Who Wins And Loses When 9 Trillion Linear TV Ads Disappear?

  • by , Featured Contributor, August 9, 2018
There’s no question that on-demand video-streaming services are radically reshaping video viewing. According to Nielsen data, streaming video-on-demand services are now used by two-thirds of U.S. households, up 10% year over year.

Meanwhile, linear TV viewing is declining at a significant rate. Also according to Nielsen data, viewing of ad-supported cable TV and English language broadcast TV is down 9% year over year.

Some might see this as a simple transference of viewing from one type of service -- pay TV from cable and satellite services -- to a new kind of service: pay TV from Netflix, Amazon Prime and Hulu.

But it’s not that simple. There's at least one enormous difference: While the vast majority of the streaming video viewing today is ad-free, all of the linear TV viewing that went away was loaded with ads.

To put the differential “ad load” into perspective, national cable and broadcast TV in total last year delivered 19 trillion ads to persons two and older (P2+ in TV media jargon), according to Kantar data. A nine percent loss in viewership means that almost 1.9 trillion 15- and 30-second video ads went away in a year. They just disappeared.



Any way you look at it, that’s an enormous amount of lost advertising opportunity. If you assume that future linear TV decline occurs at a similar rate -- a reasonable assumption, I think -- you would see the evaporation of one-half of all of TV’s ad impressions in five years: 9 trillion! That kind of loss will have enormous repercussions across the advertising and media industry.

Who will win?

For sure, Netflix and Hulu will win. They will keep winning as long as they are able to keep growing and not lose too much share to new entrants.

Amazon will win for sure, since more streamers will mean more and more consumer lock in to their free delivery service. Plus, Amazon owns the fast-growing Twitch streaming service, which does contain ads, and can serve as a complimentary ad-supported video platform.

Video content creators will win, as long as the market is expanding, since each of the services will be in an “arms race” to outdo each other, at least for the next five years, when streaming services’ growth will slow and those companies will have to normalize their business models for profitability.

Who will lose?

TV networks will be in a tough spot. By losing both viewers and ad avails, they lose revenue on both sides of their business: carriage fees and ad revenue. While most TV companies are changing operating and cost structures quickly, they are probably not changing them quickly or dramatically enough to make it to the other side profitably enough.

They certainly won’t make it unless they massively change how they price, package and sell ads, since they will need to make three to five times more revenue per impression in the future to support the kind of profitability that their investors have enjoyed.

Viewers are likely to lose. Yes, advertising can be an annoyance, but it pays for a lot of free content that may not be available in five years. Think news programming, for example. Advertising helps inform people of products or service they might like but don’t know about. Fewer ads mean fewer chances for that serendipity.

Also, if retailers become the biggest media owners -- think Amazon -- they might not have the incentives to want their audiences to know about products other than those they carry.

What do you think? Who wins and who loses when 9 trillion TV ad impressions a year disappear?

26 comments about "Who Wins And Loses When 9 Trillion Linear TV Ads Disappear?".
Check to receive email when comments are posted.
  1. Bob Coyne from Tribune Media, August 9, 2018 at 4:34 p.m.

    The ads will transfer to Advanced TV, OTT, FEP and CTV.  My opinion, in greater numbers than the lost ads on linear.  I am selling them already! By the thousand impression lot.  Targeted by Geographic market.  Targeted by channel type, like Comedy, news, or drama, and targeted by demo.  And, each ad I sell has a very high percentage view rate that I can track.  Its more intimate viewing on your iPad or mobile phone, than 12 feet away on the old wall or dresser box.  And, the holy grail, the viewer can click on the Advanced TV ad, go to the client website, and by the products right then and there.  I can get credit for the sale. Its better than linear, or broadcast advertising. Sincerely, BC

  2. Jack Wakshlag from Media Strategy, Research & Analytics, August 9, 2018 at 5:16 p.m.

    This inventory is very valuable and growing. CPMs are good too. Right now it’s a relative trickle. It will be more. But counting your own impressions and expecting the advertiser to accept that number won’t work. And you have no demos. Advertisers will use their own servers to determine if the spot is delivered. You will have to reconcile their count with yours. More importantly, there is no persons based measurement. You don’t know who is watching, or how many are watching. If it’s done on a mobile device you might, but not on a smart TV. Without third party measurement, these impressions won’t get properly counted. 

