Skip to main content
dstillery dstillery


Read the full article by Tobi Elkin on MediaPost

It’s a perfectly relevant question given today’s insanely automated media world in which programmatic media buying is the name of the game. Data scientists, predictive technology, artificial intelligence, and machine learning might seem like they’re  all ganging up on creativity. But what if they could all work together? What if the technology and the data derived from it could inform creative strategy in a way that makes sense?

What if data scientists contributed to the creative insights at the outset of campaign planning? That’s the question Claudia Perlich, chief scientist at Dstillery, a data-driven, cross-channel marketing firm, is mulling.

Perlich aims to establish a meaningful collaboration between creativity and data science. “Given the amount of information we have on individual behaviors, we have a better understanding about the true personas of consumers and who’s interested in a product. The ability to look at the different factors that get people interested in a product on a granular level should be part of the brainstorming process on imagery and messaging,” Perlich said.

Further, getting data science involved at the outset of a campaign enables all parties to experiment with ideas—to look at all the different creative messages and see how they play out. “I might be a data scientist, but I think I can contribute to the creative process as well and broaden the perspective,” Perlich said.

Real-time events like the Super Bowl and the Oscars offer opportunities to experiment and pre-test ideas. Perlich said testing enables teams to see the impact brand messages can have on everything from packaging to out-of-home media. Ad campaigns  during these high-profile events can benefit from thinking about programmatic as a tool to understand customers better.

“With events like the Super Bowl and the Oscars, we’re talking about reaching large audiences. You’re still showing the same message to everyone. With programmatic and addressable TV, you’re not showing a single message to everyone. The takeaway is that we have to move toward a more personalized way of interacting with consumers,” Perlich argued.

Programmatic, Perlich said, offers the ability to change what people see in real time, on the fly. The problem is that this doesn’t really exist for TV yet. “We haven’t been able to understand behavior with TV consumption. We don’t know in any detail what TV watchers do on different devices,” she said. But, she projected that this capability will evolve over the next three to five years.

“Brands are using programmatic ad delivery on digital platforms. I see programmatic as a great place for experimentation. Think of programmatic as a digital real-time focus group.”

Perlich offered the example of client Chobani, the yogurt brand. Dstillery worked with the brand’s creative agency last year to design messages beyond Chobani’s comfort zone. “We used those to expose them to different groups and populations. We saw people who spent time at the gym had a better reaction vs. people who went to quick-service restaurants.  There are micro populations that respond to messages very positively,” Perlich said. While the messaging was in digital media, the campaign  connected to in-store purchase via Kantar’s ShopCom loyalty card. The point, Perlich said, is to experiment with different creative messages and use programmatic to see whether they have an impact.

In addition, Perlich is interested in dynamic ad insertion: “it should be used a lot more in programmatic to gain insights into how people respond to messages.”

Since consumer attention has become a “fickle beast” and “people tune in and out of content very quickly, how do you engage with your audience? How quickly do you need to grab their attention? People are jumping from device to device. The moment ads come on, they drift away to devices,” Perlich noted. And even on those devices, ads are appearing.