Building A New Measurement Model…
Editor: On April 12, 2016, HRTS presented an in-depth look into the world of measurement and ratings. Panelists included Gian Fulgoni, Co-Founder and Chairman of comScore, Inc; Preston Beckman, media consultant and former network research head at Fox and NBC; Julie Piepenkotter, EVP of Research at FX Networks and Eric Solomon, SVP, Product Leadership at Nielsen. Moderated by Variety's Andrew Wallenstein, the spirited conversation touched on many aspects of the ratings and measurement challenges facing our industry. To add more depth and context to the topic, HRTS is pleased to provide the following white paper as well as a link below to view the entire conversation from April 14th. Enjoy!
Building a New Measurement Model for the Dynamic, Cross-Platform World
By Josh Chasin, Chief Research Officer, comScore
In March of 1930, Archibald Crossley went into the field with a telephone survey of radio listening behavior in 50 US cities. Thus began the currency audience measurement model that has held sway for over 80 years—first via surveys, then via panels, but always based solely on an attempt to draw a sampling of the user population based on small samples of people. This was Crossley’s legacy.
But if you work at a company of a decent size that is more than ten years old, you are painfully aware of the burden of the dreaded “legacy system.” In today’s technologically-driven world, legacy systems are probably the greatest impediment to any company’s future success.
Things are no different in the audience measurement space. It turns out that the inertia required to maintain the traditional, linear TV marketplace measurement model—pushed gently forward over 86 years, and in place largely as-is since 30 years ago — is anathema to the invention and development of the appropriate currency measurement system for the fragmented, multi-screen, highly targeted TV marketplace of the 21st century.
comScore and Rentrak are two companies who share a common vision of a new model of media measurement and who came together on Feb. 1 to bring that vision to fruition. We know that state-of-the-art 21st century audience measurement systems must be built around best-in-class Big Data and census-style assets, informed but not dominated by panels — and not the other way around. That’s a recipe for failure to meet today’s fragmented media measurement needs.
Rentrak pioneered the deployment of return path viewing data from TV set top boxes in the provisioning of syndicated TV audience measurement—introducing metrics of unprecedented stability, precision and granularity. comScore integrates site-centric and campaign-centric digital census data with best-in-class massive digital panel data, to provide accurate metrics for tens of thousands of sites, apps, streams, and campaigns each month. Now as one company, the new comScore is putting together all of these robust data assets necessary to bring audience measurement into the 21st century.
So what are the greatest audience measurement challenges today, and how can we meet them? We tend to zero in on these three:
- Sophisticated targeting and innumerable platform options render traditional panel sample sizes woefully inadequate.
Legacy audience measurement systems were designed in an age when advertisers wanted to reach men or women 18-49, and they had exactly three networks and a couple of dozen magazines at their disposal with which to do so. Today, advertisers seek to activate precise behaviorally-defined targets, in as close to real time as possible, across a hyper-fragmented array of platforms, screens, networks, and devices. Legacy, small sample-based solutions are simply insufficient for meeting the needs of today’s complex advertising ecosystem. Only systems based on Big Data assets can enable buyer and seller alike to understand how complex targets accrue to media content across platforms—and only these systems can provide cross-platform activation.
- It is impossible to sample your way to cross-platform unduplicated reach.
The cornerstone of cross-platform measurement is understanding unduplicated reach. This is the Holy Grail for media buyers. Most other cross-platform measures are simply additive. A thousand minutes on TV sets and a thousand minutes on tablets sums to 2,000 total minutes of viewing; but a thousand viewers of originally-scripted content on TV and a thousand viewers on tablets could be anywhere from 1,000 to 2,000 unduplicated viewers in total, depending on how many viewers watched both on a tablet and on TV. The problem is that traditional panel-dominated systems run into the zero audience problem quickly because they have no one in their sample that views a particular channel —meaning that the media buyer can’t attempt to understand how many of those cable network viewers might also have streamed that network content on their phone, or their tablet, or via OTT. comScore believes in assembling best-in-class measurement assets, and also best-in-class duplication measurement assets, so that granular overlap across platforms at the campaign level may be empirically quantified. It is functionally impossible to understand cross-platform unduplicated reach—and thus, to understand today’s cross-platform media marketplace—using legacy sampling techniques.
- Big Data assets and the war on bias and error.
Traditional media measurement panels—regardless of their quality—are subject to myriad sources of error and bias, especially as they are stretched beyond their breaking point to report on more granular consumer segments than ever before. In this age of RFID, geolocation, artificial intelligence, and algorithms, it is sometimes hard to believe that legacy measurement systems still rely on people pushing buttons every time they watch TV, in order to ascertain who’s in front of the set. And in many local markets, legacy systems still rely on viewers filling out paper diaries. Fragmented measurement means that any error or omission the respondent makes is magnified. The Coalition for Innovative Media Measurement (CIMM) published Eight Criteria for Solving Cross-Platform Measurement of Exposure to Ads and Content; in this document CIMM notes that “…the least amount of intrusiveness required by respondents is optimal for the highest quality ‘behavioral’ measures of media exposure.” comScore and Rentrak have embraced this philosophy for years; our panel software enables us to passively assign behavior to an individual consumer. Census tagging of millions of digital properties and viewing data from 40 million TV sets require no action whatsoever from the viewer— they are massive datasets, collected entirely passively.
None of this is to say that measuring today’s multi-screen consumer is an easy task. Quite the contrary; audience measurement is more challenging than ever before. Fortunately, technology has provided us with new tools and assets to deploy in the service of measurement and reporting on multi-screen, multi-platform exposure and engagement. At comScore we’ve focused on building new models of currency solutions unencumbered by the past to help both buyers and sellers navigate today’s fragmented advertising ecosystem—bringing the needed measurement assets to bear in order to best facilitate understanding, targeting, and activating today’s cross platform consumer.
Video of full panel discussion from April 14th.
Tagged as: advertising, Beckman, comScore, cross-platform, digital, fulgoni, measurement, media, neilsen, OTT, Solomon, television ratings, TV
Comments are closed.