Part of my job as a BI consultant is to help my clients to define and structure their data so they can analyse their business and make better decisions.
Although my clients are in widely differing industries, they often have certain concepts in common. Customer is one of these (although it is not always this precise word that is used). At first sight, it may seem easy to define what a customer is. If you ask different people in the business, most of them will also be completely clear about the definition, but if you ask them to answer the questions How do you define a customer? and How many customers do you have? you will often get as many different answers as people you ask.
It may seem that such a central and fundamental dimension should be easy to define, but it often turns out to be one of the hardest. Perhaps precisely because it is so central – everybody has their own view of what is right.
What should we measure?
Number of customers is usually an important indicator. We could also consider measuring number of new customers, number of lost customers, number of customers who have bought more than one product, number of customers who have only bought one product or number of customers who shopped once and never came back. Apart from numbers, we might also be interested in customer turnover (churn), which customers are most profitable or not at all profitable or what distinguishes the most profitable customers.
Whichever measure we are interested in, it is important to agree on a definition. There is not just one right answer to the question What is a customer? but in order to work out how many there are and produce a meaningful answer, you have to know what you are counting and always count in the same way.
My experience is that many people set out to count how many customers there are in their customer lists and take that as the definition of a customer. There are several problems with this: Can someone appear in the customer list even though they have not bought anything? Can someone appear multiple times with different customer numbers?
Can a customer become inactive or be removed from the list if they are no longer a customer? If so, when is this done?
I think it is better to look at actual sales. Is someone really a customer if they are on a list but have not bought anything?
Why should we measure?
As we have seen, it is not that easy to know what to measure, but it is also interesting to ask yourself why you want to measure something and what the implications might be. If you focus too much on new customers, you may not give enough attention to the existing ones, although it is often easier and cheaper to invest in those you already have. If you focus too much on the number of customers, this may be at the expense of revenue. But if you look only at revenue, you may run the risk of having too small a customer base and being too dependent on a few customers.
The fact that it really isn’t completely obvious what a customer is, or what one should measure and why, and that there may also be several other concepts and aspects that you want to analyse, can sometimes be frustrating, but it also makes my job challenging and inspiring.
I want to finish with some general recommendations. As an external consultant, I have an important role to play in helping the business, but they are ultimately the people who have to own and take responsibility for their definitions. I don’t think you should be afraid of analysing data and making use of the results, such as by looking at which customers actually have sales instead of taking all those in the CRM system. I also advocate iterative development; rather than getting bogged down in theoretical discussions, you simply test a definition, examine the outcome and then refine your definition. I don’t mean that you should keep changing, but it is all too easy to be obsessed with finding the perfect solution rather than settling on something that is good enough, which can then be changed if necessary.
Business Intelligence Consultant