How B2B SaaS Companies Can Fast-Track Growth Using “Cohort” Analysis
B2B companies studying their businesses through “cohorts” will drive actionable insights and improve growth.
Every SaaS company I talk to wants to grow faster, be more profitable, and accurately predict its future. Beating goals is amazing, and unexpected revenue misses are really painful. Not surprisingly then, companies invest a huge amount of management time and energy into forecasting and revenue analysis, and generally trying to figure out how to just be better and build better products.
Like the enigmatic advice given to Dustin Hoffman in The Graduate, I have just one word for you: “Cohorts.”
I’ve talk to a lot of companies and one thing I notice is that the best companies understand their businesses through cohorts — grouping customers by age and monitoring them over time. Studying their business by cohorts enables companies to peer into the future before it happens, and proactively take steps to improve growth, drive actionable insights, and avoid costly mistakes.
But most companies, especially in the B2B world, still do not think or act this way. Instead, they tend to look at things “in period” which is easier, and is how most reporting systems work. Relying only on traditional in-period views will mask the leading key indicators, which can truly give you guidance on what’s going on with your SaaS business; hiding the truth and leaving growth on the table. Or worse, exposing your company to nasty surprises.
I’ve found that most of the information out there about cohorts focuses on product usage, and many B2B executives are not that familiar with the approach, as it applies to revenue and marketing operations.
Below is an intro to cohorts in the B2B world. In a later post I’ll give some examples. Feedback welcome including, any examples you care to share!
What is a cohort?
In the business context, cohorts are groups of customers, users, prospects, or leads that originated in the same time period. By grouping them together in this way you can see how they behave over time, throughout their lifecycle as they age. And you can very usefully compare cohorts to each other to understand the unit economics of your business model by normalizing differences in customer age. This provides a truer way to calculate key metrics find actionable insights, and trends like average customer size, adoption rates, lifetime values (LTV), retention rates, expansion rates, and more. Cohort-analysis is a standard practice in health sciences and economics, but in business it is less common. In the business world, successful direct-to-consumer companies are more likely to be utilizing cohort analysis, while it is still relatively rare in B2B companies. But as B2B businesses increasingly sell on a subscription or usage basis and rely on account expansion as a growth strategy, cohort analysis offers a great opportunity to boost performance and reduce management stress.
To better visualize cohorts and what they mean, think of a cohort as a crop of corn plants that you plant at the same time. The number of corn plants is the size of the cohort. The age of the cohort is how much time has passed since you planted the corn. Even though they were planted at the same time, we know some plants will grow faster than others, some will produce more corn than others, and some plants will die before they can produce any corn at all. So the yield from the crop (ie, the cohort) is a function of how many plants you started with, how many survive, and the average yield per plant.
Using cohorts to see into the future
Let’s say you could plant and harvest corn year round, and each month you plant a new crop, and after it is 6 months old each crop continuously produces corn. Each month would define a different cohort.
Like any good farmer, you are going to experiment with your crops to see if you can get more yield out of them — different seed stock, fertilizer, more water, etc. But you are impatient, and 6 months is a long time to wait to see what is working and what isn’t. You want to know as early as possible whether your younger plants are on track or not.
We would naturally expect older plants to be taller than younger plants, because they’ve had more time to grow. So the fact that your new plants are shorter than their predecessors doesn’t really tell us much. We need to know if they are taller or shorter than previous cohorts when they were the same age. We also would want to know if they are surviving at the same rate. At the two month mark if they are 30% higher, and 15% more have survived compared to the previous several cohorts, that is good news, and we can reasonably predict increased yield from that cohort without having to wait for the harvest. This is an example of how cohorts can help you peer into the future and create actionable insights.
The corn analogy has probably reached its limit, so let’s quickly translate back to customers and prospects:
Corn seeds → Prospects
Corn plants → Customers
Crop → Customer cohort
Crop survival → Customer retention (the inverse of customer churn)
Corn yield → Revenue
Marketing → Their job is to find the seeds
Field sales → Their job is to plant the corn
Customer Success Managers → Their job is to grow and harvest the corn
Cohort views versus traditional “in period” views of data
To contrast the idea of a cohort, consider the alternative used in most business reporting, which is the “in period” view of the data. For example, when you look at total revenue for the company by month, this is an in-period view. A cohort view of revenue would be monthly revenue from the customers who signed up in January 2019 (ie, the January ’19 cohort), in the month of February, March, and so on. In period views are more popular, because that is the standard for financial reporting, and is how most reporting systems work (unfortunately few handle cohort data well, if at all). Ideally all versions of revenue data trend upward, but it is common that in-period views and cohort-based views tell different stories.
For example in July 2019, let’s say total revenue has been growing month over month and the company is on track to beat its plan. This in-period view gives us a big thumbs up that we are doing all the right things! However, the cohort view reveals that the March cohort is a little smaller than planned, and April and May are not ramping product usage as much as previous cohorts. The shortfall may be imperceptible, compared to total revenue and is easily missed using an in-period view, but it could spell trouble later in the year. Using a cohort view, we discover shortcomings and can take proactive steps to correct or dig into the problem, rather than wait for a more obvious shortfalls, when it is too late to do anything about it.
There are many other actionable insights that cohort analysis is great at uncovering, such as:
- Is my lead quality improving as a result of changes to my marketing mix?
- Is my land-and-expand strategy working as well as it should be?
- What are the true win rates for new business and how are they trending?
- What is my true customer retention rate? Is it trending positively or negatively?
- Is my average monthly revenue per customer increasing or decreasing?
- What is the lifetime value (LTV) of my customers and is it improving?
- How have changes to my business impacted my key SaaS metrics? (Cohort analysis as a means of measuring major GTM changes and experiments, such as pricing, contract terms, and service models.)
Bottom line: In-period views are easier but are limiting and can be hugely deceptive. With a little practice, cohort analysis usually provides better answers and can get you more actionable insights.
*************************************
As a tech industry executive and consultant with over 20 years experience in product development, marketing, and sales, I love using data to make better decisions. Join me @kechley or on LinkedIn: https://www.linkedin.com/in/aaronkechley/
