DigitasLBi has named its first-ever chief data scientist.
The Publicis Groupe agency’s Jason Kodish steps into the newly created role from his former position as SVP, North America, strategy and analytics lead.
In the new position, Kodish will manage a global strategy and analytics team of 350 and report to global CEO Luke Taylor and North America CEO Tony Weisman. He will focus on new and existing solutions, such as LookOut (competitive monitoring), PeoplePulse (audience data) and BlackBook (influencer targeting).
Speaking to AdExhanger, Kodish said that DigitasLBi has invested significant resources over the last couple years developing tools that integrate more data in all aspects of business, from planning and buying to strategy and creative.
“I’d like to broaden the practice of data science beyond traditional uses,” Kodish said. “Data science is no longer in the realm of just direct and digital, it belongs to all media. My primary goal is to insure that our industry clients, really successful marketers, accept and drive their businesses through data science.”
DigitasLBi in North America has a long heritage of analytics, but those practices have not been centrally managed. “One of my responsibilities will be to unite the agency’s global efforts, so we’re all working from the same playbook,” he said.
“A majority of the products that we build are for media, and the analytics that we do is for media, whether it be for planning or buying, or for trafficking and targeting,” Kodish said. “One half to two-thirds of my job, if not more, will fundamentally revolve around media.”
According to Kodish, multiple elements of data science help bolster the agency. One is data integration.
"Understanding how new data sources tie together to create a single view of the consumer is a really interesting opportunity to develop in the industry," he said.
Another is modeling.
“Traditionally, really sophisticated digital models have been used within direct mail, catalogs and so on, to simply understand digital," he said. "But fundamental statistical modeling, whether it be frequency or response, can now be brought to more media than ever before, such as addressable TV, digital out of home or geolocation. There’s an opportunity to make all kinds of media much more effective and much more efficient.”
A third is omnichannel programmatic buying across TV, radio, addressable outdoor and so on.
“The biggest element right now, in my estimation, is that as the inventory continues to evolve, programmatic will benefit from data science," Kodish said. "Programmatic cobbles together more data sources that allow us to bring more information into the equation to do more decisioning.”
From the perspective of a data scientist, and in light of recent a global focus on privacy, data collection and legality, Kodish added that privacy is at the core of his work.
“Privacy is fundamental,” he said. “Privacy is the first issue, and it’s the last issue that comes up in the work that we do. What good, fundamental data science will help us do, and what will drive the industry, is giving us permission to take on more types of business problems, beyond just advertising and in a way that respects privacy rights.”
Prior to joining DigitasLBi in 2006, Kodish served as director of strategy and marketing for GSP Marketing.