Over the years I’ve worked on so many market segmentations, the list needs its own segmenting: big, small, global, attitudinal, demographic, behavioural, brand-led, even segmentations of market segments. This variety hides the essential uniformity of the process that I’ve noticed in all these projects. It normally starts like this…

Amy is a marketing manager at a large multinational company. One day her boss comes to her and says that he wants to increase sales by launching a new advertising campaign for one of their top brands. Being a smart marketer Amy knows that if she’s going to figure out how to sell more, she needs to understand who the company’s customers are. After all, people are more likely to buy more if it fits their particular needs, attitudes or desires. All the current information she has feels out of date and doesn’t focus on the product she’s interested in, so she commissions a segmentation study of the brand’s existing customers to find out who they are and what they believe in.

Her agency comes back and finds out that there are half a dozen different segments currently buying the brand globally. But then Amy starts thinking: if she’s going to make a big impact on the business, she not only needs to understand current customers, but potential customers, so she commissions a second segmentation study to look at potential customers. This second study finds that there are half a dozen more different segments of non-customers.

Confident that she knows her customers inside and out, she takes her segmentation studies to the rest of her team to decide what to do next. Her team thinks the segmentation studies are great, but her colleagues raise two worthwhile points: a dozen segments is a lot of different segments to deal with, and all the information is focused on demographics and psychographics. Surely their brand isn’t just about how much money you earn or whether you consider yourself a confident person, but about your lifestyle: how you act in the real world. Feeling a little disheartened, Amy decides to do a lifestyle segmentation that would cover existing customers and potential customers which she then hopes she could use to simplify the existing segmentation.

Based on the feedback from her colleagues she commissions a lifestyle segmentation, and after plenty of back-and-forth with her research agency they settle on a few key lifestyle segments which encompass all the previous segments. Not that the old segments should be ignored, after all they contain all that detail, but the new lifestyle segments offer a way of grouping and simplifying the existing customer and non-customer segments.

Feeling confident that she now knows the customers, non-customers and their lifestyle, she decides to share the new segmentation with the local marketing teams around the world. Again, they think it’s great - but Amy’s colleague in China makes a good point. Customers in China are probably going to be different to global customers as a whole so she is going to run a segmentation just for China - as is the manager in Mexico; her colleague in the UK; and the team in France.

Amy again feels a little bit more disheartened, but while the individual countries are looking into their segmentations, she decides to brief the brand’s communications agency on the new segmentation. The brand planners at the agency love the segmentation like everyone else, but feel that in order to really get under the skin of these customers they ought to do some focus groups. A couple of weeks later the communications agency comes back with the results from the focus groups. They’ve learnt a lot about why the different segments lead the lives they do; however, one of the key segments actually seems to be two similar but in certain, very important ways, different segments. However, Amy isn’t so sure: her friend is definitely part of this key segment and she’s nothing like the people the agency talks about in the focus groups.

At this point Amy decides to take a step back. She realises that after a lot of time, money and effort she has an overall lifestyle segmentation that maps out a dozen customer and non-customer segments, which may contain segments that are in fact multiple segments. And while she has a lot of detail on the brands customers and non-customers, she’s no closer to coming up with a new advertising campaign. It feels like there’s something wrong with this segmentation project.

You may not have found yourself in a situation as frustrating as Amy’s, but most of us can relate to at least some of this. This is because while segmentation sounds great in theory, in reality it rarely delivers. A study referenced in HBR by Marakon Associates and the Economist Intelligence Unit surveyed 200 senior executives of large companies in 2004. 59% of them reported that they had conducted a major segmentation exercise during the previous two years but only 14% of them said they derived real value from the segmentation.

There are a number of reasons why we struggle with segmentations; some are problems which have always been there, and some are getting worse.

Time and cost

The first hurdle we come to with segmentations is time and cost. Segmentations are slow and expensive, often requiring serious statistical skill, especially if the segmentation covers multiple markets. Because of this size, cost and complexity, we are often reluctant to update segmentations very frequently and in a world where things change so quickly, it’s making less sense to define who your customers are only every few years.

