In his latest column (dare we say outburst?), B&T’s regular columnist and contrarian, Robert Stohfeldt, says we live in a world of data but few in adland can even scratch the surface. However, he has an ominous warning of Amazon’s pending arrival and its huge (see: we don’t need ad agencies) data in-house capabilities…
Much has been said about Amazon and its impact on not just retail, but business in general. In all the hype, one key point has been overlooked. Yes, they are believers in the power of data, but they have recognised that data scientists won’t be part of the plan.
There is a huge rush to courses teaching data science, but these courses are to genuine data analysis what a first aid course is to a medical degree.
Amazon is employing economists and human behaviourists with PhDs. Why? Science has been using “data” for hundreds of years. But you will not find a “data scientist” in the world of science. Instead you will find scientists, as in zoologists, geologists, chemists, physicists etc. and working alongside them are mathematicians. (Even Einstein, who had a pretty good understanding of maths, needed to work with pure mathematicians.)
A mathematician does not need to understand consumer behaviour. They work in a team with an expert in this area.
Statistics is applied mathematics – in particular, calculus and linear algebra. For all those people rushing to data science courses, you are really learning to be a glorified data entry operator.
Unless you did advanced maths at high school and continued with pure and statistical mathematics at university, you are just scratching the surface. Learning enough to be dangerous. Statistics is not statistical mathematics. A question for the “data scientists” out there. How do you explain i? (The square root of minus one) and complex numbers or eigenvalues. They didn’t teach you that in your data science course? How about a tangent to an affine plane? Basic calculus will teach you about the importance of the tangent line in differentiation. But that is only two dimensional. You will be dealing with multiple variables, hence the need for matrixes and finding a tangent to a plane of numerous dimensions. If all of this sounds like gobbledygook, then move onto something you understand and leave data well alone.
And off the shelf algorithms won’t be the answer. Every situation is different. You will need to be able to work out the Probability Density Function (PDF) from scratch, not just make an assumption the data fits a pre-determined computer model. Then you will need a computer programmer, not just some coding cowboy, who can write the programme for the raw data to be fed into.
Mathematicians don’t understand human behaviour. Nor do they need to. They work in a team with people who understand human behaviour – one does the maths, the other interprets the results.
Our industry has been suckered by the Google, Facebook duopoly because fooling the media agencies with numbers is like the old “taking candy from a baby”. I have sat through any number of presentations by the big media shops as they show off their proprietary analytical tools, which to be diplomatic, are total garbage that could be pulled apart by a 1st year university maths student.
We now have the media dominated by this duopoly who have used media agencies as an entrée to clients. They will be (already are) cutting you out of the relationship. No one needs a middleman these days.
I have yet to meet one person in media who has studied mathematics at university. (There may be one or two out there and if they are, they will understand what I am getting at).
Amazon won’t need ad agencies. They will do all their work in-house. They may engage specialist consultants on occasions, but these people will not be from the advertising industry.
Predicting the future has always inexact. It is fair to look at the past to predict the future, but in so many instances, something pops up out of left field (something no one could have predicted, such as 9/11) which changes everything.
Advertising has become obsessed with technology and in many ways, that is fair, as we live in a world not only dominated by technology, but one in which the speed of progress keeps accelerating. How do we future proof ourselves?
Probably the most important thing I learned from a mathematics degree is that creativity cannot be formularised. Yes, there are people spending huge amounts of time and money trying to get AI to act creatively. Even the term Artificial Intelligence is misleading. People interpret it literally. Somewhere along the line we went from calling them computers to AI, resulting in the perception they think. Well they do not. Machine learning has come a long way, but a computer (AI) can only solve a problem is it programmed to do so. If a person cannot work out how to solve a problem, there is not a computer in the world that can it do for them.
Creative thought is something AI cannot do and the day they are able to, well it will probably mark the end of civilisation. Making a machine that is smarter than us and knows we are a threat to it is not a very smart evolutionary move. (Though many neurologists argue this day will never come due to the basic differences in the working of a human brain and even the most powerful computers.).
Stick to jobs that require creative thinking, as opposed to the repetitive and formularised and you will always be needed. The move to data scientists and letting the foxes into the hen house (Google & Facebook) is edging advertising closer and closer to extinction.