In the Dark Ages, before ‘click if you like,’ the intentions of a nation were predicted by nth samples of, say, 1014 correctly profiled individuals. Tiny focus groups would lend insights as to whether a purple wrapper would sell a chocolate bar better than a red one. Quite often the pollsters got it right. Now, with access to massive sample sets, Big Data miners argue we are drawing ever closer to the Age of Infallibility, at least as far as predicting what people are likely to want, think and do.
A Wharton study, Finding a Place for Market Research in a Big Data, Tech-enabled World draws up the battle lines between Big Data science and the art of traditional market research. According to Professor Peter Fader, one of the report’s authors: “It’s a widely held belief that it has become much easier to test things in market [by saying] ‘let’s put out our concepts and see what gets clicked on the most.’ That can help you determine a winner, but it doesn’t help you design what would have been the best. By doing careful research and determining the underlying drivers that cause people to click, we can develop better products and services.” Agreed, up to a point.
Big Data presents a disruptive challenge to traditional models of market research, but is it merely adding quantum sample size to cohorts which necessarily in the past would have been much smaller? Market researchers will need to adapt to and use it in the same way that engineers, designers, film editors, music producers and, indeed, surgeons, pharmacists and clinicians needed to incorporate digital technology into their skill sets. But fully automated analytic algorithms may abstract the human element entirely.
But how do we account for ‘Black Swans‘ in the form of inconvenient mavericks who succeed without apparently researching the market beforehand? Fader blames companies such as Apple for setting a ‘bad’ example by delivering brilliant products without relying on extensive market research. “Too often we are counting on a visionary, a Steve Jobs, who really set the market research industry back years because he was so disdainful of it,” Fader notes. “[Jobs] said, ‘I am going to tell people what they want.’ And unfortunately, he was right, or maybe he was just lucky. . . . For most companies most of the time, emulating Steve Jobs — having that kind of brilliance, or arrogance — isn’t going to work. It is going to take more thoughtful data-driven approaches.”
Fader rightly cautions against launching on ‘gut instinct’ without researching the market beforehand. There are classic examples: think of the Sinclair C5 electric trike, or, for that matter, the Apple Newton. The first was driven (literally) by its inventor’s hubris. The Newton was too far ahead of its time. But whether or not they use formal research techniques, innovators do ‘research the market’ by improving on existing concepts or anticipating unstated needs. Take the iPod: Apple didn’t bet the company on the outlandish proposition that a couple of people would like a tiny device that enables people to download and listen to music on the move.’ He admired the multi-million selling Sony Walkman and made an Mp3 version. The iPhone’s model was the Nokia Communicator. Apple’s innovation was to turn it into a must-have fashion item. The touch screen interface, now standard on smartphones, was a giant leap forward, the ‘killer app’ if you will.
At almost every stage of its development, Apple, particularly under Jobs, made better mousetraps. The graphical user interface concept, which changed the face of computing, was born at Xerox PARC but Apple took it to the mass market. Ditto the computer mouse. Xerox was obsessed with market research, but it is debatable whether market research ever reveals what people might want, based on what they already have. The horseless carriage was a development of the horse-driven carriage, but Henry Ford commented that if he had asked people what they wanted, they would have probably said they wanted a faster horse . . .
Market research is a valuable component when building a better mousetrap, but generally works best when there is an existing mousetrap. Unless the sample focus group includes characters such as Steve Jobs, Bill Gates, Forrest Mars, Ingvar Kamprad, Henry Ford, Alfred Sloan, Tim Berners-Lee, Walt Disney, Mark Zuckerberg or the Wright Brothers, it is unlikely that the need for innovators will be eliminated any time soon. But the Big Data pathetic fallacy is the belief that machine learning will enable organisations to do exactly that.
According to Wharton Professor Shawndra Hill: “I have a fairly crisp definition (of market research) which is assessing consumer preferences and attitudes. Market research becomes part of big data when they marry what people are doing with what people are saying.” She adds: “while mining the data has come a long way from where it was 10 years ago, we still have a long way to go to … knowing the true meaning of what people are saying. Nothing beats knowing why people make the choices they do.’’ Quite so, but when and if we build machines to work out how to do that, we will all be Pavlov’s dogs. Where’s the fun in that?