Segmentation is all about Diversity
“Market segmentation is a natural result of the vast differences among people.” Donald Norman,
Director of The Design Lab at University of California, San Diego. The strength of this statement lies in its simplicity. It, in very few words, reminds us of the reason we engage in segmentation in the first place, to try and reach as many people as possible, knowing that each person is driven by a different set of wants and needs. Why is this important? Because we, i.e. marketers, have been segmenting populations for decades, if not centuries, but still have a tendency to lose ourselves in the trees, focusing on how to divide up our markets, while losing track of the forest, i.e. that there are actual people behind the numbers and segments.
Generalizations can have Ramifications
Here’s an example from our world at ActiveTrail: A classic segmentation parameter in email marketing is “how often do my email recipients open mails from my company?”. Now, let’s assume you have put together an absolutely brilliant email campaign, and are in the midst of configuring how your marketing system should react based on email opening frequency. Many an email marketer might set their system to act as follows:
- We start by sending emails once a week.
- For contacts who have opened emails, but haven’t within the past month, put them in a group scheduled to receive “emails once a month”.
- Classify contacts who haven’t opened their emails in two months or more into a group “email once in two months”.
The issue with the 2nd segment, is that by including “or more” people in the group, we may be doing ourselves damage, as there is a very good chance these people do not want to receive your emails anymore, maybe they even consider them a nuisance. We are forgetting that “or more” refers to actual people whom we might be turning into antagonists (against our company) with this oversight.
So many Segmentation Options
The same is true if we go too far in our segmentation. Corporations, governments and academia spend fortunes on building segmentation models to help them tailor their messages and offerings with very fine granularity, such that, today, some circles are using terms such as “single person segments” (while interesting in itself, we won’t address this oxymoron here). Technology has done wonders for segmentation as well. For example, multiple signup interfaces backed by multi-faceted data in high speed databases have allowed us to move from single parameter segmentation (e.g. by gender) to single layer segmentation including multiple related parameters (e.g. by demographics) to multi-layered segmentation, where our groupings are based on any number of layers containing any number of parameters (e.g. by demographic parameters, behavioral parameters and derived parameters).
Let’s turn back to ActiveTrail and email marketing for an example of overly detailed segmentation: Perhaps the most effective method for splitting up target populations in the context of email campaigns is by click behavior. When people click on a button or link in an email, they are telling you much about their personal preferences, and consequently, of how to group them. For example, a travel agency might try to entice clients to take a trip by sending out a promotional series of emails offering sales on vacations to certain destinations, say emails offering vacations in London, Paris, or Rome. When an email recipient clicks on one of the options, e.g. “London” you have them automatically added to the group called “London”. In this fashion, you can keep sending them offers for London, even if they don’t buy the London vacation this time around. Let’s say however, that your European sales manager becomes overly eager, and has you offer 6 different cities as options in the email. Here, by offering too many choices, recipients may not click at all or click at random (which could be worse than not clicking, in terms of segmentation). Once again, by getting over excited about what we can do, we forget about the individuals we are hoping will click on our mails.
Big Data can be a Big Help
One technology that ActiveTrail offers (as do a few others) that is helping in this matter (when used right) is data mining, or in its broader form “Big Data”. With deep analysis of historical data, we, or rather a computer, can discern more intricate and/or non-obvious patterns of behavior that allow us to peer beyond simple segmentation and to see how segments interconnect. Data mining might reveal, for instance, that women have a propensity to click on London when they open emails in the mornings, on Rome in the late evenings, and not at all when they receive mails in the afternoon. Let’s assume we determine that this is because in the morning women are more practical, and are thinking about getting away to somewhere where they know the language, in the afternoon they are busy and don’t think about vacations at all, and in the evenings, their romantic sides come out, moving them to prefer Italy. As you see, the deeper insight into the data lets us think more about who and why people are clicking on our mails, leading to better segmentation.
Art or Science?
So where’s the truth? In actuality, there is no clear cut answer to this, and segmentation is not a pure science, rather it has a very heavy artistic part to it. With that said, whatever you do and however you go about segmentation, remember the people behind the segments.