The future of advertising and marketing looks like there are a lot less Mad Men, and a lot more (mad?) machines. In a world where computer programmes write the ads and sell us the products, could marketers could become the latest endangered species?
Does the rise of the AI ad writer sound a death-knell to marketing’s creative process altogether? Or does it mean a sharper, time condensed application of data, to target messaging in a way customers actually want?
Let’s take a closer look at what an intelligent marketing future could look like. Someone hand Don Draper a whiskey, because it seems it’s already here…
The creative process doesn’t always run to a schedule. It can be one of the more frustrating aspects of writing or any other artistic endeavor. That’s before you even begin the (potentially long) process of edits and approvals if you depend on it for content.
Marketing technology uses analytics data to gain insights about customers, to understand successful ways of converting interest into sales for their clients. It’s occasionally creepy – how many times have you got an ad for something you’ve only thought about? - but it is effective. Why not leverage emerging tech to do away with the most time-consuming features of creative studio work, too?
Enter Perry Malm, a marketer who was surprised to discover back in 2015 that such tech didn’t exist already. He set up his own company, Phrasee, as he explains:
"Getting the right message had been left to the Mad Men type characters. But I wanted to apply scientific rigour to those messages."
Nothing takes the magic out of the concept of creativity quite like the words ‘scientific rigour.’ Still, there’s a lot to be said for a system of gathering all the information available on a given topic and assimilating it, both to make life easier for writers and to save time in the overall process.
The article continues:
“Standard copywriting takes place through a process of editing, argument and approval.
“Mr. Malm says Phrasee does the same thing using a technique called Deep Learning, a vast network of parameters and pre-set limits that guide the programme in the right direction.
“This allows it to bounce a slogan around, ranking its impact against raw data gleaned from many sources.
“Phrasee builds up models that generate words appropriate to each product and the target audience of consumers. Language that has featured in previous campaigns is poured into the Deep Learning model which applies values based on factors such as the look, feel or taste of a product.”
What is Deep Learning you might ask? Deep Learning works by mimicking the neural pathways in the human brain. This makes the computer “think” in the same way as a human does. Nothing eerie about that at all… By mimicking exactly how our brain would process its thoughts when presented with a slogan or wording, marketers can gauge the potential reaction in advance.
An example of how Phrasee is integrated as part of a wider consultation process with its (human) team is Rapp. The article continues:
“At Rapp, an agency that packages products and services into messages and videos that appeal to the public, Phrasee absorbs a million emails that have been fired off over the years. Then it reassembles those sentences into new messages while adding guidance from Rapp's own writers, technologists and social media experts.
“For a recent Dixons Carphone campaign, Saul Lopes (head of customer marketing at Dixons Carphone) used Phrasee to create variations on messages derived from a decade of sales campaigns: ‘What surprises me is that we still see wild cards leaping out from the list of messages it generates.’
“As Phrasee comes armed with a terrific linguistic arsenal and is divorced from the individual points of view that shape the words of human copywriters he thinks it suits our jaded eyes.”
Of course, coming up with slogans is just one use of Deep Learning. It can be used to save time in other manual, time-consuming processes. CEO of the Marketing Insider Group, Michael Brenner, had this to say about the future of deep learning in marketing:
“One area where AI is likely to overtake human marketers is in SEO. As SEO is a data-driven exercise and search engines use their own algorithms to decide which order to place websites on their results pages, this makes it a natural fit for the technology.
“There is already automatic website generation software available that tweaks design and content in real-time to optimize for SEO. This is not only more efficient than testing and experimenting manually, but also means that websites can automatically adjust for algorithm updates without the time and resources needed by a human SEO expert.”
SEO experts are expensive, and the manual process of optimization is a time burner. Could they be first in the firing line during the zombie robot marketing apocalypse?
But, there are even more options. As put forward by marketer Jason Hall in Forbes, there’s a stream of marketing work that can be cascaded through AI to unify marketing processes:
“Content generation, PPC ads, and even web design are all possible applications for AI marketing. In the world of digital marketing, AI can streamline and optimize marketing campaigns. It can also eliminate the risk of human error.
“At the end of the day, AI is not bound by human limitations. With AI, businesses can optimize their return on investment by only placing ads in front of relevant viewers. Ads can be bought automatically, then personalized at scale. Many advertisers are already using this technology.”
Of course, the biggest difference between humans and AI in the content generation sense is that a robot propagates mimicry. AI can’t create. Creation requires essence, and that’s something that comes from the living. It’s why when you read a piece of AI generated text longer than 300 words something seems a bit…off. The empathy, compassion, and storytelling that distinguish human-generated content can’t truly be replicated. At least, not yet.
Will we live to see a Cannes Lion being accepted by an Obi Wan team of marketers in the future? Only time will tell…