Market segmentation is a valuable technique to understand the commonalities between customers.

Knowing current and prospective customers’ similarities and differences allows an organization to introduce new products or modify existing ones, and more effectively communicate and sell to the customers most interested in those products.

We’ve conducted a number of market segmentations for clients and use advanced statistical techniques to identify clusters. But what leads to a successful and actionable segmentation analysis is less about using the right statistical algorithms and more about using the right data to identify the segments.

Don’t Rely Heavily on People and Product Attributes

It can be easy to think of people based on attributes: gender, age, ethnicity, education, and income. It can also be easy to think of products based on their attributes: price and features.

This attribute data is easily accessible and in many cases already collected. Consequently, these attributes are among the first variables used in segmentation.

But being easy to collect and conceptualize doesn’t make them the best variables to define segments.

In fact, pioneering researcher Clayton Christensen argues that reliance on demographic attributes is one of the main reasons around 40% of new products fail[pdf] to meet their commercial objectives—if they even get to market at all.

Focus on Needs

Instead, market segmentation should focus on the common needs of the customers and address why customers buy a product or service. Figure out what problems people have and segment around the problems (and possible solutions).

For example, in the Innovator’s Solution, Christensen talks about failed market segmentation for a fast food chain selling milkshakes. An initial segmentation built a profile around the typical milkshake customer, aggregating the most common demographics and psychographic attributes of people buying milkshakes. With this target segment, the marketers then identified what attributes of the product could be enhanced, such as price, flavor, and thickness (doesn’t it make you want a milkshake?). But the results of the segmentation failed because those product changes didn’t address customer needs.

A second analysis involved observing people buying milkshakes in the restaurant, noting what time the milkshakes were bought, along with what other food was bought and where the milkshakes were consumed. The researchers made the surprising discovery that most milkshakes were bought in the morning, without other food, and consumed in the car. It turned out a lot of people used the milkshake as breakfast to satisfy the challenges of a long dull morning commute. The milkshake did a better job than donuts, bagels, or breakfast sandwiches to address the particular needs of driving: It gave customers a free hand, lasted through the commute, and didn’t make a mess.

The researchers noted that the second most frequent instance milkshakes were purchased was in the afternoon by exhausted parents satisfying the dessert-treat needs of their children. The milkshakes competed with purchases of cookies and dessert but the milkshakes often went half-finished and were an expensive add-on.

The likely reason the first segmentation analysis failed was because it aggregated attribute data from morning commuters and afternoon parents. This led to a one-sized-fits-none product strategy. What was needed was a better understanding of the job the milkshake was hired to do.

For commuters, it satisfied the challenges of a long commute (one-handed, lasted long, and filling); for tired parents, it satisfied the challenges of looking to appease (quick, small, and sweet). Both groups had different needs, so marketing and customizing a better product for each group is key.

Knowing the job a product can do is a faster way to innovation and the better way to segment your customers.

The same idea of focusing on needs instead of attributes can be extended to almost any product or service experience such as financial software and online grocery shopping websites.

Financial Software

When designing better financial software, focus on the needs and what problem the software features might solve instead of focusing on the attributes of the people and products.

While the demographics of the small business owner and medium enterprise customer may look similar (older, male, college degree, and income above $75k), their needs almost certainly differ.

The small business owner is likely “hiring” the software to solve the problems of disorganized spreadsheets, nagging bookkeepers, and CPAs and looking to reduce paperwork. Conversely, the medium enterprise customer likely cares more about compliance, integrating with the sales software, and improving the payables workflow. A successful segmentation identifies clusters around these needs and product features; marketing communicates solutions to these needs.

Online Grocery Services

The typical profile of the online grocery shopper is a Generation X woman with kids.

But does targeting and tailoring an online grocery service for demographics (female, Gen-Xers with kids) pay off? Customers likely “hire” an online grocery service to solve the problem of limited time. It’s a convenience for people who don’t miss picking their own produce or clipping coupons.

These needs are the likely drivers of interest in the service and while many in this group happen to be women between certain ages, those demographics are not the root causes of interest.

A Better Way to Conduct a Market Segmentation

Based on the advice of Christensen and our own experience with defining product requirements for clients, here are some recommendations on conducting a more effective market segmentation:

Don’t rely heavily on attribute data. Minimize relying on people and product attributes like demographics (age, gender, education), price, and features. These are often the symptoms and not the cause of why people do and don’t use or purchase a product or service.

Identify the needs. Use a mix of methods to understand the struggles and problems customers are trying to solve—why they’re “hiring” a product.

Conduct a top-tasks analysis. A top-tasks analysis, by definition, forces you to think about what customers are trying to do. It then helps identify the essential tasks (or jobs) that customers are “hiring” your product for. It can be included as part of a larger segmentation survey.

Consider a contextual inquiry. Observing how customers use the product and in what context gives you ideas on the “jobs” customers have, even when customers can’t articulate them. A similar technique led to the breakthrough of realizing customers buy milkshakes for their commute.

With the right needs identified, the right data can be collected and then used to statistically identify market segments. While a good market segmentation is no guarantee of product success (executing the right features and marketing is essential), it can help avoid the common mistake of using attribute data that leads to a one-sized-fits-none strategy.