"On TV and Video" is a column exploring opportunities and challenges in programmatic TV and video.
Today’s column is written by Jason Miller, director of product management, brand measurement at Google.
I'm often deeply impressed by the creativity of TV ads. I loved the sentimentality of P&G's "Thank You, Mom" ads, the humor of Dodge Durango's Ron Burgundy ads and the absurdity of Old Spice's "Dadsong" ad.
Ads like these win industry awards, attract positive press coverage and generate excited social media chatter. While some ads are immediately successful, it is not known how effective most campaigns are until weeks – or even months – after they run.
This is a measurement challenge that will be solved through more census-based measurement.
Too much of the approximate $70 billion in TV ad spend, in the US alone, is based on annual media mix models. Clearly, advertising produces great results for some brands. P&G estimated that its "Thank You, Mom" campaign for the 2012 summer Olympics resulted in a $500 million sales lift. And when Ron Burgundy played a starring role in Dodge Durango ads, the SUV brand saw a 59% increase in year-over-year sales.
But faster measurement insights could be game-changing. What if a $500 million sales lift could be optimized to a $600 million sales lift, or if a 59% increase could be optimized to a 75% increase? Or, what if advertisers had access to insights that would help them intervene and turn around a poorly performing campaign?
An accountable future for TV and video will leverage the best of traditional TV and digital video measurement. GRP-based measurement of traditional TV should be extended to digital video in order to provide a complete view of reach and frequency. Census-based measurement of digital video should be extended to traditional TV in order to measure impression-level detail and interaction while campaigns are running, instead of days, weeks or months after.
The impetus to leverage a combination of panel-based and census-based measurement will grow as the technical limitations to doing both fade away. Some forms of traditional TV delivery can already measure every impression and provide that census-based measurement. The final missing piece is 100% coverage for census-based measurement across traditional TV and digital video.
Every year, the industry gets closer to 100% coverage. This empowers advertisers to measure what viewers see, think and do in response to advertising during a campaign, not after it, and will improve measurement in the following ways in the near future.
Greater Accuracy About What Viewers See
Advertisers will get a more accurate picture of what viewers see through greater access to viewability measures, an evolving GRP metric and by filtering census-based metrics according to audience segment.
With viewability measures, advertisers can favor the digital video impressions that approach the high bar set in TV where all ads have a chance to be seen on the screen. There's a long way to go here since 46% of digital video ads are not viewable.
Finally, viewability, GRP and census-based metrics like video plays, quartiles and completions will all be studied by audience composition so that advertisers can get a more accurate picture of what certain audience segments see. Advertisers will be able to understand, for example, if some audience segments see ads through to completion more than others.
Deeper Insights Into What Viewers Think
More census-based measurement will also provide deeper insights into what viewers think after being exposed to ads. For decades, the traditional TV industry has leveraged brand surveys to understand things like whether or not an ad was memorable.
With census-based measurement, brand surveys can be tied to viewable impressions so that only people who have viewed an ad can respond to a brand survey. Then, rich audience data about exposed audiences can be repurposed to describe survey respondents and allow aggregate survey data to be sliced and diced for deeper analysis. For example, an auto advertiser could slice survey data to study the difference between auto intenders and the general population.
Ideally, advertisers will be able to seamlessly attach brand surveys to every TV and video campaign and then see the results during the campaign, not after. Innovation in this area could go even further to allow advertisers to buy media based on brand lift points.
More Information About What Viewers Do
Finally, more census-based measurement will provide more information about what viewers do after being exposed to ads. With specific insight into who saw an ad, advertisers will be able to analyze activity across search, web and mobile applications after ads run, down to the minute, to interpret and determine advertising's impact on online activity and engagement. Advertisers will see which networks, programs, ad lengths and ad creatives are the most effective at motivating viewers to engage online.
Furthermore, TV will gain from the census-level insights prevalent in digital video that make it easier to go beyond measuring online activities by measuring offline purchases, sales lift, cross-device conversions and customer lifetime value.
The direction that TV measurement is headed will make these gut decisions easier because advertisers will have faster access to more data points. As a result, advertisers will make better creative decisions, ads and media investments. They'll feel empowered to alter course mid-campaign. Ultimately, they'll produce better financial results while delivering a better viewer experience.