Why Audience Segmentation Isn't All It's Cracked Up to Be

Segmentation Isn't Fit for Targeting Any More than a Screwdriver Is Fit to Drive a Nail

By Published on .

Good customer segmentation is important, but it's not the starting point of a targeted marketing communications or media plan, especially today when the clarion call is for marketing efficiency. Today the combination of available data, big-time processing power, asset management and marketing automation technology means we can really target individuals at scale, using business rules and decision-ing logic to trigger the "right time, right place, right message to the right person" ideal we've been making conceptual PowerPoint slides about for decades.

To target is "to select as an object of attention or attack." While "attack" is a little severe as a marketing term (albeit less disingenuous than "engage"), it means selecting something with precision, assuming a high likelihood of a particular outcome. Given the latter clause, I'm struck by how many marketers have conflated segmentation with targeting. "Should we create a contact strategy for the Power Millennials segment?"


Segmentation really isn't fit for targeting any more than a screwdriver is fit to drive a nail. It is a descriptive model; it classifies the base into groups with different characteristics. However, static classifications have very little to do with purchase propensity. Case in point, I'm writing this on a plane traveling at 30,000 feet and probably don't have much in common with most of the 150 others around me. However, we all needed to get from Los Angeles to New York today and bought a ticket on the same flight.

Targeting segments is also a poor experience. There are 41 solicitation emails in my personal (AKA "junk") account, and it's only 8 a.m. I'm still 1,500 miles shy of my next oil change and have an 8-mile commute, yet one of those email solicitations is for an oil change discount offer that expires in 10 days. Oh, and I just returned from a long vacation that used up most of my PTO, though I sure do wish I could take you up on that special 21-day Southeast Asia tour deal before the 15th.

And to my dear banking institutions, I am truly flattered to be deemed such an attractive potential customer, but I don't need a new checking account and I love trees more than getting lots of mail. That said, if you could send me an offer to make my cash assets work really hard in the six months or so before I attempt to buy my first home, I might move some money and maybe get my mortgage with you. (I know I've left a data trail of app downloads, loan calculator visits and keyword searches.)

On the other hand, Amazon.com doesn't need to know who I am for their algorithm to work (whether or not they do know who I am is another issue). When I visit the site or app, I fit the persona-less target definition of "people who put X in their cart in the last Y days." Even with this generic logic, my recommendations are just for me, because the company uses data, math, logic, modular design and interchangeable assets. I buy a lot of stuff from Amazon.

It is now possible to power almost any addressable medium to target in this scalable 1:1 way, whether your objective is acquisition, conversion, loyalty or advocacy. Buy data, borrow data, create experiences that generate data and start doing real targeted marketing.

What about your segmentation? Use it to define versions of the assets and content that will be deployed dynamically if you want to customize treatment for prominent segments within your target population. But target people who exhibit a behavior, experience a specific situation, exist/arrive in a particular need state or at least are predictively modeled likely to do one of those things. Your customers and your KPIs will thank you.

Most Popular