Commentary
Doesn't Matter What I Watched, If I Didn't See Your Ad
- by George Simpson , Featured Contributor, March 30, 2017
If you want to find out my television usage, you are better off asking my wife, who is quick to point out," You've been watching those goddamned basketball games for three straight days!" Trying to explain that 7 p.m. to 1 a.m. on consecutive Thursday and Friday nights is not really "all day" is pretty much throwing water on sand. Although it is hard to argue with someone who says, "Why is South Carolina still playing — is that the NIT?" (Just kidding, she wouldn't know the NCAAs from the NIT if they both bit her on the leg).
Since almost nobody has just one medium on at a time anymore, it’s really hard to calculate your time spent in front of this screen or that one. Thanks to the crushing commercial load of network prime-time shows, you need to have a laptop or your phone open to social media so that you don't put a gun to your head before the show resumes. I can comfortably get through the text-heavy Weekend Review section of the New York Times or Wall Street Journal during just a couple of TV shows. If I were asked if I'm watching TV or reading the paper, how do I respond — especially since I was also texting one of my kids?
advertisement
advertisement
In fact, if you add up my media time, it will tend to approach about 40 hours a day. And that doesn't include media that have no screens, like the car radio on the way to the gym and grocery story. Or the billboards I pass. I spend an hour and a half reading newspapers, but only one is dead-tree, the other is on a tablet. I read lots more news online or on my phone throughout the day, often breaking news on the Times or WSJ Web sites. So what do I put in the diary under newspapers?
How much time anyone spends with media is not just a parlor game, but it is supposed to reassure advertisers that their commercials are being watched. Really? Between the remote control and my cable-provided DVR, I hardly ever see commercials any more — even during must-watch live events (go, Tar Heels!!) because as soon as they go to commercial, I go to reruns of “The Big Bang Theory” or, if I am really hard-up, a soccer game. Soccer never has commercials and oddly seems to be on 24 hours a day. If I am really lucky, the fans are rioting in the stands.
There is someone in my house (who will go nameless) who routinely falls asleep in front of the TV, often until the wee hours of the morning. If there was a People Meter in our cable box, it would virtually vibrate with glee and drop her into the "heavy TV watcher" audience box, although I am not sure who buys "people who watch seven hours of ‘Law and Order’ reruns four nights a week."
Regardless of your predictive algorithms, recall diaries, minute-by-minute tune-in data, etc. I think it's impossible to keep up with how quickly and effortlessly audiences shift from one medium to another — and the steps they take along the way to avoid seeing your ads.
Correct, Goerge, and forget about cross screen measurements, we don't even have adequate TV audience measurements that tell us who and how many actually watch an advertiser's commercials---assuming that we can define what the term "watch" means. I maintain that we are asking the audience measurement companies for much too much "granular data" for them to supply with a realistic degree of precision and accurracy. Still, if we cant make any decisions without "data" to fall back on---which seems to be the case---I guess that flawed or misleading data is better than no data-----or is it?
Just a minor clarification George.
Every TV audience measurement system I know of, have a 'validation' process on the data collected. One thing they validate for is 'uncovered' tuning and viewing, and another is 'peristent' tuning and viewing.
'Uncovered' basically means that the TV set is on and tuned in, but that nobody has logged in and registered as a viewer. Any of that tuning data is removed and there is no crediting to the ratings.
'Persistent' basically means inordinately long periods of time that a TV has been tuned (or a person has viewed) to just one channel or programme with no change on the TV (e.g. no channel change even though the programme has ended and a new one started) or no registration or usage of the PeopleMeter remote. Most common cause is someone fell asleep. All ratings back to the last 'change event' are removed, and if the absolute volume of affected time is large then that household/person is removed completely from the ratings for that day.
As most audience researchers would attest to - better to not ask people they thought they did, but better to observe them while they do it.
Thank you John, enlightening. G
Thought provoking piece George! In the past, we needed to use TV program viewership ratings as a surrogate for audience. We still don't have pure ad unit ratings for TV. Thanks to VOD, mobile and online video we now can get close to understanding whether a viewer is watching or not based on whether the ad was on screen while the viewer is interacting with that screen. Linear TV is still far behind mobile and online video and the future I beleive is in VOD or PPV for tradtional TV providers. As interactivity becomes more a part of T/V (TV/Video) advertising, it will provide a better way of asking - "are you still there?"
John, theoretically you are correct,
But in practice it rarely works that way, especially for mobile and online ads and video ads. The 'digital' metrics revolve around serving ads and not seeing ads.
Take for example, one of Australia's largest news sites. It uses video autoplay embedded in articles (at this stage I am talking about content video and not ad video). The user starts reading the article and the video is placed above the fold (thus dominating the usable screen). The video takes a short while to finish buffering, load then play. The user has by then scrolled down to continue reading the article which pushes the video player off the viewable screen. The only real way to know that the video has started to auto-play is when you hear the sound. In the curreny system that counts as a full viewer and a full play.
When you move from the app based mobile/tablet world back into the desktop/laptop world (now the minority of usage) you also move into the world of multiple browsers and multiple tabs on those multiple browsers. The issue of 'focus' then comes into play. You can only 'credit' usage and time to the content that is on the active tab of the active browser. However back on the server-side, each and every piece of content is credited by the server-side analytic software is fully credited. While some of them make take into account the 'viewability' of the content (how they account for latency issues at the user's side is a mystery), none of them - just like with television - take into account the proportion of people who actually actively view the content.
In essence ALL media metrics are surrogates for audience. Some are closer to reality (ethnographically) than others. The key issue is that more data and points of measurement generally does not mean they are better surrogates, and in the main they aren't.