Commentary

How Much Should Live Viewing Really Matter To Advertisers?

by Steve Sternberg , Featured Columnist, June 5, 2017

Not long ago, we were in a television world where the most loyal program viewers were also the most attentive commercial viewers.

Today, the most loyal program viewers, those who watch their favorite programs on DVRs, are the most likely to avoid commercials. These loyal folks only represent about half of TV homes in the U.S. but the vast majority of their viewing is time-shifted — which presents a dilemma for advertisers.

Season-to-date, 40% of viewing of prime-time original cable scripted series, and 20% of viewing to original broadcast scripted series, were time-shifted via DVRs.

Since virtually every study ever done on the subject shows about 75% of commercials on DVRs are fast-forwarded, this means, for original scripted series in prime time, roughly 30% of cable commercials and 15% of broadcast commercials are skipped.

The glass half-full way of looking at this, of course, is that 70% of cable commercials and 85% of broadcast commercials in original scripted prime time series are not skipped.

If we look at all prime-time series, we see a slightly different picture. Because original scripted series are such a small part of the cable landscape, an average of all cable series shows only 10% viewing through DVRs. Broadcast, with the bulk of its programming being original scripted series, maintains its 20% delayed viewing. You can add 5% to each for traditional channel switching.

That still sounds like a lot of commercial avoidance. But does anyone really believe that digital or online ads are viewed at anywhere near this rate? Part of the problem is that there’s so much good linear television data available that you can make numerous positive or negative points depending on what you are trying to sell and who you are trying to sell against. There’s little good data on digital commercial viewing.

Does it matter that Nick-at-Nite and ION each have 96% of their adult 18-49 and 25-54 audiences live, if they only average three-tenths of a rating point and are just reaching the same small group of viewers over and over again?

Does it matter that TBS and USA have a smaller percentage of live viewing (93% and 87%, respectively) if they have 40% more live + 7 viewers? Does it matter that CBS has only 80% live viewing among these demos, if it has five to six times as many viewers?

In a sense these stats matter, simply because it’s important to know how many people are potentially exposed to your commercials.

But from a marketplace perspective, they are not as significant as they might seem. And this has nothing to do with C3, which does not account for fast-forwarding through commercials. The key to buying and selling is not rating size, it’s how networks rank compared to their competition. Pricing will always adjust to new benchmarks, but its how you perform relative to your competition that really counts.

If you look at the broadcast networks among adults 18-49 and 25-54, for example, their ranking does not change going from live to live + 7 ratings. If you look at cable networks, while the actual ranking changes slightly, the top 15 networks are exactly the same among live and live + 7 viewers. On a program basis, 28 of the top 30 original scripted series on broadcast and cable combined, were the same among adults 18-49 live and live + 7. Among acquired, off-network series, 30 of the top 30 series were the same.

Research precision is not as important as directional accuracy. In other words, knowing that program A has 10% more commercial viewers than program B is more significant to buying and selling (i.e., pricing) than knowing the precise number of people watching program A.

Until the rankings dramatically change for different streams of ratings data, this is interesting from a research standpoint, but shouldn’t really be a major concern when it comes to the buying and selling of commercial time.

advertising, ratings, tv

13 comments about "How Much Should Live Viewing Really Matter To Advertisers?".

Check to receive email when comments are posted.

Ed Papazian from Media Dynamics, June 6, 2017 at 8:07 a.m.
Steve, the usual rebuttal to your position is that Nielsen provides average commercial minute audiences as the sellers' audience currency and deletes any set where the ads were skipped. So what's the problem? As we both know, Nielsen's commercial minute ratings are not really measures of viweing, but merely that the set was on when the commercials appeared. However, to make a stronger case for live viewing, don't you need some indication that live program viewers are more likely to be attentive to commercials---or recall them---then DVR delayed viewers who, theoretically, have the same opportunity to watch them. Indeed, as the demos of the latter viewers tends to be younger and better educated than their live counterparts, this might be part of the equation. Perhaps younger delayed viewers, who are less likely to be attentive to commercials generally, have less value to advertisers when reached on a delayed basis, than when reached live.
Reply

John Harpur from Yellow Submarine, June 6, 2017 at 11:58 a.m.

So which is it? C3 does or does not account for fast-forwarding?

Ed Papazian from Media Dynamics Inc, June 6, 2017 at 12:17 p.m.

