I recently had lunch with some smart data scientists, marketers and general analytics folks when we started listing all the different ways we segment customer data, what kinds of data we search for when the answers are lacking, and how the wealth of data we have today can help us communicate with customers. 

What follows are the top eight we came up with.

8 Ways to Slice and Dice Customer Data

1. When You Know Nothing

Sometimes you have a product and no means to track your customers. For example, you are a State Lottery that doesn’t sell directly to customers, or a hot dog vendor. How then do you think about product innovation and marketing? 

The answer is good old fashioned market research. Going out into the community and running surveys and focus groups. While expensive, done right it will give you a good idea of who is buying your product and why. Back in the days before the data explosion, this was your best option for learning about the customer.

But for those with truly innovative, paradigm-changing products, people in the real world will lack the framework for conceptualizing and supporting the product before its release.

Henry Ford and Steve Jobs both created memorable quotes disdaining this type of research for this reason. However, in my mind, it was expressed most usefully by Peter Drucker in "Innovation and Entrepreneurship" when discussing innovation and new products:

“Market research does not work — one cannot do market research on something that does not exist.”

2. You Know Who Customers Are

After market research, another incredibly common segmentation technique is customer demographics. Age, race, gender, hair color, location, income, education are all examples of person describing, demographic data. 

While demographic data can be incredibly useful, such as when identifying trends that can lead to new business opportunities, it also requires care. Basing decisions on specific attributes is often illegal discrimination.

3. You Know What Customers Say

Customers will tell you all kinds of things if you know where to look for it. 

Whether it’s comments in forums, chats with customer services agents or general communications about products in the world of social media — your customers are speaking. 

Generally people only view this data as an opportunity to generate customer insights and sentiment. Tapping into these sources can help identify and segment customers based on their level of evangelism (or lack thereof).

4. You Know What Customers Do

One of the most exciting advances of digital marketing is you can now know what people DO rather than what they SAY. Expressions of customer interest can be a strong positive signal, however, they pale in comparison to the incandescent light of actual purchases.

It’s not just purchases you can look at. Today there is data on everything from what people search for on the intent, to the ability to track how customers traverse your website. And mobile data is starting to make its way into common usage.

Learning Opportunities

As a variation, you can also segment customers on the actions they didn't take. In WWII, British engineers up-armored airplanes based on where the bullet holes weren’t on returning planes, rather than on where they were. Planes hit in areas that had statistically fewer holes never made it home. 

Similarly, we can look at what did not happen as a source of inspiration. Abandoned shopping carts, pages with high rates of abandonment or other soft spots in a customer's journey to help us group customers.

5. You Know Which Customers Look Alike

Since you know everything about a particular person as well as everybody else in your data warehouse, you can start looking for similarities. You can choose between two common methods of finding similarities: The first is finding common features to use to predict specific desired behaviors, such as purchases. Those potential customers who look like existing customers can then be given a nudge in the desired direction.

The second way to leverage customer similarities is to filter on the point of similarity, and then begin the whole segmentation process over. Do you have fundamentally different segments for people already engaged in an activity from those that aren’t?

6. You Know Who People Are Friends With

Social media has opened up a whole new dimension of data to think about: who your customers interact with, who their friends are, the people in their digital network. Following these links can be just as predictive and insightful as finding similarities among customers.

7. You Can Let the Machines Decide

The tandem improvement in computing power and volume of data means algorithms now have the ability to identify meaningful groupings in the data without human interaction. The challenge with the machine output is that it is often incredibly complex and not meaningfully interpretable by people.

8. You Know What Images People Share

My closing segmentation approach is image processing. While currently only used in limited contexts, it's an area with new opportunities being explored and a rapidly expanding number of solutions.

There’s No Wrong Way to Slice a Carrot

This is only a small taste of what we can currently do with the tools at our disposal. Our capabilities as marketers appears to be continuously growing. 

So what do you think of our top eight? Did I miss one of your favorites? Please share in the comments below.

fa-solid fa-hand-paper Learn how you can join our contributor community.