Stop Calling It Segmentation

When Words Get In The Way…

Decision-First AI
Corsair's Business
Published in
4 min readOct 11, 2016

--

Segmentation is a big word for “cutting things into little pieces”. It really isn’t the right noun for the process known as Marketing Segmentation or Customer Segmentation. Analysis is a hard word for “cutting things into simpler pieces” and has origins in common with “loosening” — so it may be mildly better, but still doesn’t describe the process.

Segmentation is synthesized — which is a big word for built or created. It is also a process of recognition and discovery.

Stereotyping is a flawed word on many levels, its etymology is overly complex and confused (not surprising that its modern meaning also reflects that). Profile is little better, as it is an outline — hollow and thin. Type is better. It means form or kind in Latin. In Greek it means to strike — thus type-writer. Eventually it became symbol or emblem. Latin also gives us model, from mode, manner, and measure. Now we have something…

Models and Types define tools that are symbolic and measured. They speak to manners and forms, simple but descriptive.

In model theory and related areas of mathematics, a type is an object that, loosely speaking, describes how a (real or possible) element or elements in a mathematical structure might behave. — Wikipedia

The mathematicians got this right. So why have so many marketers and salesman screwed this up?

Honestly…

… just doesn’t roll off the tongue. It isn’t very sexy and to the less informed, it doesn’t seem very meaningful. So rather than embrace a terminology that is succinct and allow the power of the tool to build the mystique and interest, we embrace big words that confuse and obscure. Segmentation sells! But only in the short term…

Introducing MTMs — A new way to recognize and target the types and kinds of people who distinguish your market and customer base, utilizing a model that is simple and measurable.

Was that so hard? Still missing the sexy? MTMs could still have all those sexy marketing adjectives, the ones that still seem overly complicated but at least they are descriptive. We could have:

  • Behavioral MTMs
  • Demographic MTMs
  • Psychographic MTMs

Personally, I prefer Actionable or Effective MTMs. Feel free to go with whichever adjective makes you and your marketers feel better — so long as it is accurate.

MTM is a better term because it better describes our aim and our process.

MTMs are built by discovery and recognition of key manners, forms, behaviors, or attributes of the individuals in our population. We don’t look to carve up a large pool into smaller segments (which despite claims to the contrary may prove more difficult to manage… there is less volatility in large populations). We look to build (or aggregate) important types of people. These types become the elements of our model.

These types effectively segment similar types of people. That similarity should be far less forced and ultimately far more likely to make the group relatable, understandable, describable, and yes — marketable and manageable. The latter will only be confirmed once adequate testing and measurement are performed.

We also don’t attempt to stereotype every individual into a neatly sliced segment. Those who don’t fit are collected into a type characterized by their difference from everybody else. In the end, like all models, the MTM becomes a tool for learning, measuring, and adjusting. So many marketing segmentations become so complex and convoluted in their attempt to force every member into a pre-defined segment, learning becomes impossible (understanding as well).

MTMs are humanizing. They identify virtual communities of customers. These communities may not formally interact but they share common attributes, ones identified by humans that find them important. Machines are great at finding patterns (and should have a role in building MTMs) but we market to humans. We create experiences for humans. That makes humans uniquely qualified at selecting the attributes that should matter. Machines then get to tell us if we were right. They are good at that… but they aren’t very good at synthesizing… not yet.

So stop segmenting. Recognize the role analysis is supposed to play on either end of this process. Don’t stereotype or profile. Build a model that identifies the types of customers in your market in meaningful, human, and effective ways. Leave ambiguous wording to the marketers. Leave mindless cutting to the machines. Embrace MTMs and build your model for stronger customer relationships.

--

--

Decision-First AI
Corsair's Business

FKA Corsair's Publishing - Articles that engage, educate, and entertain through analogies, analytics, and … occasionally, pirates!