John, unless something has suddenly changed---and I doubt it----C3 ratings are for "commercial minute viewers" and eliminate any dial switching avoidance by "live" viewers as well as any fast forwarding by delayed "viewers". An interesting question---which I don't have an answer for ---is how Nielsen handles sound muting during commercials---assuming that this aspect is measured.

Steve Sternberg from The Sternberg Report replied, June 6, 2017 at 12:33 p.m.

Hi John-

C3 does not account for fast forwarding. I've written a number of articles on the subject. Email me and I'll send you a copy of The Sternberg Report - Premium Edition, which explains why C3 does not measure fast-forwarding. It ordinarily costs $2,000 per year to subscribe, but I can cut out the appropriate section for you. I think I also wrote a TV Everywhere column for MediaPost on the subject, you can look there.

Ed Papazian from Media Dynamics Inc, June 6, 2017 at 1:41 p.m.

Steve, maybe we are getting caught up in semantics. Are you saying that a Nielsen C3 commercial minute rating, which includes delayed viewing, does not eliminate those DVR viewers who fast forwarded past---in effect, skipped-----the commercials? Or are debating something more technical here?

Ed Papazian from Media Dynamics Inc, June 6, 2017 at 1:45 p.m.

Add a "we" to the question at the end of my last post. Sigh!

Steve Sternberg from The Sternberg Report replied, June 6, 2017 at 2:11 p.m.

C3 does not eliminate viewers who fast-forward through commercials. It has to do with how Nielsen measures the average minute. Thay use the plurality of seconds viewed during the minute, and apply that as viewing the whole minute for that channel. During DVR playback, there is only one channel being viewed, so any viewing, even just a few seconds, becomes the plurality. So you could see the first 10 seconds of a commercial minute, fast-forward thrugh the rest, and be counted as viewing the entire minute.

Ed Papazian from Media Dynamics Inc, June 6, 2017 at 3:25 p.m.

Steve, I understand the technical issues Nielsen faces in tabulating its average commercial minute ratings, however, I was referring to the general understanding of the ratings by buyers and sellers, which is that DVR delayed "audiences" who "zap" commercials are not included in the so-called average commercial minute ratings. It is my impression that while Nielsen has made a compromise in order to tally its findings, that this does not make that much of a difference. If you have evidence that, say, 25-35% of those who are not reported as average commercial minute delayed viewers actually see the ads with some degree of impact, than this point might be worth raising with Nielsen. In like manner, if Nielsen is, somehow, counting viewers who did not have a chance to see the ads as part of their "audience"A, that, too, would be worth noting. In my opinion, if a perfect measurement and tabulating system was available, I doubt that we would see a meaningful difference----but I could be wrong.

John Grono from GAP Research, June 6, 2017 at 9:45 p.m.

I'm unsure as to what meters Nielsen uses in the US, but if it is 100% audio-matching based (as we have in Australia) then as FFWD has no sound and a muted TV has no sound the TV, the tuning can not be matched to the audio reference source and therefore no rating is counted. This applies to both content and advertsing minutes.

Steve Sternberg from The Sternberg Report replied, June 6, 2017 at 10:40 p.m.

That wouldn't matter. If only 5 or 10 seconds of a commercial minute is viewed, the entire minute is counted. The entire commercial minute would need to be fast forwarded for it not to be included.

John Grono from GAP Research, June 6, 2017 at 11:18 p.m.

Steve, it depends on the validation rules. In any minute, a large number of channels can be tuned in to as a person flicks around looking for something to watch. There are generally two rules. The first is the 'rule of dominance' - the channel with the most seconds in a minute gets credit for the entire minute. The second is the 'middle minute' rule. The channel tuned at the 30th second gets the credit for the whole minute. Middle minute was introduced when channel numbers started to proliferate. From memory the processing rule also requires 30+ seconds of a minute to be tuned in order to credit the whole minute.

But the point behind audio matching is that if the TV sound is muted, or the DVR is in FFWD mode durigng playback (i.e. no sound) then there is no matching and credit for that duration.

Steve Sternberg from The Sternberg Report replied, June 6, 2017 at 11:32 p.m.

John, that's not how it works.

Ed Papazian from Media Dynamics Inc, June 7, 2017 at 7:31 a.m.

As I pointed out in an earlier post, how big of a difference are we talking about in this discussion. If, on average, it is very small and the variations are more or less random, then it doesn't matter that much---does it?