Over the last 30-odd years, marketing has evolved from a rough collection of untestable hunches into something closer to a science. This didn’t happen overnight. The rise of the internet and personal computing ushered in the world of data-driven marketing. As consumers generated more and more data, marketers could track customer behaviour, pick out trends and measure campaign success with frightening accuracy. By the time smart phones became ubiquitous, there wasn’t much marketers couldn’t learn about their customers.
But like most industries based on number-crunching, the problem quickly became one of overabundance. By 2020, it’s estimated that 1.7MB of data will be created every second for every single person on Earth. Put simply: there’s too much customer data for the average marketer to analyse.
Which is where artificial intelligence comes in. Most experts see AI marketing as the new frontier. The holy grail of marketing. Technology that will somehow fuse traditional marketing techniques and advanced data science, turning them into actionable, customised, ruthlessly efficient sales tools.
But what does that actually look like? And where’s the line between AI marketing and science fiction?
Salesforce leading the way
Just like it did in the 1990s with SaaS technology, Salesforce is on the cutting edge of AI marketing. Their Einstein AI has recently been unleashed on the Salesforce Marketing Cloud. Why is this a big deal? Because Einstein goes way beyond simple EDM automation and bespoke customer journeys. Salesforce can now measure all sorts of things. How many emails each person will accept before getting EDM fatigue and losing brand engagement. The optimal send-out time for each individual customer, based on web and browser activity. Automatically tagging libraries of existing content, then feeding those into relevant EDMs with AI image recognition.
Real time possibilities
Big data allowed marketers to get inside customers’ heads. But a better analogy for AI marketing is like watching someone’s thoughts. Seeing the neurons flash, then tailoring business activity to match. This might take the form of machine learning analytics, which map customer activity and engagement in real time. Or it might be smart social media listening tools, which let marketers crawl the web at supersonic speeds – listening to brand conversations and monitoring for positive or negative campaign feedback. Ben Ellis from BT and Microsoft has talked extensively about this before. “In a typical week, your brand and competitors will generate 20 major peaks,” he says. “I’d say an analyst spends around three hours conducting this analysis – in other words, a whole morning. AI would take less than two seconds.”
Product and content recommendations
As SaaS companies gradually took over the internet, the idea of personalised content became almost mandatory. Unfortunately it took a few years for technology to catch up. The data was there, but crunching it required the sort of processing power (and intelligence) you only get from GPU-driven AI. This started with companies like Amazon, Netflix and Spotify using AI to generate suggested content. But some companies are pushing the concept into strange new worlds: Sky Cinema has already implemented AI marketing that recommends movies based on the viewer’s current mood.
Efficiency at scale
AI is already been used to improve account selection in certain account-based marketing companies, especially when ABM is carried out at scale. If your B2B company approaches thousands of potential customers each year, then any tiny, iterative improvement will slowly accumulate into massive advantage. VP Jessica Fewless from American personalisation company, DemandBase, says AI helps them quickly filter unsuitable clients, or clients who are statistically likely to jump ship and join a competitor, saving the company money in the long term. “We took the elements that made those customers churn and removed them from our model,” she said.
Google has known about the power of AI marketing for a long time, particularly in regards to SEM. They started innovating with RankBrain, a machine-based learning algorithm, back in 2015. RankBrain made Google’s search more relevant to searchers and (therefore) more profitable for marketers. And other companies have followed in Google’s digital footsteps. Amazon has already built smart search indexes and AI models into its search bar, spitting out more relevant products (based on customer browsing patterns and purchase history) to increase overall conversion. Any e-commerce company can easily do the same.
What you see is what you get
AI has also opened up a whole new world for marketers. One that early 2000s marketers could only dream of – image recognition and visual search. The idea that you can search for something that looks like something else. AI marketing tools have become adept at quickly seeing an image, analysing what it is, then offering similar alternatives. We’ve already seen this pop up all over the marketing world, from Pinterest (which is basically one giant image recognition and synthesising program) to Google Lens and ASOS. For any marketer that deals in visuals, this is like been gifted a sixth sense.