Every business recognises the importance of understanding their customers. Many make a concerted effort to gather data through customer research, asking questions about what their customer’s preferences are so that they can improve their products or services. But how can you be sure that it is safe to act on the results of that research?
A famous example of research going wrong was Coke’s change to its formula. In blind tasting customers preferred the new formulation, so new Coke was launched with a huge fanfare and tremendous confidence… and it crashed and burned. Six weeks later traditional Coke was back on the shelves.
More recently Walmart surveyed its customers and asked them if they would like less cluttered aisles, and customers said yes. So millions of dollars were spent on store layout, clearer aisles and shortened shelves. The result? Sales went down, and it has been estimated that they lost more than a billion dollars. Ouch!
How could companies as big, as thorough, and with as much money to spend on customer research as Coke and Walmart get it so wrong? The clue is in another example.
Researchers (Nolan, Schultz, Cialdini, Goldstein and Griskevicius) surveyed householders in California, asking them what would make them change their environmental behaviour. Would it be to help the environment, benefit society, save money or because their neighbours are doing it? They all said that the last thing that would make them change their behaviour was what their neighbours do, but when their actual behaviour was measured that was the thing that influenced them most.
When you ask a customer a survey question, they do two things. They use their rational minds and intellects to assess what they would do, but don’t engage their emotions; and they rationalise their assumed actions to conform with their idealised views of themselves.
So with regards to Coke, when tested customers preferred the new flavour. But when they stand in front of the shelves, all the emotions around the brand come into play (memories of childhood, Coke being a piece of American culture, whether they identify themselves as ‘coke drinkers’, etc). Intellectually, Walmart customers want less cluttered isles. But in practice, they shop for a bargain and associate a certain ambiance with low cost shopping. Householders think it would be shallow to be swayed by what their neighbours do, but conforming to social norms is an overpowering instinct.
Which brings me to the question of zoos verses the wild. Asking your customers survey questions is treating them like animals in a zoo; it is an artificial environment. Sure, you can learn a lot, but you have to be careful about what the data says. This type of research is best for gathering hard data (‘did you go on holiday last year’) rather than opinion data (‘why did you go on holiday last year’).
The key, whenever possible, is to measure customer’s actual behaviour – watch them in the wild. If you are testing something new and it is impossible to watch real behaviour, at least set up scenarios where something approaching real behaviour can be assessed. Think of this as a safari park!
And if all of this is making you think of your customers as animals, just don’t tell them that they are chimpanzees. According to my research, they don’t like it!