Digital Revolution! Stay Tuned for Details

You'll see up there on the top of the page that this column-blog is called the Online Metrics Insider.  But really now, isn't "online" as a concept so Five Minutes Ago?  I mean, are you ever OFF line?  I know I'm not; right this very minute, as I work on this column, I'm downloading music, streaming an episode of Lost, and tweeting. At comScore we tend to think in terms of digital media as opposed to "online" media.

I mention all this because this column is about the digital medium that, in today's U.S. media marketplace, commands the most adspend; of course I'm talking about TV.  While it may be tempting to place TV in the analog, or "old media" box, thanks to satellite TV and digital cable, maybe half of all TV impressions in the U.S. are delivered digitally.  While newspapers that still print are facing double digit circulation declines, the medium of television has undergone a subtle but profound change over the past five to 10 years.  The distribution of TV has changed technologies.  What makes that fact profound is that while the printing press is a quintessentially industrial-age technology (and so, one could argue, is broadcast), chances are the television content that comes into your house gets there through information-age-ready, digital tech.



Not surprisingly, all the excitement in TV metrics is on the digital side right now.  Canoe's David Verklin has talked about the convergence of television with digital interactivity to create "t-commerce."  TRA, the latest venture of my old pal Bill Harvey, just raised a second round of funding from a group of investors including Arbitron; TRA combines set-top-box data with shopping and consumption data to get at causality and ROI.  TiVo has turned the data it has on viewing behavior -- and I mean behavior; the pausing and rewinding and fast-forwarding that we all do now when we watch TV -- into a sandbox of analytic possibilities.

Back in 1997, I went to an early ARF conference about the Internet, in Monterey (I had a really nice dinner off campus with agency vets Alan White and Fred Sattler.) At that conference, we talked about how the Internet had to "grow up" and make itself look more like TV before anyone could really take it seriously.  But a funny thing happened on the way to Internet maturity: 12 years on, we've all discovered that it was the other way around all along, that all media will be digital media eventually, and that one day TV was going to end up looking like the Internet -- the addressable, interactive, on-demand Internet.

Certainly, TV metrics are evolving in that direction. Set-top-box (STB) data becomes the site-centric server data of television, with strengths and weaknesses that mirror site-centric server data.  Like server data, STB data is machine-level, not person-level, so it has the same "Last Two Feet" challenges arising from not knowing who the people are doing the watching (these challenges including demographics, disentangling reach versus frequency, and some dirt around session end times.). And like server data, STB data provides the benefits associated with census-level detail on the granular (second-by-second) flow of data between a set of machines, with that data flowing in two directions.

In the Internet space, we're moving toward a panel-centric measurement hybrid construct, where machine-based server data and person-based panel data are integrated to provide the best of both worlds in a single measurement system; I expect that you'll see a similar measurement construct emerge in the TV space.  And, too, I expect that TV is about to confront what my colleague Manish Bhatia over at the Evil Empire calls the "problem of plenty": digital media lend themselves to multiple measurement systems, and history tells us that media buyers and sellers experience comfort when there is a single set of data, even when the preferred sport of both sides is criticizing that data.

Another exciting possibility that emerges from digital TV metrics is one of true cross-media measurement.  Media companies -- in this column, we generally call them "publishers" -- want to understand how the Internet is affecting their traditional platforms.  The MPA has done some good research showing how magazines work well supplementing a TV buy.  All this digital data gives us some great opportunities to develop new, empirical tools that will allow media companies, advertisers and agencies to understand how the Internet (and online video, and mobile/cellular) can extend a media brand's reach, how the different digital platforms work together to surround the consumer and deliver holistic marketing messages: right message, right platform, right time.  True media-mix ROI will be within our grasp; we will be able to understand how shifting incremental money from traditional TV to online and mobile video, for example, affects branding metrics and sales.

All this and more was swimming through my head as I rubbed shoulders at 770 Broadway week before last with TV's digerati at Mitch Oscar's STB throw-down.  There is a lot going on in this space right now, a lot that we'll all want to learn very quickly.  It was heartening to see so many of the right people in one room; both long-time colleagues and new ones, talking and thinking about the coming revolution in TV metrics.

Sure, sometimes it feels like we're all a bunch of expectant dads in a ‘50s sitcom, pacing the delivery room, chain-smoking, and wondering, "When's the baby gonna get here already?" Well, I'm no doctor, but I am a digital metrics insider.  Make no mistake; there will be labor pains, and more labor pains.  But I'm here to tell you, that baby is going to show up sooner than we think.  And when the baby shows up, everything changes, forever.

2 comments about "Digital Revolution! Stay Tuned for Details ".
Check to receive email when comments are posted.
  1. John Grono from GAP Research, May 12, 2009 at 6:54 p.m.

    I couldn't agree more Josh.

    Here in Australia we're forging ahead investigating a panel/server hybrid to produce online audience metrics - there sure is a lot to learn. Later this year we should also see the first results for a hybrid measurement system for subscription TV.

    The most important learning though is that while panels are stressed as each medium fragments, they still provide an extremely robust measurement of people and their behaviour, even if it is a gradually deteriorating measurement of the quantum. Criticising panels or criticising server-centric is self-defeating - the key is to leverage BOTH data sources.

    Interestingly, we are alos launching a hybrid measurement system for Out-Of-Home (OOH) media later in 2009. When we set out on this task we came to the realisation that it was impossible to design a sample or panel that would measure all OOH - one that would capture: drivers and passengers; pedestrians; commuters; shoppers; flyers and so on. What we did realise is that we had an abundance of other data; traffic counts; airline ticket sales; bus and train ticket sales; footfall counts at shopping malls and so on. We were then able to use other data sources such as government "Household Travel Surveys" as well as bespoke surveys to allow us to calibrate the level of duplication between each OOH format as well as over time. After some sophisticated mathematics, as well as adjusting each potential contact into a "likelihood to see" we have robust estimates of audience for every panel and every OOH campaign. Of course these data are based on averages and likelihoods, but while the medium is sold on a typical four-week campaign basis they are at the precision level that the buyer and marketer want. Of course, with the advent of digital signage there will be some impact, but as long as the content is sold on a 'rotational' basis and not a 'spot' basis the data will hold up.

    So, what's next? I envisage that there will be a "hybrid of the hybrids". That is, a system that takes each medium's audience measurement and melds them together (retaining each one's 'currency') to produce de-duplicated audience data across all media - for planning purposes. At it's hub it will be very product and brand focussed - and funded by the marketers? - but use the common harmonised demographics across all media. I can't see it having the necessary precision and granularity to measure ROI. My belief is that will be best done by non-linear multivariate time-series econometric modelling by the marketer (or their agent) - but this an altogether different topic.

    John Grono, GAP Research, Sydney Australia.

  2. Kevin Mirek, May 12, 2009 at 9:29 p.m.

    Personally, I think we're looking at more than ten years of continuing labor pains. The profound inability of any TV station to monitize its "other than broadcast" efforts (internet, mobile, etc.), the total lack of high quality video content from anywhere other than TV, the pending litigation issues of piracy and intellectual & real property, and waiting for the 12-24s to have any spendable income to make any difference will all contribute to those labor pains. A totally "free and on-demand" world will need several epidurals and an episiotomy.

Next story loading loading..