Using Advanced Tech In Advanced Television: Q&A With DataXu's Catanzaro

Sandro Catanzaro, founder and Chief Innovation Officer of DataXu, describes himself as a serial entrepreneur. He was one of the co-inventers of Dataxu’s real-time optimization algorithm, based on research he did at MIT. Here’s an overview of his company, the world of advanced and addressable television and the data work being done in the industry.

Charlene Weisler: What is your definition of advanced TV? Is it the same as addressable?

Sandro Catanzaro: Advanced TV is TV advertising purchased on an impression basis using advanced audience data and software automation, creating additional value for both the buy and sell sides of the transaction.

Addressable TV is a form of advanced TV where households are targeted on a one-to-one basis via cable and satellite set-top boxes, smart TVs and OTT devices, but not all advanced TV is necessarily addressable. Other forms of advanced TV may be based solely on automating the purchase process, but still display ads on a one-to-many broadcast basis.



Weisler: What are the challenges of advanced TV?

Catanzaro: One of the primary challenges in advanced TV right now is the ability to target and provide attribution for OTT campaigns. This form of TV is accessed by the viewer via internet-enabled televisions, and streamed either live or on-demand. The connected nature of OTT makes it very similar to digital video, but as these ads run on actual TV screens and not traditional digital devices (PCs and mobile phones), the typical digital markers (cookies and mobile IDs) are not available for identification, making advanced targeting difficult.

Weisler: How can these challenges be overcome?

Catanzaro: This issue can be overcome through new forms of identity management made possible by cross-device graph technology. A device graph is the unification of several otherwise separate devices, such as a laptop, mobile phone, tablet and smart TV under one unique household. In the real world, these devices don’t exist in a vacuum; they are linked through ownership and usage and can be used to understand the full context of a person’s digital footprint.
By including smart TVs and OTT devices in a device graph, marketers are able to leverage advanced audience data, built using legacy digital IDs, to enable addressable targeting on televisions, even though these legacy IDs are not present.

Weisler: What metrics do you use?

Catanzaro: Advanced TV is typically purchased on an impression basis as opposed to ratings, but it is possible to provide traditional TV metrics such as GRP. However, we find that marketing professionals are also frequently leveraging more detailed metrics to prove success, such as lift studies which compare conversion rates of exposed populations versus control groups. This is made possible through addressable forms of advanced TV and device graph technology, where specific viewers are directly targeted and others are intentionally excluded, in order to compare their actions across all devices — and, in the real world, after having viewed the ad.

Weisler: What is Dataxu?

Catanzaro: Dataxu is a software company that helps marketing and media professionals use data to improve their advertising, using AI to optimize ROI on marketing investments. Dataxu ingests first-party data (e.g., customer purchase information), matches it up with many other kinds of data across devices and identifiers and creates a customer machine learning classifier for each campaign that invests more budget into what’s driving acquisition and less into what isn’t.

Weisler: What will the future of TV buying look like three to five years from now?

Catanzaro: We expect that in this time period all major TV programmers will have made their content available via connected devices and most will be providing marketers access to this advertising via programmatic channels. The largest portion of TV buying will still operate through legacy methods, such as upfronts, but advanced TV will soon constitute a much larger piece of that pie in the near future.

Next story loading loading..