A guest post by Stephen Upstone, Founder of Loop Me
From being the preserve of sci-fi movies, artificial intelligence has suddenly burst through into the mainstream. Once a subject only discussed in universities or by data scientists, artificial intelligence has become the ‘must-have’ of 2017, with companies across all sectors investigating how it can be incorporated into their business models.
The advertising industry, which has long prided itself on being agile and forward thinking, is one of the first sectors to bring AI into its product offerings. The amount of data available to those working in digital advertising – particularly mobile advertising – is what gives AI so much potential in this area. Artificial intelligence thrives from large amounts of data, without data it cannot build models or test its assumptions.
While much is made in the press of the potential of AI to develop advertising creative, or even of the use of chatbots, the real area for AI potential in the advertising industry is in the delivery of digital advertising. Ad platforms process trillions of data points each day and this data, properly analysed and with the right mathematics applied, can be used to predict which users are the most likely to undergo a change in opinion. This enables adverts to only be served to the users displaying the highest probabilities of delivering the desired outcome. As the campaign progresses, the new data created can be incorporated into the AI engines using machine learning, refining and improving the system over time.
The value to the industry of using AI in this way may not seem immediately apparent, but it has the potential to massively improve the value of advertising. It is possible, by using AI, to double the efficiency of campaigns – rather than serving, for example, ten adverts to find a user who undergoes a change in opinion, it is only necessary to serve five. If the UK industry was to adopt this model, it would effectively save £4 billion per year.
This means for every campaign run by an advertiser, they can expect double the performance than if they ran a campaign which does not use artificial intelligence. If the campaign uses AI to optimise against a user’s intent to purchase, this can at least double, if not more, that result. Research shows a 2% rise in purchase intent leads to a 1% rise in footfall – if purchase intent rises by 100% thanks to AI then 50% more consumers could visit an offline store. These are tangible, measurable results which could potentially have a massive impact on brand revenues.
Why isn’t this happening already? The first challenge is that, while AI needs to feed off a large data set, the vast majority of existing platforms are not built to store data. The second is that for AI to be implemented in real-time across campaigns, decisions have to be made in under 5 milliseconds. This means that the AI system has to be built alongside the data management platform, not just bolted on to existing systems, in order for it to work correctly.
Despite these challenges, artificial intelligence is and will transform the advertising industry in 2017, it is already happening. The technology is simply moving faster than the established industry players. Watch out for the innovators, the new and nimble technology companies, to deliver on the promise of artificial intelligence.
Stephen is currently Chair of the MMA UK and member of the MMA EMEA Board. He has worked in mobile advertising since 2006 as Managing Director Europe for Ad Infuse (sold to Velti) and then VP Sales and Business Development Europe for Velti after the acquisition.