Advances in artificial intelligence (AI) are grabbing headlines more frequently than ever. The most recent leap to make global headlines was Facebook’s announcement of ‘automatic alternative text’, where they are using AI to help blind people ‘see’ Facebook.
This is possible because of Facebook’s object recognition technology, which is based on a neural network that has billions of parameters and is trained with millions of examples.
Other artificial intelligence, designed to benefit humanity by surpassing our abilities in highly complex tasks – such as diagnosing illness, researching pharmaceuticals, managing power grids and protecting against cyber threats – could rely for its success on deep learning and the unpredictability that seems to be a necessary part of it.
These breakthroughs in computer technology are rightfully earning the curiosity of marketers who are keen to understand how AI will revolutionise the way media is planned, bought and optimised to enhance the customer experience. It will also mean a break with some parts of marketing tradition.
Back in business school, we were taught that marketing worked by developing a product, assembling the appropriate focus group, testing some messaging, identifying different segments and executing a campaign.
This approach worked in traditional marketing for the past couple of decades. In fact, it is still the dominant approach today. We use this old-world methodology to inform how we run sites, how we advertise on mobile, and even how we approach search. However, it is far from optimal. It’s an example of how analogue thinking has obfuscated our ability to see the full potential in digital.
Audience segmentation is a case-in-point. We still use demographic and psychographic data as our starting point, and then we look at behavioural data like it’s just one more bit of information instead of an entirely new information space. As a result, we see campaigns that were once built around two or three or four segments “explode” to incorporate eight or nine or 10 segments.
It’s one of the world’s favourite statistics that ‘90% of the world’s data was created in the last two years’. Oft quotes it may be, but it does point to the proliferation of consumer data that’s already swamping our human abilities to comprehend consumer behaviours, let alone exploit them.
As marketers, sooner rather than later, big data will get so big that we’ll need brute computing force to experiment, explore non-intuitive links and tell us what is working.
But unlike the days of yore, we wont get the results weekly, because we wont be taking a punt on month-old, third-party data sets. We’ll be using live consumer data combined with real-time second party data to adapt, optimise and learn in the moment.
Every ad call, every bid, every click, every interaction can now be better understood on its own merits. This type of power is compelling enough alone to encourage an industry-wide rethink of digital strategy.
This real-time moment scoring will allow AI platforms to govern the direction of any digital marketing initiative. But it requires a leap of faith, and a trust in technology that we’re yet to see from the marketing profession en masse.
AI means the transparency, the insights and the opportunities are all more available than they have ever been. In my opinion, there’s never been a better time to break with tradition.