The myth of uniqueness

We often like to think that our brand or the market we work in is different from the other brands in the category and the other markets in the world. However in ‘How Brands Grow’, Byron Sharp writes about the research done at The Ehrenberg-Bass Institute which suggests that consumers within a particular category don’t tend to differ very much by product or brand—in other words, the person buying a BMW is the same sort of person as the person buying an Audi.

Likewise, in my experience, while differences across markets do occur, they’re not as frequent or dramatic as you might expect, or again to give an example, the person buying a BMW in the USA isn’t a million miles away from the person buying a BMW in China. Our tendency to believe that our customers are different somehow often leads us to do things like customer and non-customer segmentations when we should be doing broader category-led or global segmentations.

The naming trap

Segmentation is all about putting people into groups, but we often don’t think about the consequences of the process itself of putting people into these boxes. How we group people together and in particular what we call our segments changes our approach to them. We can be led astray by our own invented terminology. For example, segments are often given memorable names to help embed a segmentation within a business. These names often anchor themselves on a particular key attribute of the segment, so we end up with names like “Thrifty Freddies” for a value led segment. The consequence of this is that we over-focus on that highlighted attribute, thinking that our “Thrifty Freddie” segment is focused predominately on getting a cheap deal, even if the data shows that longevity of the product is just as important to this segment.

The friend problem

We also have a tendency to compare market segments to people we know in real life. The problem with this is that, on average, the segment accurately represents everyone within it, but in reality the segment doesn’t accurately describe any single individual. This disparity between the person we know and the segment average often causes people to question a segmentation or apply what we know about that person in real life to the entire segment.

Reductionism

One of the other influences acting upon segmentation is a whole range of cultural and technological changes currently occurring around context and personalisation. Whether it’s through location-aware mobile phones, on-demand printing or automated CRM, there’s a push towards increasingly personalized experiences for consumers. This trend is encouraging segmentations to be more detailed, segmenting people not only by demographics and attitudes, but also by behaviour and location.

However segmentations struggle with a tendency towards reductionism. Segmentation is all about a splitting a market into sub-groups - but where do you stop? How many different segments is enough? More segments give us more detail, which means we can better serve people, surely? Within limits this is true, but there comes a point where a segmentation is too detailed to be of any use to a business. In fact, the segmentation itself may be fine, but its complexity prevents organizations from acting on it.

This reduction often leads us to miss things. Imagine if I take all the sentences from a novel and list them all out alphabetically and then read through the list of sentences. After going through the list would I understand what the book is about? Of course not: I wouldn’t understand how the sentences fit together, what the plot is, or anything else that actually makes the novel meaningful to a reader. Well, what about if I went into even more detail, listed out all the words in alphabetical order, then went through the words with a dictionary to make sure I understood exactly what the words meant. Would I understand the book then? Again, no adding a level of detail to the understanding of parts still doesn’t help me understand the whole.

We know things are often more than a sum of their parts. This is true for market segments, and it is why clever behavioural segmentations like the Amazon and Netflix recommendation systems tend to offer us too much of the same because they are based on the assumption that people’s behavior is consistent both individually and as a group. For example, if you watch a Sci-Fi films at a weekend because you’re looking after your sisters children and they love Sci-Fi, Netflix keeps on recommending you Sci-Fi films to watch even though you secretly hate Sci-Fi.

One tool for every job

The final—and maybe most common—problem with segmentation is that we try to use it as a one-stop solution to a number of marketing problems. Segmentations focused on demographics work really well for media planning, but aren’t necessarily great in helping develop communications and are rarely useful in developing new product ideas; segmentations focused on psychographics and attitudes help a lot when developing communications, but again aren’t as useful for media planning or product development; and while segmentations that focus on behaviour tend to help product development, it’s not so great for planning media spend or coming up with a new advertising campaign.

None of this means we should stop doing segmentations. Just that we should stop expecting them to answer questions they were never designed for.