Evolving from Traditional Segmentation to Segments of One

Blog Post
January 25, 2018

Traditional segmentation and targeting are all about effectiveness and efficiency.

We are all familiar with a classic use of segmentation and targeting: offering a luxury product at a high price to the affluent and a stripped-down, low-cost version to the masses. 

Alfred Sloan of General Motors provided us one of the first examples of segmentation when he developed a strategy to compete with Henry Ford’s "You can have any color as long as it’s black.” Sloan created brands to serve different segments better than the black-only Ford—Cadillac for the wealthy, Buick for the doctors, Oldsmobile for the executive, Pontiac for the sporty, and Chevrolet for the masses.

But there are diseconomies of multiple brands. During GM’s 2009 bankruptcy, Oldsmobile and Pontiac were discontinued. Buick was saved only by its popularity in China. It was just too expensive to have so many offerings. 

Now we are entering a period where the plummeting cost of computer power and the dramatic increase in artificial intelligence—in self-programing computers and the ability for computers to learn—are changing the economics of segmentation. It’s not a question of if but when will we be able to create and serve segments of one with our communications. 

Segments of One in The Boutique

The problem with traditional segmentation is that it is static, but people are dynamic—they can be in one segment one day and a completely different segment, solving a different problem, the next. Each person can be one segment at the beginning of their search and then find information that shifts them to a completely different segment and process.

At Harte Hanks, we are pushing the frontier of segment-of-one customization of service offerings to address this dynamic segmentation. Through our efforts in our daily meeting of The Boutique (our version of a marketing war room), we analyze the traffic on our website and blog to discern clues about searching behavior—and use these clues to speak to our visitors more contextually. Real examples include:

  • For a prospect who was searching for information about a decision we inferred from the time spent on a specific topic, we responded by sending in-depth information about the topic.
  • For a prospect who returned several times to explore a series of related subjects, an invitation for a real-time chat led to a conversation with a product manager and eventually to a customized consulting engagement.
  • A visitor with an immediate, pressing need, chose to speak with a representative on his first visit. The need was identified, and a quote for a customized offer was made and accepted within hours.

From Manual to Automated

We are just moving from a completely human (read: manual) process to the first steps of automation. We'll be using algorithms to identify browsers and shoppers who visit our digital boutique. So far, the humans and he machine agree 60% of the time, which is a great start. By having a machine identify the browsers and shoppers, it will free up the Boutique members to spend their time trying to understand the journeys and to decide the next best conversation.

Moving through the year, we will be deploying machine learning and AI to have the machine take on more and more of these steps—not only identifying the journeys, but making decisions on what the potential client is looking for and deciding on what conversations should happen next, whether that is customizing the web page they are on, sending an email or direct mail outreach, or even directing a business development person to contact the visitor directly.

When we achieve that level of automation, this will become a marketing process than can be done at scale. And at each step, we free up our human marketers to do higher level work, evaluating the conversations taking place, identifying content gaps in these conversations and introducing new types of conversations to be used in the process.

The paradox here is obvious. We are using automation and AI to make the dialogue between a potential customer and us (and ultimately our clients) feel like it is personalized in the moment—just like like the visitor is speaking to a sales person in the store.

Online retailing, for example, has made great strides in using AI to improve the process of buying books, home furnishings, the perfect gift—so many categories—online. It is just a matter of time before AI enables the creation of customized, complex offers in an entirely automated process. And the time is drawing near.

If you liked this post, check out some of our other Boutique content to see how we're evolving towards more human marketing.