RTM: Can you give an example?

LS: Sure. One of my former students, Nirmalya Kumar, now a professor at IMD, did some consultancy for a caterer who was putting bars and restaurants into an airport. A typical demographic analysis would say: 'well, we have four groups - business people, groups, airport employees and tourists. So let's build facilities for each of these.'

But Kumar did a needs analysis. This showed that what was really important was the amount of time you have. Those with time will typically want something which is more service intensive. Those with less time want faster food. His solution addressed the underlying needs of travellers in a way which a demographic analysis wouldn't have done.

RTM: So you are saying that demographic terms like 'small and medium businesses' are really useless?

LS: Absolutely! I have been trying to disallow SMB for decades! Most companies dismiss SMBs as the small opportunities which they can reach through mail order or direct mail and focus their salesforce on large accounts. In practice, many SMBs may have a greater need for service than some of the large accounts!

RTM: So needs based analysis just produces much more elegant and satisfactory solutions?

LS: It also enables you to see much larger markets. Most IT companies segment by vertical market. Let's say you supply a logistics solution to a large pharmaceutical retailer like Boots. That solution meets Boots needs on inventory turn, logistics and reporting. Instead of searching around for another Boots in the pharmacy vertical, you should say: 'OK, so what other companies in other industries have similar needs?'

RTM: I guess the big question then is how you actually carry out such a needs analysis?

LS: Yes. The trouble is that customers are an unreliable information source, particularly on new products. Here I would like to pick up on something I heard from David Pottruck, chief executive of online stockbroker Charles Schwab. He made the point that customers can not tell you what they want, they can only tell you what they don't like about your products and services. They can't tell you what they want, because they haven't actually seen the new product. And they are hopeless at talking about what you should be charging them. They just don't have the skills to talk about these things.

Typically, companies are bad at listening to complaints and criticism from customers - the area where feedback is most useful and quite reliable. If they were better at this, they could quite easily make big improvements in what they do.

RTM: Hmm, yes. Rather than listen to complaints we tend to want to go for the next step and hold focus groups, where we ask customers what they want next.

"Frankly, people spout garbage in focus groups..."

LS: Yes, and focus groups are disastrous when used like this. A focus groups can create a hypothesis, but it can't prove a hypothesis and yet this is how they are normally used!

They are notoriously influenced by group dynamics and strong personalities. And, frankly, people can spout garbage in focus groups - they talk about warm, fluffy things and feelings, which don't really mean much.

RTM: So how do you segment customers' needs?

LS: There are quantitative research techniques - conjoint studies which can coax these things out. I have been using them for two decades and they do work.