Nailing Your Cohort Analysis

Cohort analysis - Much discussed, much written about, but honestly, pretty simple.

It’s so simple that you can get an excellent assessment of your business in one single visualization. Green, yellow, red -- Just like a stoplight.

We’ve brought together some of BuildGroup’s “thought leadership” in the cohort analysis department to bring to you a simple explainer and, of course, some templates.

We won’t make you scroll all the way down for the template. (But read the explainer. Seriously. Read it!)

Download The Template

What is a cohort analysis?

• A cohort analysis segments your customers into groups and assesses the performance of those groups over time.
  

• The conventional segmentation is the time at which a customer joined, be it a month (January 2020), quarter (Q1 2020, AKA “the corona quarter”), or year (2020, “let’s not do this again”).
  

• You could also potentially segment by vertical or by channel. Could be interesting!

So...what does that look like? What do I need to build this?

• You need revenue data by customer by month (incidentally, that’s one of the ingredients for our #1 hit, the KPI Dashboard).
  

• You can build your own spreadsheet - an excellent exercise for aspiring SaaS CFOs, VPs of Finance, and BuildGroup Interns.
  

• Or you can use our template! ​
  

• It will sort customers into cohorts:
  ○ It will identify what the initial revenue generation was for that cohort in its first time period
  ○ And then it will analyze the revenue performance over each subsequent time period, as a percentage of the initial revenue generated.

Let’s walk through a very simple example first.

Here is a sample customer set, with revenue by customer by quarter (for simplicity).

Customers have been grouped by the shared attribute of being onboarded in a specific quarter, and appropriately color coded:

• Customers 1,2,3 are in Cohort 1
  

• Customer 4 is in Cohort 2
  

• Customers 5 and 6 are in the Cohort 3 Customer 7 is in Cohort 4.

This is a cohort analysis output (we’re color coding the cohorts so you can see how they map to the individual customers).

We’ve provided this small example and a larger dummy data set for you to play with here.

What am I looking for here?

We look for strong cohort performance:

• Cohorts generate at least 100% of the initial revenue generated over subsequent time periods
  

•  ...and they do so specifically through a renewal cycle. Usually this means 100%+ year over year
  
  
• This means that your product/service is valuable enough to renew, at least, and hopefully to upgrade as well. Perhaps your customers love your product so much they purchase additional seats for employees, or upgrade to your premium offering. Your feature roadmap rocks. Heck yes.

If your cohorts are shrinking or performance is uneven​:

• You’ll need to take a hard look at product, pricing, and value proposition. Basically, something’s not right. You’ve managed to get them through the door, but something about what you’ve offered and what you’re delivering doesn’t match up:
  

• The product is not essential
  

• The product is potentially essential, but does not meet the promises made
  

• The product is episodic in nature (maybe you shouldn’t price on a SAAS model?)
  

•  Product onboarding or customer support is not working

Some things we notice:

• Perhaps early cohorts don’t do well, and your later ones have improved. That’s okay. Sometimes it takes a bit of time to figure out what your sweet spot is.
  

• Cohort revenue concentration. Sometimes, you have a really big quarter or month, and they’re doing the “heavy lifting” for your business.

Big picture on cohorts.

There’s a limit to scalability of ARR if net retention is less than 100%. You’re building a leaky bucket that gets harder and harder to top up if your net retention isn’t great.

One of our lead Builders, Klee wrote ​here about the churn vampire​. It’s accurate. I don’t want a leaky bucket.

How can I improve retention?

Don’t let customers leave, like a B2B SaaS-themed house of horror. Kidding!

The main question you have to ask yourself is, “how valuable is my product to customers?” If you make their lives dramatically easier and it’s used everyday, (e.g. slack, zoom), they will stick around and pay you more (upsell!).

If you can answer “very valuable” to the main question, you’ll need to understand why customers cancel or downgrade. Sometimes, the answer isn’t obvious so we recommend doing a customer survey, as a starting point. We ascribe to the school of NPS and more importantly, understanding why the NPS score is high or low.

We’ll conclude with a visualization of Slack’s cohort retention, as an inspiration. Cohort #goalz

‍