    This is was the dream of those of us pushing TV Everywhere all those years ago. The idea remains a good one. 

  3. Dave Morgan from Simulmedia replied, August 9, 2018 at 5:35 p.m.

    Bob, very good points. Thanks for the feedback. I agree that there is no question that we are seeing more and more video ad availabilities in lots of other channels, and those will help absorb some of the advertiser demand that TV itself can't satisfy. However, those units are not being created at the same quality - 30 second of interuptive linear attention - and not at that scale - trillions. To add some perspective, as I like to write, Judge Judy every day delivers more audience ad minutes than all of the videos on all of YouTube in all of teh US all day.
    Of course, as we get better at targeting and optimization, we will make the fewer spots much more relevant and more valuable, and monetize better, so that the loss of ads in total might not be as devastating as it sounds.

  4. Ed Papazian from Media Dynamics Inc, August 9, 2018 at 6:34 p.m.

    Dave, I'm struck by your statement that Nielsen is showing a 9% decline in the audience of boadcast TV and basic cable on a year to year basis. The quarterly time spent reports Nielsen publishes on its website show a far smaller attrition in average weekly time spent with "linear TV"---live and delayed---which is averaging about 1-1.5 %. Indeed the latest report for the first quarter of 2018 for adults shows a drop of only a few minutes per day over a comparable period last year. I fear that you may be referring to average primetime rating losses for adults aged 18-49, which is true for the average broadcast network TV show but not for "pay TV" in its totality. The broadcast TV networks' primetime content accounts for a mere 9-10% of all TV consumption and adults aged 18-49 are a fraction of the broadcast TV audience.The broadcast TV networks' rating losses are both to SVOD channels like Netflix as well as  many other viewing options out there including  hosts of cable channels. In other words, I don't see the huge loss in commercial "impressions" that you are speculating about---not yet anyway. Did Kantar say that its 19 trillion figure represented a 9% decline over the previous year in ad impressions? If so, I'd like to see how they derived that estimate.

  5. Dave Morgan from Simulmedia replied, August 10, 2018 at 9:31 a.m.

    Ed, yes. I mislabeled the year-over-year number and will fix it in the piece over the weekend. The year-over-year decline for linear (including delayed) cable and English-language boadcast viewing according to Nielsen is 4%. The 9% number, as you correctly pointed out, is Nielsen's number in the adult (18-49) demo. It shows the same decline rate both for prime-time and all day. The 19 Trillion number is derived from stitching the Kantar ad occurrence data to Nielsen AMRLD data at the minute level. Year over year, that data actually shows a P2+ year-over-year decline above the Nielsen rate, closer to 6.5%. Projecting that out steady-state, it would suggest a 7 Trillion impression decline rather than 9 Trillion. Given the several year acceleration from classic pay TV bundles to skinnier SVOD & vMVPD bundles, both are probably within a pretty fair range of the likely linear (and delayed) TV ad impressions over the next five years. Thanks for catching and pointing out the mistage!

  6. Ed Papazian from Media Dynamics Inc, August 10, 2018 at 12:06 p.m.

    Dave, whatever the actual numbers are, you raise an interesting question. Even if the loss of ad "impressions" for linear TV  is not as great as some expect, some loss is virtually guaranteed. How will the sellers deal with this? Already they are relying to a great extent on retransmission fees and other non-ad incomes and this will only increase. However, if we look at "linear TV' as a media buy, I believe that it is vastly underpriced compared to digital media. Typically, a TV advertiser who uses a mix of broadcast and cable and all dayparts, pays something like 1-2 cents per viewer. That's way too cheap. In addition to all of the "advanced" TV schemes, which are predicated on charging far more per viewer under the guise of better targeting, the networks and major cable players, once they get their acts together, wil also move agressively to increase their CPMs---by virtue of the better targeting gambit or by selling a high proportion of short break ads at a CPM prmium, or by other means. The net result will that advertisers will have to get used to paying considerably more per viewer, which, in my view, is perfectably justifiable, considering what the going rate for digital search and DR buys seems to be. Also, this raises the question about corporate vs brand by brand buying. Will advertisers finally appreciate the value of modern day targeting and allow their brands sufficient freedom of choice to take advantage of what is available right now---even if the mechanics are somewhat flawed and the promises over hyped? One would think that that's a "no-brainer".

