.png)
If you want to gain a deeper understanding of your customer data, you may find a cohort analysis to be a valuable asset.
Cohort analysis is a framework that allows you to organize your customers and understand all the ins and out of their behavior. It allows you to make accurate business decisions on each type of customer, instead of bunching all of your customers under one assumption. Since customers are essential in the SaaS and subscription business, you’ll want to know every opportunity possible regarding increasing retention and LTV and reducing churn.
First, let’s break down cohort analysis. A cohort is defined as a set of people with a common characteristic. In SaaS, a cohort could be made up of customers who share a common denominator such as acquisition date, pricing plan, or geographic location.
Cohorts are then tracked over a set time period and compared against SaaS benchmarks to gauge performance of your business. Analyzing your cohorts will put your data into context and make you aware of what is and is not working for your business.
One of the most common approaches for a cohort analysis in SaaS includes measuring either customer retention or churn against a cohort characteristic. However, there are endless possibilities of cohorts and benchmark comparisons available to help you dive deeper into your data.
.png)
There are countless combinations that you can subgroup your customers into as long as the cohort users share a common denominator.
Some of the most common SaaS cohorts are:
Acquisition Date - How does seasonality or timelines impact business?
Acquisition Source - Where does the lead come from? (Event, sales, paid ad, newsletter, referral, social media, etc.)
Device Source - What device was the visitor using? Desktop, tablet, or mobile device?
Geographical Regions - How do geographical regions impact customer engagement?
Customer Size - How does the product perform for individuals vs. large firms?
Pricing Plans - How do pricing models impact churn?
Sales Cycle Length - How does sales cycle length impact business?
Customer Size Example:
The example above is formatted with cohorts depending on the size of the customer and the number of users within the company ranging from individuals, small businesses, to large corporations. Here, we are measuring customer retention rates over time against the size of the customer. Under the # column you see the amount of newly acquired customers from each size cohort and their corresponding retention rates after each month after acquisition. If we look into the 25-100 size cohort, we see that 87% of the original 63 customers acquired remain after 4 months - that's 54 customers left.
Customer size cohort analysis is a great way to gauge what audience your product will be the most successful based on the size of customers. In this example, we see that there is about a 20% difference in retention rates between smaller size customers and larger customers. Based on the analysis, this company is much more successful in retaining customers with a company size of 25 and up compared to customers with a size of 25 and below. At this point, you can decide to adjust your sales and marketing strategies to cater toward higher retaining cohorts.
Cohort analysis helps you see trends, which are helpful when making decisions for your business. You can use a cohort analysis to see the direct impacts that cohorts have on retention, churn, and customer lifetime value (LTV).
Customer Retention and Churn
Given the nature of the SaaS business model and recurring revenue, SaaS businesses depend on sustaining and retaining their customers to maximize revenue. The metrics that correlate with this would be customer retention and churn. The goal is to keep retention high and churn low.
A cohort analysis will allow you to see which cohorts meet this goal. Depending on how you set up your cohorts and format them, you should be able to analyze who, why, and when customers are churning.
For example, creating cohorts based on user location will let you analyze how your business is doing in different regions. Through a cohort analysis, you may realize that a certain location has higher churn compared to another. This answers from where customers are churning and you will be able to make effective business decisions to reduce churn in that specific region.
The same goes for using cohort analyses on customer retention measurements. If you group your customers based on the service plan they receive (original package vs upgraded package), you will be able to see how each different type of service plan impacts retention. Cohort analysis will allow you to see which product features retain the most customers and highlight your customers’ needs. You’ll be able to identify your most successful packages and make decisions accordingly - maybe it’s time to get rid of that package plan after all!
In SaaS, the customer’s lifetime value is important when talking about profitability. You want your customer’s average lifespan to be longer than your CAC payback period since all revenue acquired afterward is profit from that one customer.
Cohorts with high retention rates likely means their LTV is higher, which in turns means profitability. Whether you’re a CEO or CFO, you have a say in how and where to allocate your resources and money. Cohort analyses can be great indicators of which cohorts to double down on investments and focus your time toward.
