“Know your target market” is one of the mantras of modern business. It is shorthand for saying that its a good idea to design your product and communications so that they can be found by, and appeal to, the customers that you think are the best prospects.
If you are an established product working out your target market can be as simple as profiling your customers and can be done using predictive modeling. However, if you have a new product, or, are struggling and suspect you are targeting the wrong target market, you need to segment your market prior to deciding who to target.
Market segmentation is the term that marketers use for breaking a market up into parts and developing different strategies for the different segments. Almost all businesses segment their markets. Nevertheless, there are two quite different logics of segmentation and it is important to have a clear view of which you are trying to employ.
The classic marketing approach to segmentation focuses on identifying customers who need different things. This has lots of different names, but in this post I’ll call it needs-based segmentation. While needs-based segmentation focuses on understanding customers, its ugly big sister is customer value segmentation, which focuses instead on identifying the customers that offer the most potential value and targeting these customers.
My earlier posting on predictive modeling and the 80:20 rule covered the basics of customer value segmentation, so I won’t repeat them here, other than to emphasize that the core logic is one of identifying how customers differ in their appeal and then targeting the most appealing customers (i.e., those that provide the highest value). This logic underlies how most consumer-facing service companies work. If you are a big customer of a telco, bank or airline, they make you feel special. Even small businesses do this. If you go to the same place for coffee every day, they learn your name and maybe give you a rewards card so that you can save some money.
Customer value segmentation works in situations where the low value customers are, in some way, intrinsically less appealing. For example, if you are selling data storage solutions the company that needs to store many terabytes of data is intrinsically more appealing than the company that only needs to store 100 megabytes. That is, customer value segmentation works when some customers have “smaller” needs than others.
Customer value segmentation breaks down in situations where the reason that people are low value customers is because they have different needs, rather than smaller needs. And this describes most of the truly innovative new products of recent years, such as iPhone, Dropbox, Nespresso and Xero. Think about Dropbox. Most of its 100 million+ customers were the minnows that were never approached by IT companies looking to sell Virtual Private Networks (i.e., networks that you can access remotely). What makes Dropbox so useful is not just that it is cheap, but also that it is much simpler than running a “real” network. That is, Dropbox addresses a different set of needs to those addressed by traditional networking solutions.
Needs-based segmentation involves identifying the groups of customers in the market that have distinct needs, where each of these groups is known as a segment. There are two basic approaches: intuition and statistics. The intuition-based approach is the main one that people use. Indeed, it is the one that underlies DataCracker. The core intuition behind our product is that we think that many of the companies that do surveys do not have the technical-know how to analyze these surveys and, consequently, need software that will empower non-experts to analyze their own surveys. Thus, we have chosen to segment the market for survey analysis into two segments: experts and non-experts and target out software at the non-experts.
Of course, the danger with intuition is that it can be wrong. For every failed new business there was an entrepreneur confident that their product was meeting a need. One solution to this is to use your intuition to segment your market and then use concept testing to check that your idea has legs. The other solution is to use statistical methods to create needs-based segments. The basic approach for doing this is to:
- Conduct a survey with at least 300 people that could, potentially, be in the market, and ask a whole series of questions designed to work out what needs people have. Working out what to ask is the hard bit. Check out mktresearch.org for some general tips on how to do a good survey.
- Use a statistical technique called latent class analysis to hunt through the data and automatically find segments in the data. Latent class analysis is available in most modern data analysis programs, including some easy-to-use programs like DataCracker (in which it is conducted by selecting Insert > Groups/Segments). If you cannot access software that does latent class analysis, a fallback is to use cluster analysis, which is the same thing but cruder.
Once you have worked out what segments are in the market, be it through intuition or statistics, you then move onto the fun bit, which is to work out which you want to target and to develop new products for these segments!
If you find this post useful, share it.