In his book, The Undoing Project; A Friendship That Changed The World, Michael Lewis describes work carried out by the Oregon Research Institute into an algorithm. The algorithm was designed to test whether someone had cancer. It proved to be better at identifying cancers in patients than doctors – every time. Doctors, in theory, could be replaced by an equation created by people who knew nothing about medicine and had simply asked a few questions of doctors.
An algorithm is a step-by-step set of rules which performs the task of solving a mathematical problem. We use them in market research. In fact, we are using them more and more, especially to predict which segment a company should be in.
We have tried to develop killer questions for sales people to ask their customers so that they can classify them into different segments. This has proved unreliable. Sales people either don’t ask the questions because they think they know better or they skew the answers with their own biases. The objectivity of an algorithm beats the sales person every time just as the algorithm proved more reliable than the doctors’ diagnosis.
Algorithms are becoming more and more important in marketing. As well as using them for placing companies in segments, we use them for checking data. They reliably look over a set of online survey data and work out who has flat-lined (given the same scores to every question), who has entered gibberish to open ended questions and who has completed the interview in 3 minutes instead of 10. Amazon uses algorithms to let us know what other people have looked at following a search we have made for a product. They have been used for years to match people on dating sites. Lastminute.com uses programmatic technologies to show the impact of the weather or the change in the clocks on customer’s demand for travel. The internet of things is taking shape so that algorithms will have even more power over our buying habits. It is already happening with smart fridges which automatically orders milk when it is running low.
The human touch will still be required to deliver insights and creativity in marketing but it won’t be long before the recognition of market demand will be better served by technology. Stand aside people, the robots are coming.
