Juniper Research (www.juniperresearch.com) has found that machine learning algorithms used to enable more efficient ad bids over RTB (real-time bidding) networks will generate some $42 billion in annual ad spend by 2021, up from an estimated $3.5 billion in 2016.

Machine learning — a subset of AI — has arrived at a point where it is both accessible as well as affordable to a wide range of stakeholders. Juniper anticipates that the technology will eventually permeate into nearly all industries in the next five years.

In the case of the media industry, machine learning is being used to develop so-called “bots” and digital assistants, as well as maximize returns on digital advertising. In the case of the former, companies such as Facebook and Google are leading the drive, with the likes of Rocket Fuel and Datacratic developing innovative solutions for the latter use case.

Juniper’s study found that current RTB bidding mechanisms are based on segmentation, as opposed to the individual, while rules for determining bid amounts are often rudimentary. Machine learning promises to transform this segment of the digital advertising market, in that algorithms are able to predict the success outcome of an impression, and thus adjust bid amounts dynamically, according to the research group.

For example, the fact that the end-user has recently been delivered a similar advert may impact their likelihood of clicking on another. Meanwhile, other behavioral and contextual attributes are used to predict how successful an impression may be – on an individual level.

“Typical RTB allows the advertiser to target demographics or various population subsets,” said research report author Steffen Sorrell. “Adding machine learning into the mix effectively allows RTB networks to target the individual. This is a much more powerful tool.”

In addition, the study found that machine learning is likely to lead to an era of “fully personalized”ad delivery. This concept was demonstrated last year by M&C Saatchi via their adaptive digital poster in London. Meanwhile, advances in machine learning will enable computer systems to understand both the images and text delivered to web pages, offering an opportunity for unique ad campaigns, according to Juniper.