Know Thy Customer: When to Use Cohort Analysis versus Customer Segmentation

The marketer’s core responsibility is to know thy customer. Who is this person buying this thing I’m selling? Can I get them to buy it again? Are there other people like this person out in the world who might also want to buy my thing-a-magoo?

That’s the basis for market research and analysis, which is the driver of all marketing decisions — implicitly or explicitly.

Market research can inform your understanding of a person: who she is, what’s she like, what are her needs, how she acts. Historically, qualitative and quantitative research would be used to build what’s called a segmentation, or a breakdown of your customer base into groups of similar people. This was useful for a few reasons:

  1. Targeting: You can identify, based on measurable behaviors and attributes, what this group of people is like and try to go find more people like those in this group
  2. Communications: You can tailor your communications to different groups of people, because hey, not everyone is into the same thang
  3. Business Intelligence: You can measure the health of your business and even define your business along the lines of your segmentation. With a good breakdown, you can often predict growth patterns and even understand why you’re having a bad
  4. New Market / Product Development: You might identify an unmet need that you can capitalize on.

In the last few years, though, segmentation seems to be going out of vogue. At least in the startup world. Perhaps its because the field was historically dominated by expensive consulting firms who would say that you needed to pay $500k in 6 months to get a picture of your customers. (Segmentation, FWIW, can easily spiral down a path of “boil the ocean” analysis.)

Instead, a new kind of analysis came to dominate how businesses are measured and how customers are understood: cohort analysis. The methodology is great for understanding how your business grew to date and what the future might hold.

Typically, you look at a certain important behavior you want to measure, and group customers who took that action within a certain time frame. You can slice the rest of the temporal pie in equal chunks, and analyze and forecast based on those groups. It’s a very useful way overcome the difficulty of trying to understand your business in one homogenous sum.

Cohort analysis is great for certain applications, like:

  1. Communications: You can target communications to people who are at similar points in their journey with your business / platform / app
  2. Business Intelligence: You can see patterns in data that might otherwise go unnoticed, particularly anomalies based on things that happened in a certain time frame
  3. New Market / Product Development: Based on patterns in the data, you might learn that certain features fall flat or others are big growth drivers

Cohort analysis is critical for truly understanding how your product is working and whether this is changing over time. But it’s not the same as truly understanding who comprises each group. Cohort analysis is company-first analysis, segmentation customer-first. Both are critical, but they are not the same.

My big caution to all startup marketers with limited time and money to “know thy customer”: don’t think that knowing how your customers are engaging with your platform is the same as understanding who they are. And when in doubt, there’s nothing more valuable than talking directly to your customer


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: