AT CES today, IBM announced Advertising Accelerator with Watson, using the brand’s famous AI capabilities to predict optimized creative while also uncovering new audience segments. As DMN reported, IBM Watson Advertising rebranded in 2017, from the digital assets IBM acquired when buying The Weather Company. (The Weather Channel, it should be noted, is owned separately as a cable network, by Entertainment Studios.)
In an exclusive interview, Randi Stipes, CMO of IBM Watson Media and Weather, told me that this new platform, currently in beta, “offers a new approach to engaging and converting customers.”
She added, “It’s a really important part of our client strategy. Clients are aware of the importance that weather plays in consumer behavior. Weather intelligence helps advertisers and brands understand the right to run campaigns and target audiences, how in Texas their response might be different than in Georgia.”
And of course weather news is an extremely valuable form of content that attracts a wide, engaged audience. It generates customer insights far beyond just weather-related behaviors. These insights are used to help IBM clients convert customers during a campaign.
“We know audience conversion is the holy grail,” Stipes stated. “Even if you reach consumers, it’s tough to convert.”
This is where dynamic creative optimization (DCO) comes into play, along with audience segmentation. Advertising Accelerator with Watson measures the response that specific pieces of creative achieve with individual audience segments. The platform’s AI continues to learn and predict what creative will work best with the most responsive audience, and helps advertisers plan for future efforts.
“Brands are craving more insights and spending so much money on campaigns,” Stipes explained. “It can’t be a one-and-done. Advertising Accelerator with Watson learns from campaigns and uses insights to apply to the next campaign or strategy.”
According to Stipes, IBM ran an alpha test campaign, using its own radar app, Storm Radar. The test ran 23 days, using app installs as the primary KPI. During the test, they found that specific creative performed at a rate of three times higher when shown to certain segments in particular regions. The segments were broken down by market, what time of day (the morning commute, for instance) and what device they were using, along with other variables.
The DCO worked by using a standard three-by-three-by-three technique, using all combinations of three different headlines, ad backgrounds and calls to action.
“All those permutations are put together on the fly, and then we test,” Stipes said.
The beta offering launched at CES includes a limited number of brand partners, included LendingTree. All campaigns will be executed across IBM Advertising’s owned and operated weather properties, “so we can test and learn and have that exist within a controlled environment,” Stipes told me.
Later in 1Q, brands and agencies will be able to use the technology with other publishers.
“Once we really refine the AI, then our philosophy is to give brands the ability to scale that technology and run their campaigns wherever they see fit,” she said.
“Brands are still looking for ways to use AI in their business and that’s what Watson Advertising is all about, taking AI and making it more scalable,” she added. “And similar to what we’ve done with weather and weather insights, we wanted brands to use AI to uncover insights about their business, and [Advertising] Accelerator [with Watson] is the next step.”