Good Segmentation Schemes Often Start Out Sounding ‘Just Crazy Enough’

Image by Lluís Ribes Mateu via Flickr.

Making Sense of the Messy Middle

“It just takes some time

Little girl, you’re in the middle of the ride

Everything, everything will be just fine

Everything, everything will be all right, all right

~The Middle by Jimmy Eat World

One of the benefits of having coached project teams through our needs-based segmentation process hundreds of times across dozens of industries is that we can spot patterns that are difficult to identify for even the seasoned industry practitioner who has typically developed a limited number of segmentation schemes, probably in just one or two industries.  One pattern that we often see is a natural clustering of customers based on size and complexity – with large customers demanding a high degree of customization for their unique requirements and small customers buying just one or two standard products serving their far simpler needs (or if they do have greater needs, we cannot afford to serve them because of their purchase volume).

While this type of segmentation works in addressing those two extremes, the problem is ‘the middle’ – the medium-size, medium complexity customers.  Leaving them all in one segment is often not a good option, as they can comprise 60 – 70 percent of your revenue. Trying to further refine them along these two dimensions, by creating ‘small-mediums’ and ‘medium-larges’ for example, might be initially satisfying, but it is usually not informative in terms of changing segment strategies, and it can be misleading.

We have always said that segmentation is messy and iterative. So, if you run into this type of situation, don’t give up.  Or as the song says “don’t write yourself off yet.”  Remember that in segmentation, we are looking to exploit asymmetries – the underlying factors that drive different needs among otherwise similar customers. Here are some things to consider when trying to tease out important differences in these medium customers:

  • First, ask yourself if you have the right segmentation variables. While size and complexity may seem initially appealing, do they really explain customer purchasing behavior and underlying needs? Step back and ask yourself what are the key differences in needs relative to your offering that we might exploit in a segment-based strategy. Explore other possible segmentation variables that might better explain differences in needs, such as: business model (i.e., innovator vs. follower) or step in value chain (i.e., Tier 1 or Integrated OEM). While you may have already invested in developing size and complexity as way to segment your markets, don’t settle on a segmentation framework unless it clearly ties to underlying needs.
  • Second, make sure you are thinking about all potential customers and the whole customer, not just taking an internal view of what we sell today.  Too often, segmentation starts with internal sales data.  Obviously, this gives you no visibility to customers who don’t buy from us.  But more importantly, for this topic, you may define a customer as medium when in fact we represent only 20 percent of their purchases.  They may actually be a large customer who only values us as a secondary supplier.  Their needs and the implications for how we serve them might be very different than those of a truly medium-sized customer where we have 100 percent of their business.
  • Third, if you keep coming back to size and complexity, see if you can further define these segmentation variables.  Size and complexity can be interpreted fairly broadly. You get a deeper insight if you add some granularity to these definitions.  For example, is size number of people, square footage of facility or number of locations rather than just annual purchase volume?  This can be important, as purchases are an outcome while these other characteristics are often one step closer to the underlying need.  The same is true on the complexity dimension, is it number of product lines purchased, number and variety of sensor inputs or frequency of changes to the system?  All of these can contribute to complexity but the implications may be very different.
  • Fourth, see if there is some other structural dimension that you can add to the segmentation framework that might tease out differences within the “messy middle”. Is there another dimension along which customers vary that could be driving differences in needs and therefore buying behavior?  For example, two customers may be similar in their technical product requirements, but one is a mature company with an experienced engineering group that does their own integration while the other is a start-up with limited experience who needs to buy a turnkey system and a support contract in order to accomplish the same end. In this case, adding a dimension of technical maturity to your segmentation framework might be powerful and more directly linked to needs.
  • Lastly, think beyond the product to see if sub-groups of these middle customers might be served by a different business model.  For example, if you could make your product easier to self-configure or re-package your offering ‘as a service’ which customers would be most interested in it and why? This thought exercise may help better identify and define differences of underlying needs within the messy middle.  Are there additional services that some customers might be willing to pay for but that others would not? This is always a good sign that you are getting closer to an actionable segmentation.

To see how these can all come together to create a real insight, we recall a client of ours who sold Personal Protective Equipment (PPE) to industrial facilities around the world.  They quickly saw that their customers followed the size/complexity pattern we have described here.  They had very large, demanding customers who wanted custom programs and products and shopped their PPE spend in large and very detailed multi-year tenders that required aggressive pricing to win.  At the other extreme were small customers who bought only one or two categories of PPE at volumes that were most economically served on-line – no personally selling, customization or discounting made sense for this segment.

The problem was that these two extremes combined described just over 30 percent of their potential business – the vast majority of their sales (and even more of their profit margin) came from the medium-size/medium-complexity ‘middle.’  As they dug deeper, they realized that they had lumped into this medium bucket some customers with a single relatively large location and other customers, with a similar annual purchase volume, but spread across many individually small facilities. 

This turned out to be a critical distinction because large single sites usually have a dedicated Environmental Health and Safety (EHS) Manager.  These managers generally viewed themselves as the PPE expert – they knew their facility, did their research, defined exactly what they wanted and how much they would buy. They typically tried to play vendors off against one another in a process that was fairly similar to what we described for ‘large’ customers.

On the other hand, the medium customer with multiple small sites likely had just one EHS manager serving all of them.  Many of these customers would value a supplier who could help them track the different regulatory requirements across facilities, set up inventory management programs to make sure key items never go out of stock, and provide summarized usage and compliance reports.  All of these are additional services that represented incremental revenue opportunities for our client, and would also help make them ‘stickier’ as a supplier with this newly identified segment. So, if you find yourself stuck in the middle, you could do a lot worse than to take the advice of Jimmy Eat World: “Just do your best. Do everything you can. And don’t you worry what their bitter hearts are gonna say.” Go back and try the suggested paths above: maybe you are too internally focused, maybe your segmentation variables are wrong, maybe you need to further define your segmentation variables, maybe you need to add a dimension. In our experience, there is almost always a way to subsegment this messy middle, if you are willing to step out of comfort zone and try something different. Remember, segmentation is an iterative process and good segmentation schemes, like good strategies, often start out sounding ‘just crazy enough.’

Image by Lluís Ribes Mateu via Flickr.
Yellow-Red-Blue, 1925 by Wassily Kandinsky

Leave a Reply