“Just because someone walks in and buys a high-margin good doesn’t necessarily mean you should prioritize that person as a customer.”Jason Stanley, Element AI Design Research Lead

In spite of your best efforts to acquire users who match your best fit customer profile, you’ll probably notice that:

  • Some customers are engaging much more than others.
  • Different goals and use cases have started to emerge.
  • Some customers are spending more than others.
  • Users from other segments with different needs are getting value from your product.
  • Your homogenous group of customers will start to appear less homogeneous.

As your business grows, it’s normal to start noticing a wide range of usage and consumption patterns.

You might have built your product with the aim of attracting a certain type of users, but if over time you notice that other profiles are getting more value from your product, you would be foolish not to adjust your targeting.

Amplified, today’s edge cases can become tomorrow’s best opportunities.

Researchers at EPFL University in Lausanne, Switzerland have concluded that 73% of startups ultimately pivoted to other markets once they realized that their initial market wouldn’t be able to support their growth.

Are you sure you’re focused on the right users and customers?

At this point, it makes sense to start looking at the composition of your customer base. You may already be leaving money on the table.

Finding the Segments in Your User Base

To start discovering the segments in your user base and do behavioral segmentation, you should define success along three key dimensions:

  1. Revenue: Customer lifetime value (CLV)—how much revenue is generated over the entire duration of the customer relationship—is often a good way to find your highest-spending customers.
  2. Engagement/Retention: A good way to evaluate engagement might be how often core product actions are performed, habituated users, or 7-, 14-, or 28-day retention rates.
  3. Referral/Word of Mouth: For word of mouth, it can be the number of referrals sent or completed.

Looking at the breakdown of those results, create three buckets:

  1. Your best customers: the top one percent in terms of engagement, revenue, and referrals;
  2. The next best: customers ranked within the 2nd to 10th percents;
  3. Your worst customers: the bottom 10%.

Depending on how technical you are, you might be able to find these users with SQL, a CRM, database exports, or by looking at people analytics in tools like Mixpanel, Amplitude or Intercom.

You should focus on these three groups because the top one percent represents your product’s “fans”, your advocates.

The next 10% gives you a good comparison point. It can help reveal low-hanging fruits.

The last 10% helps you define who you probably shouldn’t be targeting.

Learning About Behavioral Segmentation Through Customer Interviews

Randomly recruit 15-20 candidates in each of these buckets. Schedule 20-minute interviews.

You’ll want to understand who they are, what need(s) the product solves for them, and what value they perceive.

You can get started with the following questions:

  • How would you describe your role as [ Role ]?
  • What does success look like for you?
  • Why did you initially sign up for [ Product ]?
  • Did you evaluate other tools?
  • What was your previous experience with [ Solution Space ]?
  • What made you decide to [ Buy / Use ] the product?
  • Can you walk me through how you use [ Product ]?
  • What is the main value you feel you’ve received from the product?
  • Why do you keep using the product?
  • What is the main [ Problem ] you feel the product solves for you?
  • If we took away [ Product ] from you, what would be the things you miss most? What would you use as alternative?

Through these interviews, you are trying to evaluate different positionings.

What are the commonalities in the stories of your best customers? What use cases tend to lead to sustained product usage? What needs do your best customers have? What value do they seek? How do they explain your product’s core value?

It’s not uncommon for similar analyses to reveal that your best customers are actually quite different from who you thought they were.

Once patterns are starting to emerge, consider expanding your research by sending a survey to a larger group of customers. You can ask open-ended questions like those asked during the interviews. If you feel like the patterns are clear, consider providing multiple choice answers, always making sure respondents have a way to input free-form answers.

Example of Behavioral Segmentation

At LANDR, where I used to work, through a similar analysis, we noticed that occupations seemed to have a big impact on the CLV.

We used the basic occupation groupings that we had uncovered through interviews, and surveyed a large part of our user base to understand what kind of work they did.

Since we weren’t sure about the groupings, we made sure respondents had the option to add other occupations.

Although we weren’t able to get answers from all users, the much larger sample of respondents helped refine our behavioral segmentation.

Kieran Flanagan, VP of marketing at HubSpot, recommends using regression analyses—a statistical process for estimating the relationship between variables—to see what users in your best-performing segments did in their first week after sign-up, how they got onboarded, and how they ultimately became customers. The insights gained from these analyses will help you improve overall performance.

If you intend to use the segments you uncovered to create personas—fictional profiles based on interview data used to summarize information about customers—you should definitely validate them by surveying larger populations.

Once you know which segments perform best, you can make prioritization decisions. The fewer segments you target, the clearer your messaging will be, and the more certainty you’ll have that your pitch will connect.

– –

This post in an excerpt from Solving Product. If you enjoyed the content, you'll love the new book. You can download the first 3 chapters here →.

Categories: Customer Research Technique