The latest feature of Google Analytics, cohort analysis is a new asset of choice to integrate into your web-marketing toolbox. Cohort analysis can allow you to detect new levers to take off your website traffic, in particular by analyzing in-depth the evolution of user behavior on your site.
Cohorts: What are they?
Cohort analysis can seem confusing at first, and for good reason, the latter is rather unusual in the Google Analytics ecosystem. First, the cohort analysis is dynamic, unlike most of the features of Google Analytics. The analysis of cohorts is done continuously over time, unlike features such as the number of sessions, or page views, that you generally analyze over a given period of time. Second, the analysis of cohorts alone will not allow you to identify obvious growth levers: you will necessarily have to cross these analyzes with other KPIs to optimize the performance of your site.
What is cohort analysis?
If this definition still seems nebulous to you, don’t worry, let’s take the time to dive into this feature in more detail by taking a concrete example:
Suppose you are the manager of a clothing store. You decide to launch a promotional campaign by offering 20% discount coupons to certain customers. One of your customers, Martin, receives this discount voucher, and takes the opportunity to visit your site, consults your various items, and finally decides to buy jeans. As a manager, do you want to know how many customers like Martin will come back to your site after purchasing these jeans? How long does it take for these clients to come back to you? You ask one of your associates to analyze the behavior of Martin and all the customers who are like him. You have just created your first “cohort”. Some months later,
This is where the cohort analysis begins. Still in our example, imagine that, out of the 20% of customers similar to Martin who actually returned to your site, an extreme majority of them added a product to the cart, but stopped at the “Shipping costs” page. It seems that all customers similar to Martin have been held back by your shipping costs, so maybe you should run a “free shipping” type of campaign with these customers. Congratulations, you have just analyzed your first cohort, and activated a growth lever!
Three keys to understanding cohort analysis
Key 1: Define a cohort
A cohort is always supervised within a given period, during which a certain group of users have browsed your site, and/or committed specific actions. The first thing to do is therefore to define this period of time to be studied. A bit like a race, which has a beginning and an end, what interests us about our cohort is to know what behavior it had during this “race” on your site. More concretely, you could – for example – analyze the behavior of a cohort during an event like the sales. Beyond the period chosen, and the profile of your cohort, you can of course sectorize a little more by choosing to follow, for example, only users arriving on your site from Facebook, Google, or others…
Key 2: Define a latency
Now that our cohort is defined, in terms of period and persona, it is time to define the duration during which you will follow your cohort. If, for example, your goal is to understand the behavior of your users who return a month after their first visit: the duration of your analysis should therefore extend over one month. There is no right or wrong way to define the period over which you will follow your cohort, it all depends on your activity, the events that are taking place, the industry in which you operate … We advise you to do so which makes the most sense to you, based on your own expertise.
Key 3: End date of your analysis
If you want to analyze a cohort of visitors who visited your site between January 1 and January 7, with one month of latency: you will get your analysis results on February 7. Once this date has passed, the users you observed will no longer be considered as belonging to this cohort. Be precise in choosing the date on which your cohort analysis must end, otherwise, you will notice incomplete information.
Current limitations of cohort analysis
If the cohort analysis can reveal interesting behaviors to detect to guide your strategic decisions, this functionality of Google Analytics still has some limitations. The first is that it is difficult, at the present time, to determine whether several visits come from one and the same user or not. Ideally, you should individually track each user corresponding to the cohort in question, and analyze their behavior on the desired latency. Then you aggregate all that individual information to get a clear and precise cohort analysis.
Another problem stems from terminating your scan. Google Analytics determines that a user “comes back” to your site if they reconnect at least once 30 days after their first visit. Let’s say we start a cohort analysis the week of January 1-7, and this analysis ends on February 7. If a user visits your site on January 2, then returns on February 3, Google Analytics will not consider it as a “return”, since the user has returned 31 days after their first visit … This is a bias analysis that should not be neglected since Google Analytics will record a “new visit”, which is not really one.
How do I use cohort analysis for my business?
It is important to understand one thing before embarking on the cohort analysis: the information that you will collect, and the levers that you will activate thereafter will take time to produce their effects. Unlike a remarketing campaign, for example, where the effects are relatively immediate, the cohort analysis focuses on the behavior and habits of users. These are aspects of marketing that are more difficult to pin down. Between the start of a cohort study period, the analysis at the end of it, the launch of actions supposed to be beneficial, and the analysis of these actions to determine their real effectiveness: it could well take between 4 and 6 months for you to see tangible results.
If this period of time is quite unusual in web marketing, the effectiveness of a good cohort analysis should not be underestimated either! From our point of view, the cohort analysis can provide key information when it comes to measuring the effects of a promotional campaign on your activity, or even measuring the engagement of your consumers in the long term.
In e-commerce, you have to go fast, much more than in traditional physical commerce. Tools like Google Analytics make it possible to obtain almost real-time information that goes in this direction. However, you won’t last long if you fail to engage your consumers for the long term. As such, the cohort analysis is proving to be a key tool for immersing yourself in the behavior of your users and thus identifying areas for improvement that will allow your brand to establish itself over the long term. If the cohort analysis can be discouraging, due to its complexity and the time required to analyze and act accordingly, you should certainly not neglect it: a good mastery of this tool will be your best ally to improve your user experience.
Utkarsh is the founder of Esite Bucket. Esite Bucket is a leading software company that provides the best web development solutions in countries like the USA and Canada.