Spotlight On: Jason Kodish, Global Data Scientist, DigitasLBi

kodish-jason-slightly-croppedJason Kodish (JK) leads DigitasLBi’s 350-person strategy and analytics team. ClickZ caught up with him to hear his thoughts on the intersection of data, analytics, and content.

ClickZ (CZ): On the Digitas Distillery blog you wrote that new devices are creating “new types of analytics for understanding audience behavior.” What are these new types of analytics? How are you going to use them?

JK: In the past, the industry has used surveys or focus groups for inferring consumer behavior. But now, with the advent of many new mobile and non-mobile technologies, for the first time we get really good insights into actual behavioral data. Most of our clients are trying to drive a behavioral change — transactions, relationships, retention. Now the data that devices are enabling, stuff like geolocation data, social media, content consumption, the use of search terms, and other unstructured data, gives us a much more precise view of behaviors that drives a better understanding of our consumer. What are you doing with all this new data? Historically, the response to data has been to use that data for programmatic buying in the digital arena. What’s now more interesting than just programmatic buying is that by using this behavioral data, we actually have the ability to do more intelligent and precise media planning. By media planning, I mean all channels: the development of content, the development of media plans, the development of environments and CRM plans in offline environments.

CZ: How does new data interrupt old planning strategies?

JK: Most media planning companies have really smart, sophisticated media planning tools, but the problem is [those tools] sit on survey data that asks questions like, “When you bought the pair of shoes that you’re wearing today, what was the role that radio played in your purchase decision?” As much as you want to be able to answer that question, the fact of the matter is, you can’t. You don’t know what role radio played. But by tying multiple sources of behavioral sources of data, media exposure, social data, geolocation throughout the day, we get a much more fundamental picture of what your experiences were before the purchase of those shoes and start to craft experiences that are more valuable for people in the market for shoes to engage in all channels, not just digital channels.

CZ: What are your biggest data points?

JK: We’re in the process of putting together a really interesting asset for Digitas. It combines multiple sources of first-party data in a safe, compliant environment. We run five sources of data: geolocation, digital data, TV exposure, social data, and transactional data and combine all five of those different views into one macro-environment stripped of personally identifiable information (PII). It gives us this reservoir of data with really great insight. We do a lot of better targeting of creative that way, better establishment of offers and better development of content that’s more in line with what people want to consume.

CZ: What’s your KPI process like?

JK: We don’t have the luxury of saying “I don’t know.” The main reason we are integrating data across multiple sources is because not only does it give us the ability to do the front end targeting and development of content and creative, but just as importantly, it aligns all the data that we need to put exposure and transactional data into one place. This reservoir of data allows us to take the notion of attribution away from just being digital. Understanding which people are exposed to different touch points, either proactively or reactively, we know who’s being exposed to what online and offline. We know their behaviors. Integrating transaction data with exposure data gives us the ability to attribute business success to any of the points of engagement with the consumer. It’s a nascent process, and we’re seeing some really interesting interactions between the different media.

CZ: What kind of interactions?

JK: There are a lot of things that we see that show the effectiveness of search when broadcast is being run as opposed to when broadcast is not being run. Some of the really interesting overlays that we’re seeing right now are interactions between digital consumption on either a mobile device or a desktop computer and TV content consumption. It’s redefining in some ways whether things like search and digital media are truly success events or if they’re actually just capturing success events that are generated from other channels.

CZ: How do you use these interactions to create quality actionable deliverables to your audience?

JK: The more people are exposed to [these behaviors] the more they’ll see [the relationship between search and broadcast] is not coincidental. The Golden Globes are a great example of the fact that people were following fashion online as they were watching the Globes, which gave us the ability to link households and serve mobile and social ads to devices while that was happening. What’s going to be more interesting than that is when that goes the other way. As it stands right now, only about 2 percent of TV is bought at a programmatic level. But as that number grows, and that number is going to grow significantly over the next through years, the dynamic inclusion of ads in TV will mirror what people are doing online.

Related reading