What is the 90/10 Measurement Trap, and How Can Advertisers Avoid It?

chair in front of hundreds of TVs

Eliminating Tradeoffs Between Counting and Outcomes

By Chris Kelly, CEO, Upwave

Ever since the Media Rating Council board halted its support of Nielsen’s national rating service, TV media execs have moved from discussing the technical limitations of Nielsen’s panel to the deeper limitations of the current Nielsen-centric framework for measuring TV.

The VAB’s Measurement Innovation Task Force has taken a strong leadership role in rethinking measurement for the CTV era. As part of this initiative, NBCU EVP of measurement Kelly Abcarian recently announced a renewed measurement framework  that prioritizes brand building–a position we support.

“It’s time for us to declare measurement independence and build solutions that will serve all consumers, advertisers, publishers and platforms for the next century,” Abcarian wrote in the blog post about architecting a new blueprint.

As Abcarian pointed out in this interview with Beet.TV, TV has long suffered from a misalignment of analytics priorities. Whereas digital media measurement spending has historically been 90% on outcomes and 10% on counting, TV measurement has focused on the inverse: 90% on counting (that was never even that accurate) and 10% on outcomes.

The path that led to this 90/10 paradigm, both on the digital side and the TV side, happened naturally. For decades, counting on TV was extremely difficult and therefore warranted an abundance of energy and budget, to do. Advertisers believed top-down that TV worked, so they were comfortable simply assuming it drove outcomes.

Conversely, on digital, counting was always simple; all you needed to do was look at the ad server logs. However, advertisers didn’t assume that digital could be as effective as TV, so they focused energy and budget towards measuring outcomes.

The rise of Connected TV (CTV) has encouraged many to think of TV more like its digital counterparts with an obsessive mind towards addressability. The stakes are high as TV fears it will continue to bleed dollars to digital, which has been the home of performance marketers.

The explosion of CTV means that just because you can count the ad server logs doesn’t mean you know how many people saw it, like when a family all watches a program together in the living room.

With the fragmentation of video consumption, the assertion that the misalignment—and the assumption that TV “just works” has to change, but it’s not a question of percentage tradeoffs. Counting and outcomes are not mutually exclusive and shouldn’t be considered zero-sum. Therefore, we need both counting and outcomes. TV measurement dollars should — and can — be 100% spent on both.

But it’s not just about combining counting and analysis; we need to make sure we’re measuring the right thing.

To accurately measure the performance of an advertising campaign, you already need to identify individuals who are exposed to the campaign, as well as those who are unexposed.

Traditionally this exposure data has been used by measurement providers to produce brand lift studies–measuring the outcomes of ads by comparing exposed individuals with similar, unexposed individuals. This is problematic though, because it measures brand lift without audience estimation. With the right technical expertise, that same measurement provider can use the exposure tracking data to count the size of the population of exposed individuals as well.

Combining this data lets you compare campaigns based on their true outcome, producing much more meaningful insights. Previously, if you had two campaigns with the same cost: one that lifted 2% of exposed individuals and another that lifted 3%, you would assume the second campaign was more successful.

However, if you add in campaign data such as on-target reach, you may find the first campaign reached 20 million people in the target audience, while the second only reached 10 million, changing your conclusion about campaign effectiveness.

Due to pressure on CMOs and agencies to show financial ROI on advertising, the industry is becoming too focused on imprecise and backward-looking measurements like Multi-Touch Attribution (MTA), which measures the short-term sales correlations of advertising.

Any time you employ backward-looking measurement, you will undervalue media and content that supports long-term brand building, that is, TV content.

TV is and has always been the domain of total immersion, where brand advertising builds sales and your brand over a long period. Focusing on immediate sales misses the point.

Of course, Nielsen’s recent stumbles have allowed many new measurement providers to recast how it should work and broadcasters and content providers, long aware of the deficiencies of historic TV measurement, are more than happy to get fresh voices into the conversation. While the VAB’s Measurement Innovation Task Force created a unique measurement framework, in partnership with NBCU, it welcomes others to weigh in on what will work best.

It’s not surprising that an executive from a content provider is the face of this call to action. With the continual decreases in C3 and C7 ratings, broadcasters are rightfully looking for improved ways to measure their advertising. Otherwise, the quality of content will deteriorate rapidly.

While it’s true that people increasingly watch news, shows and sports with a second device in their hands, it remains a lean-back experience. Focusing just on cross-platform attribution will fail to account for the brand-building TV advertising for which it has always been known.

But it’s not about sacrificing counting for outcomes or vice versa. The future of TV measurement is one platform that can do both. What is required is moving away from single solution providers towards campaign analytics tools that track results and continuously optimize TV, digital and cross-media campaigns towards both full-funnel lift and targeting accuracy.

No medium gets the same attention combined with reach as TV in all its forms. But if TV providers want to maintain the value of their content, they must fight for the right measurement that does an adequate job of tracking how many people see the ads on their content and how those ads impact their audiences.

Got a Question? We’ve Got Answers.