The AI future of marketing: Could Don Draper be replaced by a robot?

TV programmes like Panorama have explored how service industry workers are being replaced, more and more, by robots. But, we are told, marketing is a creative exercise and it requires a creative mind – we’re safe in our seats. Except that isn’t true. Automation is a big part of marketing already and it is likely to become a much bigger feature. There are jobs in marketing that can be replaced by automation, and there are some that perhaps won’t be just yet – but they will be.

“The Oxford English Dictionary” describes marketing as “the action or business of promoting and selling products and services”. Notice how that definition does not say anything about winning awards for clever commercials or amazing people with your design skills. It is about selling stuff. The way to sell stuff is to find all the people that might want to buy that stuff and then show it to them. Why can that not be automated?

In fact, there are several areas of marketing that are already automated. There’s Automated Insights, a software program that helps Associated Press to publish company financial announcements or search marketing reports without human input. Marin Software manages thousands of ads and keywords, making micro-changes constantly without human input, while corporations have begun using automated triage for customer enquiries – using algorithms to read the context of emails and route them to the right people. 

Large retailers also have clever software that automates content and pricing depending on user behaviour. Why show every customer the same products on your home page when you can automatically show them what they are more likely to buy, without having to pay hundreds of human merchandisers?

Consider Don Draper, from Mad Men. His ability to understand what pushes the emotional buttons of a customer leads to advertising campaigns that get results. Nowadays a computer could get better results. If we analysed results from one of his ad campaigns, then created an algorithm to analyse big data to sell the same product, the algorithm would eventually out-perform him. This has been proven time and again. Chess computers started out being worse at chess than the world’s best grandmasters. Now we have chess programs on our phones that can beat the best in the world.

WalMart famously discovered, by analysis historic trends, that the biggest selling item before a hurricane was Strawberry Pop Tarts rather than water or beer. Would Don Draper have been able to predict that? And what will an automated marketing future look like?

Self-driving cars are safer than cars driven by humans. This is because computers are better than people at any repetitive, mundane task. The creative Don Drapers and artistic directors who invent clever marketing campaigns may not feel threatened by algorithms, but not taking algorithms seriously is a threat to them. So we are a few years from being able to create machine-learning algorithms that can beat Don Draper at his own job. If you don’t believe me, consider this.

IBM’s Watson AI platform uses natural language processing and machine learning to answer questions and make recommendations in many of today’s applications. Retailers, health companies, and recruiters use it to personalise the experience of users automatically. It does this by accessing big data – masses of unstructured data.

The key difference between Watson and Don Draper is the data. Draper’s data is based on the knowledge and experience he has gleaned through his own years of doing the job. Any decisions he makes about things he has no data for rely on educated guesswork. Watson, on the other hand, has billions of data records from all manner of sources. It can process that data to look for correlations that enable it to make accurate predictions about the kinds of things that will have a positive impact.

Read on to find out the potential jobs it could substitute.

Share this story