By “Why Segmentation” we don’t mean “Why is segmenting your market important”. There are a dizzying array of resources available to answer that question, and frankly, there is little we can add to that. What we mean here by “Why segmentation?” is “why are you doing the segmentation”.
There are a lot of good reasons/applications of segmentations. Think of these in the broadest terms possible. For example, “product” may be a service or some intangible asset.
These are just a sample, but they make up the vast majority of the good reasons for segmentation.
Note that “the existing segmentation is a few years old and it needs to be refreshed” isn’t on the list. That is a very market research-oriented approach, and it inevitably leads to an academic exercise that collects dust on someone’s shelf. This is not just a marketing research exercise nor is it just an analytics exercise. Both a necessary, but not sufficient, for a successful effort. The point of all this is that segmentation is inherently a cross-functional process that requires leadership from research, but also requires input from all stakeholders and support, if not direct involvement, from senior management.
So how do you do this, how do you set yourself up for success?
First, make sure you have a cross-functional team. We don’t want everyone in the room, but we do want every area of the organization that will be directly responsible for making the segmentation come to life represented. Bonus points for getting someone with a CXO title.
Second, the team needs to determine what business activities are to be driven by the segmentation (this is the "why"). It’s tempting to try to make a segmentation that works for everyone in all circumstances, but this isn’t realistic. You’ll end up with segments that are undifferentiated and make no one happy.
Once the key business activities have been defined and prioritized, the team should lay out a hypothesis of what the segments will look like and how they would change what everyone is doing today. These are critical! This step is the most time-intensive and also the most common step to miss. Don’t make that mistake. This step has two purposes: it sets up the data collection (maybe not only from surveys) and it allows creation of a pro-forma report and ancillary communication pieces.
Now, we can get to data collection. Usually this involves survey research, but it may also include internal or secondary data. The segments you get are always a reflection of the data you choose to collect. Sounds obvious, but this is often overlooked. The work put into the previous (third) step really pays off here.
Once the data are collected, some analytics are performed to identify segments. The details of the analytics are the topic of a future post, but suffice to say that all the sophisticated analytics in the world can’t save a segmentation from other poorly executed steps.
Next, the segments must profiled and evaluated to identify which one(s) will be the target and what the implications on operations will be. Everyone must be on the same page. This is usually done in the post-research workshops.
The key to creating an effective segmentation is in the up -front preparation. If you can't get a cross-functional team to own it and its implications, not much else matters. The first three steps: assembling the team, defining the purpose, and hypothesizing segments, are often skipped and thus a missed opportunity for effective segmentation.
Brett Matheson is the founder and principal of EMS Analytics. He has experience in a wide array of multivariate techniques and their application to real-world business issues, including: choice modeling, segmentation, forecasting, and simulation.