As artificial intelligence (AI) starts to be used in digital advertising – an industry famed for its tech-speak smoke and mirrors – it’s safe to say that we’re in for a confusing few years when it comes to defining what exactly it is we’re talking about. Many people still think that AI should be a term reserved for technologies that can mimic or surpass human intelligence. However, technically speaking, you don’t have to have built a sentient being in order to lay claim to having AI technology, and machine learning – the practice of using algorithms to parse data, learn from it, and then make a determination or prediction about something – is regarded as a subset of AI. One of the ad tech companies positioning themselves at the vanguard of the AI in advertising conversation is LoopMe, a mobile video DSP headquartered in London. Here Stephen Upstone, LoopMe’s CEO, discusses AI and the impact he thinks it could have on the wider ecosystem.
Artificial Intelligence (AI) is a term that confuses many people and there’s a lot of confusion about how AI can be used in advertising. When you talk about using AI in LoopMe’s technology, what are you referring to?
True artificial intelligence can be applied to advertising in a variety of ways, incorporating chatbots into the creative, or even developing the creative itself. LoopMe uses artificial intelligence to calculate the probability of a user being influenced by the advert at that moment, if the probability is high enough the ad is served. The AI then uses machine learning techniques to improve the quality of its calculations over time. In our case AI allows us to deliver better results, for example a recent native video campaign delivered 67 percent video completion rate (VCR) because of AI optimization, compared to 20 percLent VCR for the manual optimization control group.
LoopMe recently launched a new product aimed at boosting brand metrics like purchase intent or in-store footfall. Could you explain how this works?
The PurchaseLoop product works by surveying two populations of user, one which is not shown the ad and another which is, and asking them a question about the brand. This could be related to purchase intent, favourability, recall – whatever the brand wants to measure. The AI engine then uses the survey responses of both the control and exposed groups to build a predictive model to enable the system to directly target the people most likely to change their minds after viewing the ad. This model is then use to target the audience going forward. The effectiveness is measured using a third survey on those people who have received a targeted ad. For foot traffic, location data is used in place of the surveys.
Is LoopMe still committed solely to mobile video advertising? Do mobile-only solutions still have a place in an industry that is increasingly cross-screen?
Mobile is the most important screen, not only is it the one we use the most often and therefore should take pride of place in a brands’ advertising strategy, but the data it provides is invaluable. Mobile is the only device which can deliver attribution through surveys, foot traffic or online and offline purchases, proving ROI for brands.
Our platform already sees traffic from desktop and connected TV, and our data science team is exploring how to use data from mobile to improve ads on other screens.
As a buy-side company, what would your advice be to companies on the sell-side seeking to compete with ‘the duopoly’?
Clients are looking for the best technology available to the market today, as well as complete transparency, viewability and brand safety – offer something really special and deliver on those three aspects, and the company will be in great shape.
Google and Facebook both have extremely credible advertising businesses, but they aren’t the only big competitors. Other major players are those who hold data – Verizon in the US, AOL, Yahoo, Amazon, Baidu, France Telecom, BT/EE, Adobe and Salesforce, all have access to vast quantities of data and with the correct application of AI could transform the industry. Any companies looking to disrupt the market would be wise to also look closely at these businesses.
However advertising is incredibly fluid, ten years Facebook was still in its infancy, we’ve seen companies like Snapchat become billion-dollar businesses in just a few of years, so a new rival to Google or Facebook is perfectly possible. Companies looking to compete in the space should look to develop complementary offerings to the big players, or create a truly unique technology product.