The beauty of a cohort analysis is the ability to isolate factors and see the direct impact of those variables on your company. When analyzing your customer’s LTV, creating cohorts of individual customers vs enterprise customers will highlight the differentiation of LTV between the two entities.
.png)
We already know how cohort analysis allows you to make strategic business decisions based on data. However, it can also serve as a platform to test some of your business ideas. You can test how implementing certain product features, pricing plans, or marketing strategies will have an effect on churn, retention, and overall your business.
Creating an experiment with cohort analysis is as simple as asking a question on how one cohort will impact one benchmark or metric:
How will this product feature impact retention?
How will introducing a new pricing model impact my churn?
Would a custom on-boarding program lead to improved retention?
Once you have your question, you can test it and collect your data to help you make an informed conclusion about the cohort and its impact on your business.
For example, suppose you were interested in increasing your prices for your product. Your cohort would be the customers who are in the new higher pricing plan and the metric you would be measuring is churn. After tracking the churn rates in this cohort over some months and comparing it to the churn rates of the cohort with the original pricing plan, you should be able to see the impact of increasing your pricing plan. If you see a significant increase in churn in the higher pricing plan cohort, you may conclude that increasing your prices will cause customers to leave your business.
Correlation does not mean Causation
In any experiment, creating a controlled environment is key to determine direct causation - but this is a very difficult task to do. Isolating your variable from any outside factors is tricky, especially in businesses when there are different components that can impact success.
When formatting your cohort analysis experiment, try to account for confounding variables that may impact your results. For business experiments, it's better to err on the side of caution and remember correlation does not mean causation.
Even though correlation does not mean causation, cohort analysis experiments are a great indicator of your business performance and will lead you toward effective decision making.
Cohort analysis is all about making data-driven decisions utilizing the additional context that is given through a cohort. Still, there is always a piece of the story that data cannot shed light on - that’s where segmented cohort analysis comes into play!
Segmented cohort analysis adds another layer into your data for you to further investigate.
There will be more context to your data, allowing you to have a better understanding of your business and the impact your cohorts have.
Let’s say you want to evaluate your marketing strategies and see if you should keep investing toward paid acquisitions. You plan to run the same paid advertisement on LinkedIn, Facebook, Twitter, and Instagram. Instead of simply tracking your retention rate of the new customers who were acquired through the paid advertisement, you should segment each customer by platform from where they viewed the ad. Segmenting your cohorts by social media platform will allow you to see how effective paid acquisitions are overall, but also the success of each platform specifically in retention. If you notice your Twitter cohort has the lowest retention rate, this might be time to reallocate investments away from Twitter and double down on your higher retention cohorts, such as LinkedIn.
Cohort analysis is a must for any business. There are plenty of resources available to help you get started such as Kissmetrics, Amplitude, Google Analytics, Excel, Google Sheets, and more. There are even dozens of cohort analysis templates available to guide you.
However, you must remember: your business is unique - your cohort analysis will be unique as well. Make sure to pick cohorts and segments that are key for your business. When using templates you found online, try to tailor that set up so it fits your company’s specific needs.
Pssst... KPI Sense can build you tailored cohort analyses along with all of our modeling, reporting, and dashboard services!
.png)
Cohort analyses are beneficial… only when they’re done right. Watch out for these common mistakes:
Correlation does not equal Causation - as mentioned before, keep in mind that the trends a cohort analysis produces may be factors of correlation and not causation. Cohorts add context, but it's not a complete picture.
Not differentiating Annual vs Monthly plans - when measuring churn, make sure your monthly and annuals plans are separated or else your data will be skewed!
Confounding Variables will skew observations - you may not be able to control for all outstanding factors, but keeping this mind when making business decisions will save you a lot of time and confusion.
Miscalculating retention or churn - your data tracking and calculations set the tone for how accurate your analysis will be. Consistency and discipline are essential when it comes to creating your cohort analysis.
If there’s one thing to take away from this post, it’s that you should be using cohort analysis! The ability to make data-driven decisions starts with cohorts and isolating variables to see the real impact they have on your business.
As much as a cohort analysis is beneficial, setting up your cohorts and tracking your data is just as time consuming. To help with this, KPI Sense makes customized cohort analyses with accurate trends so you can focus on what really matters. For more information click here.