The 2019 IBM Think Conference in San Francisco offered several glimpses at how the venerable technology company and its partners envision the future of advertising. With blockchain, artificial intelligence and more, IBM expects a more efficient market as it becomes easier to track the flow of ad dollars from brands to consumers and as consumers become easier to track across multiple devices and online identities. All of which is sprinkled with what CEO Ginni Rometty called “random acts of digital and AI.”
Here are three key takeaways from the conference:
Blockchain-enabled advertising network moves forward
In January, IBM and Mediaocean rolled out an initiative for brands, agencies and publishers to better track their campaigns, expanding on a pilot program last summer. Although results won’t be announced until the Cannes Lions festival in June, there was a lot of optimism at the conference.
“With the blockchain, potentially, you’ll have a dashboard for every marketer to see every part [of their campaign spending],” said Babs Rangaiah, executive partner, global marketing, IBM iX.
The program uses a custom-built blockchain developed by IBM integrated with Mediaocean’s campaign-management platform. The initiative includes partnerships with eight major brands, including Kimberly-Clark, Johnson & Johnson, Pfizer, and Kellogg’s, whose CMO, Gail Horwood, said early results were encouraging.
“I don’t expect a solution in June,” Horwood said, “but over time, we will get closer.”
By allowing marketers to better track their spending, the program aims to make ad spend more efficient and enable compliance with regulations like the European Union’s General Data Protection Regulation.
Building regulatory-compliant consumer identity
GDPR, along with a similar bill passed in California regulating online privacy, was a hot topic at IBM Think. Heather Blank, general manager, data products at MediaMath, saw a chance to leverage an “identity stack” for advertising as a way to remain compliant with regulators around the globe as well as to increase efficiency.
“There’s a lot distrust on the consumer side,” Blank said. “We as marketers and ad tech have also made a mess—fragmented device recognition, data loss, inefficient media spend.”
MediaMath began the project a year and a half ago by creating deterministic user matching—only identifying a user across multiple platforms when it could be certain. But by moving to a probabilistic model—using AI to make high-degree-of-certainty estimations of user identity—the company had more luck. “We analyzed 400 campaigns and found a 19 percent reduction in cost per acquisition from probabilistic versus deterministic,” Blank said.
Modeling user identity lets advertisers reduce costs by narrowly targeting their messages, said Ravi Shah, IBM Watson marketing and commerce, product director. “If a wallet company is bidding on my impression when I want to buy a wallet, it should bid more than a purse company.”
Using Watson predictive models, IBM is piloting an automated system, a media optimizer that identifies users in real time while bidding for online ad space. “We applied it to our own corporate advertising at IBM,” Shah said. “We had a 65 percent reduction in CPA on average in two weeks.”
AI works its way into more aspects of business
More than 2 billion messages passed between brands and consumers last year. It’s a space IBM sees artificial intelligence reshaping. Since 2016, IBM has run advertising experiences using Watson AI capabilities, including in the food, travel, toy, and home and garden sectors.
The Behr Paint Company, whose vice president of digital marketing, Tanuja Singeetham, reported on how her firm used IBM Watson ads to build a color-recommendation engine for consumers. “Watson asks questions—why are you painting your place? How do you want the room to feel? The overall look?” she said. It then uses natural language to deliver its recommendation.