Actionable Segmentation Solutions: All the Little Things

Making a Great Segmentation: Wizardry or Wisdom?  

Segmentation often conjures up images of fancy math and indecipherable algorithms. At times, non-analysts are more inclined to wave away the explanations and call it magic than to dive in. It’s as if a wizard lurks in the background, brewing up a mysterious solution of segmentation in a cast-iron cauldron.  

It’s silly, but as one of these so-called “wizards,” let me share a secret. Any analyst can run a clustering algorithm, but producing a great segmentation is about so much more than the algorithm or the parameters one chooses. Yes, the algorithms are important, and the analyst must know what they’re doing, but that’s not enough. It’s only one link in a long chain of factors that must be properly executed to produce a great solution.  

Segmentation Strategy Alignment: Starting with the Business Objectives  

Before a single question is written, it’s important to align with the client and their internal stakeholders across the organization on the business objectives and outcomes of a segmentation. From marketing and sales teams to operations, product development, and beyond, hearing from all groups who plan to use the segmentation is important. If this step is missing or improperly carried out, the greatest algorithm in the world won’t make up for it. You’ll end up with a perfectly crafted waste of time.   

To avoid this, we begin every segmentation with stakeholder interviews. In these interviews, we discuss the business objectives of each stakeholder, how their departments plan to use the solution, what they hope to learn, and who they believe their potential customers are. We then use these discussions as guiding lights throughout the entire process. Clarity at this stage helps ensure segmentation success and actionable output.   

Discovery: Uncovering what is known, assumed, and unknown

After alignment, we move to discovery. We begin with an information inventory, in which we pore through the client’s previous research and customer data to uncover relevant attitudes and behaviors. We undertake secondary research to better understand the category and ground ourselves in industry trends. This foundational review is also paired with qualitative interviews to capture the consumer’s current perspectives. 

Naturally, there are times when the client’s viewpoint differs from the viewpoint of their customers. These gaps aren’t things to shy away from. We prefer to lean into these, as often times a successful segmentation solution will be built and activated to close these gaps. Without this early discovery phase, we run the risk of designing a solution around the wrong topics and missing critical information about our client’s customer base. Again, no algorithm in the world will make up for this mistake. 

Survey Development: Identifying the Attributes that Differentiate Customers 

Once we’ve aligned with the client and discovered what’s important to their customers, it’s time to draft a survey. Obviously, you can’t throw everything into a questionnaire and hope for the best. If you ask too little, you’ll miss the nuance that often differentiates customer groups. If you ask too much, you’ll exhaust the respondent and get substandard data. A solution is only as good as the data you use to create it, so it’s important to be highly selective when it comes to the survey.  

To strike this balance, we start with a long list of potential attributes and run tests in our proprietary panel to weed out highly correlated or non-differentiating attributes. We then discuss the remaining attributes with the client to finalize a survey that’s manageable for our respondents and aligns with the client’s business objectives. The best segmentation surveys include unique attributes that divide respondents, helping to create meaningful groups.  

Data Collection: Building a High-Quality Segmentation Dataset 

To emerge with reliable segment data, you must also control data quality during collection. Again, no matter how good the algorithm is, there’s no way to make up for bad data. If you slack on this step, you could end up with one segment of straight liners who both “take their time making decisions” and “make quick decisions without much thought.” Trust me, that’s an uncomfortable group to explain.  

To address this, we heavily monitor our sample. We include red herrings, ask contradictory questions, and use various statistical methods to identify near-straight-liners and outliers. We look at completion times to remove speeders and review open-ended responses for thoughtfulness. We set the bar high, but it’s worth the effort to have good data.  

Solution Selection: Creating Realistic Customer Segments 

After the analyst does their work, there are usually several segmentation solutions with strong mathematical properties. But mathematical strength alone does not make a solution viable. The model still must produce distinct customer groups the business can use. 

Before an option reaches the client, it must clear four tests. It must be: 

• Reproducible: meaning stable across reasonable variations in the data 
• Sizeable: with each segment representing an addressable portion of the client’s customer base 
• Differentiated: with every segment meeting a minimum level of differentiation from every other segment 
• Actionable: so the client can do something meaningful with/for each segment 

These criteria narrow the field to a set of defensible, viable solution candidates. 

Math can get us that far, but it can’t choose among those candidates based on real-world usefulness. That is where business judgment and business acumen take over. If you simply choose the solution with the strongest mathematical score, you risk overlooking the one that makes the most business sense. The analyst’s job is to recommend a set of viable candidates, and then the full team and the client determine which one best fits the business.  

Once we reduce the list to three or four viable candidates, we set up a half-day working session to review the options with the client. In these sessions, we discuss each candidate in detail, weigh the tradeoffs, and agree on the one that makes the most sense for the business. While we always arrive with a recommendation, it’s surprising how often the client sees value in an alternative. Our firm belief is that the client knows their business best. It’s our job to explain the tradeoffs and push back when necessary, but ultimately we trust the client’s judgment. 

Activation: Turning Segments into Business Strategy

One of my favorite parts of the process is the activation workshop because this is when our solutions come to life. You see, my team could create the greatest solution in the world using the very best algorithms, and we could still completely fail if that solution remains unused. Our job isn’t done until the solution is socialized throughout the client’s entire organization and each stakeholder is using the tools we provide to their fullest potential. I’ve been in far too many client discussions and heard the statement, “We have a segmentation, but no one ever uses it.” What a shame! Spending all that time and effort on something that’s never activated. But an unused segmentation is not merely an activation problem; it’s an unsuccessful segmentation.  

To address this, we developed the activation workshop. In this working session, we meet with each department of our client’s organization to socialize results, provide thought starters, brainstorm use cases, and help the stakeholders come up with products, marketing, messaging, or customer service strategies based on each segment. You’ll recall that our first step is alignment with these stakeholders. It’s only natural that our last step would bring us full circle as we hand off the solution to the same group, along with the necessary tools to use the solution for years to come. 

The Real Magic of Segmentation: Collaboration 

I hope this sheds light on the many important things that must be done correctly to produce a great segmentation solution. I also hope you noticed how little of this actually comes from the analyst. It comes from a dedicated research team, solid client relationships, engaged stakeholders, and careful planning.  

Ultimately, great segmentation solution is a team effort. It requires alignment around a clear outcome, a thoughtful process, and a shared commitment to getting it right.