A seasonal retailer is scrambling to advertise a new Mother’s Day promotion. The company only has two weeks to make sure its creative hits the spot with its target audience, but lacks both the time to conduct qualitative research and the money to broadcast ads that don’t work.
Enter Video Creative Intel, a tool from Omnicom Group’s data and analytics division Annalect. Using machine learning and artificial intelligence, powered both by Annalect’s propriety data sets, Google’s data and a client’s own metrics, Video Creative Intel helps marketers know which creative is most effective before a campaign launches. It also helps refine creative assets as the campaign is underway.
“As pieces of creative are running in market, you can start to understand when a piece of creative is starting to wear out [and] when you need to change the creative video,” said Annalect’s chief experience officer Clarissa Season.
Making more data-driven decisions
The launch of Video Creative Intel comes at a time when the line between media and creative agencies is starting to blur. Marketers are looking to make more data-driven decisions, even about creative assets. At the same time, the data that they have relied on for over a decade is disappearing as a result of cookie deprecation, making targeting trickier and the need for compelling creative more urgent.
Season said Annalect has tested the tool with seasonal retailers and entertainment clients (the latter face tight time frames around theatrical releases). Using the tool, clients have found they need to show their branding, as well as introduce humans, earlier in the creative. Another lesson was to include both visual and audio branding, Season said.
Powering these data insights is a partnership with Google. Annalect is the first of the holding companies to get access to this creative intelligence data from Google, Season said.
“Our advantage is that we have taken what Google has done and we’ve added on and created an interface for our clients to get insights out of,” Season said.
In line with the Google partnership, current insights are based on lessons from video assets’ performance on YouTube, Season said. Annalect is working to test the tool’s predictive capability for eventually digital non-video assets—and all places video creative can be utilized; Season added the tool has already been found effective for showing which creative works on connected TV.
“YouTube is cheaper, faster place to get a read,” she said.
What distinguishes Annalect’s new tool from predictive ad tech from other holding companies is that it focuses on optimizing the creative, rather than the media format, said Jay Pattisall, principal analyst at Forrester.
“The audience-first approach to using data to create audiences was something not being used inside the creative agencies,” he added. “In the last two years, they’ve worked very hard in terms of catching up in data maturity and data literacy to media [agency] counterparts.”
And in contrast to other ad-tech-based creative services in the market, Video Creative Intel can take advantage of a wide breadth of data from both Omnicom and Google, he added.
“The potential to save or use your media dollars in a more efficient way is an advantage any advertiser would want,” Pattisall said.