Garbage In, Garbage Out

"I think that in many research companies there are too many statisticians and social psychologists and not enough business people, so market research is done in isolation from business, with no opportunity for actionable output. In the business itself, research is seen as the domain of the marketing department, especially advertising, whereas it should provide guidelines to every employee."

Watermeyer says the problem is two-fold: in the research design and in the application.

"Most research tends to focus on what happens without taking into account why it happened or what the business did to make it happen. You have to ask the right questions. The second challenge is translating the results into operational language for the entire organisation in order to give guidance to every department, from R&D, to sales, from Human Resources to the factory line. You have to operationalise the needs of people into an organisation."

Watermeyer gives the example of the concept of cleanliness in a retail environment. "Research will find that consumers want cleanliness in a store. But what is a clean store? It means different things to different people - it could be a clean floor, or neat staff uniforms, or fresh smells at the fish counter. Research has to concentrate on these specifics so that every employee within the system - from storage to till packers - understand what to do to satisfy this need for cleanliness."

When Johan Bolt & Associates was working with one of its clients, a major retail chain, the consultant and the client drew up a questionnaire with 88 specific 'need-states', such as a polished floor, friendly staff, fresh-looking fruit and vegetables. "We then looked at the issue from the retailer's point of view - what he can do to satisfy these need states," explains Watermeyer.

Good research, therefore, has to focus on specifics. "Market research is a science and as such has to understand the causal links between dependent and independent variables, between means and ends. You cannot establish spurious links between concepts that are not meaningfully related," Watermeyer reiterates. "It must understand the difference between product features and benefits, consumer needs and wants, product and people attributes, purchase occasions and shopping channels, cause and effect."

Watermeyer gives health as an example. "If you ask a consumer if he wants to buy healthy products, it is pretty meaningless. It is like asking, 'What do you prefer, health or absence of a sore throat. Do you want a healthy life or to be HIV-negative?'" she says. "You must ask questions like: 'Is your concept of healthy food, low-sugar or low-fat?' Then you will know how to design your products accordingly."

Sticking to the healthy theme, she adds: "So often research companies tell a client their brand is sick, but the client doesn't know if it has AIDS or a sore stomach! As a brand consultancy, we cannot doctor 'sickness'. We can only doctor the cause of the sickness."

Watermeyer suggests that market research must concentrate on understanding consumer or shopper needs, and the behaviours they engage in to fulfil these needs. Only once you understand these needs can you decide what to do to fulfil them. Most market research concentrates on demographics and psychographics, while there is actually very little correlation between these two variables and consumer behaviour. You also cannot translate psychographics onto an assembly line.

Manufacturers of big brands find themselves faced with contradictory results from competing research houses. "Part of the problem is that conventional segmentation groups similar people, whereas effective segmentation should be about grouping people with similar problems. To give an example: two people from vastly different LSMs might both have the same need to relax at the coast during the Christmas period. The lower LSM individual, however, might caravan to Durban and stay at a campsite, while the higher LSM person would stay in a Five-star hotel in Cape Town."

Circumstance also determines needs and behaviour. "The two key attributes of a successful brand are relevance and differentiation," says Watermeyer, "and these apply equally when conducting research. You have to find out what is relevant to a specific shopper or consumer - what is important in terms of the need-states to a specific target market - because what is important to a shopper in a local store in Butterworth is very different from what matters to someone who shops at Sandton City. Then you have to find out how the client's store is measuring up against their competitors, in other words, how to differentiate."

Useful research also has to take into account that consumers and shoppers behave differently in different circumstances: this is a variable that is often not recognised. In the retail environment, for example, "one shopper has different needs from a convenience store and a hyper store and behaves differently in each. At the corner shop, where she does her top-up shopping, she wants convenience, speed of service and easy parking. At a hyper-store, however, which is a destination shopping experience, she wants range of products and competitive prices. This is a rich area of research where very little has actually been done by the well-known research houses who continue to implement measurements and methods of yesteryear."

Watermeyer makes the point that the consumer is a notoriously unreliable source of information about his or her needs or future behaviour, if they are asked directly. "An excellent example is that if consumers were asked 20 years ago if they want a home computer, they would have said no. At that stage computers were room-sized mainframes. IBM famously misread this situation. However, if you had asked them if they need a convenient way to do their business from home, they would have said yes. To put it another way: if Thomas Edison had done consumer research, he would have invented a bigger candlestick, not the electric light bulb!"

"I think that part of the reason why a lot of research misses the point is that useful research sometimes raises more questions than it answers," Watermeyer explains. "It gives answers the client doesn't have, instead of confirming him in what he already knows, and this is frightening for the client and the research company. That is why in focus groups, you often find generic or leading questions, instead of coming at the issue from an oblique angle.

"An illustrative example of this was when a couple of progressive stores installed cameras to watch shopper behaviour within the shop and at different categories. Shoppers were filmed and then, once they had exited the store, were asked questions about how they had behaved. Most of them were wrong!"

The only sure thing when predicting the future is that you will be wrong! says Watermeyer. "You have to come in from a different angle, to try to understand the basic drivers of consumer behaviour. Only then will you produce useful research that can be put into operation in order to satisfy those drivers."