The machines are coming. Whether it’s Siri on your phone, the Amazon Echo, or fighting cancer, AI is here in the now and entering our lives in all sorts of unexpected ways. Even Barack Obama acknowledges its impact, noting in a recent Wired interview: “that it has been seeping into our lives in all sorts of ways, and we just don’t notice”. But what is it? How will it impact the work of marketers? And will it change ad tech as we know it?
Algorithms make the digital world go around
The most familiar component of AI, which already impacts everything we do, is the humble algorithm. Algorithms are at the heart of our modern economy but what are they? Basically they’re a sequence of steps that lead to solution of a specific task; the most famous being Google’s search algorithm. So in effect, an algorithm is a recipe with a series of stages and an outcome of tasty product at the end. Yet, unlike Jamie Oliver, a digital algorithm will go through a recipe 7,000 times per second (and not take longer than expected). The role algorithms play in the world economy is not to be underestimated. Take the trading that’s done on the Nasdaq: more than 60% is done through algorithms. The wolves of Wall Street are now more likely to be data obsessed Ivy League scientists than Gordon Gecko style mavericks.
In a sense, algorithms really do make our world go around. They are the engines that drive digital giants Google and Facebook. It’s through their use of algorithms that users are able to view what is most relevant. A huge impact on the work of marketers. Think about how important SEO and a Google ranking is to a company. Or how crucial it is to make sure your Star Wars battery ad appears to someone who actually cares on Facebook. For BannerFlow CTO Magnus Villson it’s clear: “The world is going into an algorithmic economy – some of us are aware of this. It is not controlling our lives but helping us live our lives”. Without algorithms our digital lives would be very different. But how can we utilise them further?
AI and algorithms will likely help marketers by increasing our ability to personalise. Currently Data Management Platforms (DMPs) can choose ads based on the weather, location and time of day but by using AI you can go deeper. Algorithms could use an ad publisher’s data to offer unique insights into a specific audience. Gender, age, profession and interests, could all help to select the best ad to suit a particular individual. According to Jamie Evans-Parker, founder and chief executive at wayve: “Combining the best ad based on the publisher data set and layering that with dynamic data sets — time, location, weather – is the closest the industry’s come to bespoke advertising yet”. For marketers, targeting and retargeting just got that much more precise.
Machine learning: elementary my dear Watson
IBM’s Watson is no doubt a machine you’ve heard of: remember back in 2011 when it competed on the US game show Jeopardy? It beat the human contestants. Watson is one example of machine learning; a computer system that is able to learn and use analytics to adapt its own algorithms to make predictions. Watson also has the ability to understand natural speech, hence why IBM selected Jeopardy to show off the technology. Watson is able to call upon hundreds of millions of web pages, and through machine learning, change its algorithms, hence it could get a handle on the tricky questions of Jeopardy.
Today, Watson lives in the cloud, its computing power coming through apps. And since 2011 it’s been implemented across a variety of industries, with IBM tailoring Watson for a range of applications. The system now helps North Face customers choose a new winter jacket, works as a Hilton concierge, and assists doctors making critical diagnoses. It’s even designed a dress. The cognitive skills that Watson learned for Jeopardy have, in a short space of time, found real-world purposes. For ad tech Watson is an example of how a complicated system can be learned then refined by a programmable machine. Imagine Watson assisting you in finding the right type of images to suit a specific audience in your next display campaign. That’s on its way already.
Neural networks offer a glimpse of the future
Artificial Neural networks are the next level of machine learning. Like the name suggests it is a form of AI that tries to recreate how the human brain works: made up from many connections, linked algorithms and layers of complex data. Neural networks require training and data to perform the (usually) complicated task they perform. The best example of neural networks in action is in the self-driving car. All self-driving cars, from Google to Volvo, rely on a neural network to make sense of the world around them. At their core are different types of algorithms that make decisions based on the data supplied. And this isn’t science fiction! Last Autumn Tesla cars received an update – this patch allowed the cars to drive themselves. Overnight they became semi-autonomous.
One day neural networks could help marketers with their online campaigns. Picture programming an ad campaign with an AI assistant. Google's DeepMind is a tantalising glimpse of that future. It can not only recognise and analyse images in a video but in one publicised example, discovered cats. Yes, through its deep learning algorithms, without human guidance, it figured that YouTube is the home of cat videos. It identified 10 million of them (no doubt including my all-time favourite: keyboard cat). While this is amusing now (and creepy if you visit its offshoot Deep Dream) it could help revolutionise ad tech. Imagine a tool that could take the data it receives and understand why a campaign is not performing. It could make recommendations; it could even make suggestions about content; it could even write that content! But marketers need not get worried just yet! Harry Armstrong, senior researcher at Nest, predicts that the changes to how marketers work will be a “subtle shift rather than a sudden disruption”. Phew!
Whatever developments occur in the future, marketers need to be ready for the opportunities and challenges that AI poses. It’s clear that ad tech will benefit from further embracing data science, while the benefits AI will bring to programmatic advertising and tools like BannerFlow will be immense. Though one comic word of warning: some tech might be too smart its own good, as suggested by a rogue Amazon Echo. Turns out it’s not Skynet terminators we need to watch out for but thieving voice activated speakers.