard to distinguish any real differences. Of course, I could just look at the percentages, but that’s just not as convenient.

More importantly, though, we can only create cohorts based on Acquisition Date at the moment, which is a nice start, but I do really hope they’ll start adding more Cohort Types. Just the Acquisition Date is really not enough, for me at least. I’d love to see cohorts of people buying a specific product (category), for instance.

#Summing up

The cohort analysis can definitely give you some insights that weren’t readily available before. However, it does still require more than just basic knowledge of Google Analytics and might be a little confusing in the beginning. So I’m not completely sold on this feature yet, but to be fair; it is still in beta, so who knows how much better it will get right?


Other Useful Links : Use this spreadsheet for churn, MRR, and cohort analysis
Source :


  1. website design studios
    Great! We will be connecting to this enormous post on our site. Continue the good writing.

  2. They delivered a website that's been well-received both internally and externally by top web design firms in the world; it has successfully increased the number of prospective recruitments.
    user experience expert

  3. Pretty! It was really a wonderful blog. Thanks for the provided information.
    free url opener for mozilla

  4. Responsive and transparent, website design firms San Francisco meshed seamlessly with the internal team.

  5. Awesome work! That is quite appreciated. I hope you’ll get more success.
    UI UX design firms


Share Your Inspiration...