AI in Digital Marketing: ‘This is the Land Grab Era’
7 March, 2019
An insight from the ‘AI in Advertising’ panel at MWC featuring our Global CEO, James Hilton, by Junior Research Analyst, Anna Fitzpatrick
Searching the term ‘AI’ is met with endless articles of out of reach possibilities which have no real-world application. Behind this smoke screen, AI is having an actual, tangible, impact on the advertising industry. Our global CEO, James Hilton echoed this sentiment at MWC where he joined the AI in advertising panel. “I want to dispel the myth that AI is some futuristic thing that we’ve got to spend a fortune on developing”, Hilton explains- “it’s not. We’re using it”.
AI is infiltrating technology in ways not obvious to the consumer. Netflix and Spotify are key examples, using machine learning to personalise the user experience. Netflix makes recommendations based on ongoing behavioural predictions, going beyond segmentation and demographic groupings. As Hilton notes, quality personalisation is now inseparable from consumer expectations. “We don’t want to go on to Netflix and spend ten minutes finding the content we want to watch. That’s natural” he continues, “and over time, we are going to expect it more.” Consumer expectation now necessitates the use of machine learned personalisation. Following this expectation, AI is not only being used but is becoming ingrained into how we experience technology. As Hilton reminds the audience, “this isn’t going to be here today and gone tomorrow”, AI is here to stay.
From a logistical perspective, AI is crucial if marketers are going to combine the vast amount of data that they collect. At MWC, Hilton explained that marketers must use machine learning ‘because we are collecting so much data…which has to be processed in milliseconds.’ Take the example of Netflix again. The streaming service has 125 million members. To keep content timely and relevant for viewers, it must achieve personalisation at an unprecedented speed and scale. The same logistical challenges apply to ad fraud- an operation that is set to cost the advertising industry an estimated $50 billion by 2025. It is unrealistic to expect ad fraud will ever be completely eradicated, as Hilton points out, ‘everything we do, they’ll do as well’. However, it is the dynamic nature of ad fraud that makes machine learning, as Hilton continues, ‘intrinsically valuable’ in combatting it. The evolving nature of fraud can be challenged by the predictive capacity of machine learning- which allows us to not just detect ad fraud effectively but begin to prevent it. ‘If we could freeze time’, Hilton notes, ‘we could do this manually. But we can’t.’ The use of data is accelerating at a rapid pace. Marketers need AI to keep up with the sheer volume of it. If you don’t use machine learning to optimise your marketing, the question is not if, but when you will. AI is inescapably part of the present and it is, as Hilton states, ‘part of everyone’s futures.’
The implications of AI for performance marketing go beyond predictive personalisation and preventing ad fraud. As machines learn, it is crucial to ask who is teaching them. Hilton reminded the panel of this at MWC, pointing out that the learnings of AI are largely concentrated in the hands of tech giants Alphabet, Amazon and Facebook. As digital marketers necessarily rely on machine learning to survive the information economy, they will therefore rely on the services of the tech oligopoly, imbedding their dominance even further. As Hilton pointed out to the panel, ‘this is the land grab era’. This is when the rules are getting created; and how we code the future is up to us.
If you would like to hear more about how we harness AI to deliver the best performance marketing, get in touch today.