MediaWave Actionable Insights and Industry News for Media Professionals
Programmatic attribution modeling: Attributing TV Programmatic

Programmatic Attribution Modeling

August 24th, 2016   ||    by Charlene Weisler   ||    No Comments

In an environment of cross-platform media and the Internet of Things, marketers are grappling with programmatic attribution modeling. Attribution, according to Cynopsis, is defined as “the process of identifying a set of user actions that contribute to a desired outcome, and then giving each of those actions a specific value.” Ideally, good attribution would enable marketers to get the correct and fair sequencing of events that truly influenced a consumer’s behavior.

Last Touch Attribution

Most common is last touch attribution, which gives full credit to the last-touch point before the desired action takes place. This essentially negates the value of all previous actions leading to activation. The IAB is addressing programmatic attribution modeling, specifically on last-touch, with an update to their 2012 guidelines. According to Benjamin Dick, director, industry initiatives, IAB in a recent Mediapost article, the update was to, “account for the many developments in cross-screen measurement, data collection, and usage, and ultimately provide a common language when discussing attribution tools, technologies, and methodologies.” The intent was to show the limitations of, and alternatives to, last-touch methodologies.

Vishal Agarwal, a staff writer at TechCrunch offered a good overview of attribution, suggesting the following alternatives to last-touch:

  • Equal Weighting: Gives credit to all channels in the decision process
  • Time Decay: Provides the most weight to the most recent channel
  • Algorithmic: Uses millions of data points, which are then weighted by purchase path

Application to Programmatic

As programmatic become more prevalent, it may change the way we look at and process attribution. But just how fast and how extensive this change will be is open to opinion.

Andrew Feigenson, managing director of digital at Nielsen is optimistic and notes that, “Marketers still need to move product, which means they need to reach enough people and with the right mix of media. The next generation of attribution will account for tonnage and also for a huge diversity of options for ad formats, devices, and targeting across digital and traditional media.”

Attribution Will Take Time

“It will take a while for the TV programmatic players to get past content data, direct match targeting, granular post-evaluation to ROI by attribution, single source, mix modeling, and/or A/B,” says Bill Harvey, chairman, RMT. “When they do get to ROI they will tend to take something off the shelf, such as a well-known attribution platform, then realize that accuracy is all-important and will reopen consideration of single source, A/B testing, combination approaches. Ultimately, they will wisely seek a programmatic solution that covers all media, especially linear TV and all forms of digital.”


Caroline Horner, SVP at comScore, believes that how people process TV will determine how fast programmatic attribution advances. She says, “Except for a handful of small TV panels, the actual person in front of the TV is not explicitly known. It’s a challenge because TV is often the dominant influence in the sales funnel. If we accept the household-level data construct for the analysis, there’s ample data from which to model attribution. Attribution requires a view into all inventory that the person is exposed to. Most programmatic TV buying platforms represent only a portion of inventory so it’s challenging to believe that current programmatic TV offerings could have a comprehensive attribution model.”

Optimistic or not, industry acceptable attribution must occur in programmatic to advance the business. There are great minds focusing on this issue. That, along with advancing technology, may mean it’ll get there sooner rather than later.

Tags: , , ,

Share this page:

Share on Facebook Share on Twitter Share on LinkedIn Share via Email