When the Market Changes and Your Segmentation Doesn’t

By Mark Ailsworth

Segmentation. Personas. Targetable audiences. These are all examples of the way a company views its customers.

And the way a company sees its customers is the way that it sees the world, its market, and the opportunity to grow.

So many important business choices are based on this point-of-view—marketing, product, sales, hiring, operations, etc. If your segmentation is off, then many of your core decisions will be wrong too. This is why it’s so important that your company’s segmentation practice accurately reflects the market.

Segmenting customers into groups isn’t easy and requires historical data to guide the process. But what happens when there are dramatic shifts in consumer buying behavior and historical data is worthless for segmentation?

How does a company move forward when it knows that its segmentation is faulty—and thus, its market perspective and all of the cascading decisions made based on that point-of-view are wrong?

You can’t have static customer segmentation in an ever-changing market

The current global pandemic has disrupted every industry.

Because of the radical shift in behavior between pre-Covid times and today, most segmentation models will need to either be retrained or restarted completely.

One of the core tenets of simMachines’ segmentation approach is that it is dynamic (versus static) and as new signals come in about customers, individuals are moved between segments. For example, if you are a national restaurant chain and there is new data showing that your customer, “Jane Smith”, now has children, she might move from a Upscale City Spenders segment to a Trend-Conscious Moms one.

But, even a dynamic segmentation approach isn’t flexible enough to handle the extreme circumstances that we see today. Consumer behaviors have fundamentally changed and the attributes your segmentation uses to differentiate your customers’ behavior is likely now completely irrelevant. The models that segmentation are based on need to be changed in the light of COVID…

But how?

With consumer behavior changing so frequently, how does a company properly segment its customer base fast enough to keep up?

When the market changes and your segmentation doesn’t

Let’s revisit Jane Smith. In your national restaurant chain, she’s in the Trend-Conscious Moms segment. Before the COVID pandemic, Trend-Conscious Moms stopped by once a week before or after they picked the kids up from sports practice to buy dinner for the family. This is a valuable segment because they come in after the After Work Crowd clears out, the average order value is high with plenty of food and even high-margin desserts to satisfy everyone at home. This segment doesn’t buy drinks from the bar, so alcohol and apps promotions don’t move the needle—but curbside pickup is something that will.

Pre-COVID, you knew how to market to Trend-Conscious Moms and what specials could increase revenue against this segment. But during the pandemic, the Trend-Conscious Moms segment is completely different. There’s no more soccer practice or ballet lessons. Money’s tight from layoffs, furloughs, and general uncertainty. Many of these moms are at home now and have more time to cook and stretch their food budget by grocery shopping online.

What do you do?

Do you still try to chase this segment? What promotions will even work on them any more? Are Trend-Conscious Moms even a viable group to focus on any more?

The reality is that new types of customers evolve whenever the market shifts. As cities and states open up on their own timelines, outdoor patio seating becomes an option (as it did here in Chicago last week). There’s a younger crowd that has been antsy from quarantining and the last thing they want is another home cooked meal. Let’s call them Restless Socialites. They want to see their friends in person and not just over Facetime or Zoom. They are looking for options to eat nearby that serve alcohol during the day. They might sit there for hours chatting, ordering drinks, apps, and meals. Such a segment is high-ticket and very valuable for your struggling business.

How do you identify this segment quickly and answer important questions that follow? 

• How do you figure out the right promotions to run and the best media mix to reach them?
• Who are your competitors? Are they the same group pre-COVID or has that changed?
• How do you measure each location’s success against Restless Socialites?
• How do you confidently guide your regional managers and franchise owners to switch their focus to this new segment before your competitors figure this out?

So many questions to answer with so much on the line… Marketers need a new approach to handle these challenges.

Segmentation modeling speed is the new requirement of business

You need a segmentation approach that can work quickly. You don’t have time for the business analysts to take six months to research and come up with conclusions. The truth is that we’re in such foreign territory they wouldn’t know where to even start.

And what about the waves of COVID that the scientific community is warning us about? They will each usher in new changes in consumer behavior that will require yet another segmentation model. By the time your analyst team has started crunching the data, the next shift in behavior will have already occurred.

Companies need a segmentation modeling system that can work quickly. simMachines’ approach uses explainable artificial intelligence built on a proprietary high-speed similarity engine that brings a new level of not just accuracy—but speed—to segmentation. This is the new requirement of business.

We use a predictive modeling approach which ties segments to business metrics. In the national restaurant chain example, KPIs such as average order value, new vs returning customers, etc. are taken into account. Old school descriptive modeling uses too many manual assumptions that are more like personas than true customer segments. With predictive modeling, segments are tied to business outcomes that the AI forecasts.

simMachines’ modeling can do in hours what an analyst team could take months to accomplish manually. And once you have your segmentation model, you can test it and learn. Segmentation is—and should be—an iterative approach. If the segments behave the way that the model predicts, then you know your segmentation is spot on. If it needs some adjustments, simMachine’s machine-learning algorithms can take in the new data and retrain the model.

There will always be waves and market fluctuations—the question is, are you ready?

COVID-19 is just an extreme example of a market fluctuation. But the truth is, there are always market fluctuations that impact your business. Seasonality. New competitors. Old competitors pushing forward with bigger budgets. Changes in taste and style.

Speed is crucial because, in business, there’s always a timeclock.

Customers are not gained. They are won and lost between competitors. If your competitor can align their operations faster than you, then they can beat you to the punch.

In the case of Restless Socialites, once they find their favorite new spot, they’re going to keep going back. You have to be first. It’s much harder to wrestle away a loyalist from a competitor than to win over a new segment. Think about it this way: would you want your competitors to already be up and running with an updated segmentation model while your company struggles without one? 

Marketers have known that artificial intelligence and machine learning was going to forever change things. If there’s ever been a time to react quickly and embrace the new way of doing things, it’s now. No one knows how long COVID-19 will be around and what the next 18-24 months are going to look like. The only thing a marketer can bet on is that there’s going to be change.

Is your company prepared for the rapid change that is inevitable?

Here’s my email address: markailsworth@simmachines.com. Let’s chat.