  7. Douglas Ferguson from College of Charleston, August 10, 2018 at 12:23 p.m.

    You completely lost me at "advertising can be an annoyance" when the irrefutable fact is that unsolicited advertising most certainly is an annoyance and waste of everyone's time. Except whose livelihood is threatened by the demise of advertising. If advertising was so wonderful, we would not check our email at every commercial break or use the DVR to avoid it.

  8. Sean Wyseman from several, August 10, 2018 at 2:17 p.m.

    I thnk the missing element is the breakdown on sectors. For example: how much of that is lost on News Network related advertising?, How much on sporting events?, How much lost on feature length movies?, I'm guessing the declines can be attirbuted to more granular representation of where the greatest percentages and greatest dollars are lost. Without that we cannot comment intelligently -- another reason people are abandoning media -- the media no longer cares for much beyond headline memes. Like the loss of the Hillary candidate -- the Democrats cannot believe that their candidate lost because she was a bad candidate. We need to know who the worst advertising revenue drops are. We can tell you a lot more when we know which sectors are carrying the most attrittion. Also pricing is probably needed to be reviewed. Gone are the days when the media's prices are not driven to the last drop by market demand or lack thereof.

  9. Dave Morgan from Simulmedia replied, August 10, 2018 at 2:36 p.m.

    Doug, no question that ads can be annoying, but they also pay the bills ($70B wiorth) to create and deliver content in the US. Plus, ad-support is particularly important to get video content to the 90+ million Americans without broadband. Also, it works. It certainly informs people of products and services that they then decide to buy.

  10. Craig Jaffe from Baruch College, Zicklin School of Business, August 10, 2018 at 2:37 p.m.

    Dave, your statement that streaming video-on-demand services are used by 2/3 of households sounds like it may be a penetration number, and not a number based on actual viewership. If that's the case, then it may not be appropriate to compare this with the other Nielsen data you collected for linear TV which you say is down.

  11. Dave Morgan from Simulmedia replied, August 10, 2018 at 2:47 p.m.

    Sean, here are some recent trends by type of programming, which may continue as linear TV declines: sports, down slightly; national news up slightly; dramas, down a lot; unscripted, down less; big networks, down big; small nets, flat to up; morning shows, down; daytime game shows, flat. I hope this is helpful. However, there is no question that the loss loss of so much ad-funding will undermine the platform of linear TV to support all kinds of programming. It could go into a death spiral.

  12. Dave Morgan from Simulmedia replied, August 10, 2018 at 2:52 p.m.

    Craig, that’s correct, but I think it’s fair to suggest that there is a correlation been the penetration of streaming services and the declines of linear tv viewing. It would be better to also show streaming usage by content type and household, but that data is very fragmented and sparse and not available in robust, representative sets.

  13. Ed Papazian from Media Dynamics Inc, August 10, 2018 at 4:01 p.m.

    Guys it's a really big mistake to equate the broadcast TV networks' primetime fare with the totality of TV. For example, even if the broadcast TV networks' primetime dramas are down more than some other genre---due, in such a case, largely, though not exclusively, to competition by Netflix and a few other OTT sources, drama reruns constitute a larger piece of the drama audience pie via syndication and these are much less affected. Also, and this is proven by one of the comparisons in "TV Dimensions 2018", as more and more programming becomes available total time spent with TV increases, though not proportionate to the increase in content. In other words, if there is 25% more content to watch, the average person's total viewing time may grow by 5-6%, which means that Netflix, Hulu, etc. have not only been competing with certain forms of "linear TV" content but have also generated additional viewing which is not necessarily at the expense of what is already out there.

  14. John Grono from GAP Research replied, August 10, 2018 at 9:58 p.m.

    Dave, you should NEVER confuse correlation with causation.

    There is also a correlation with population growth and the decline in linear TV viewing.   So does that mean, that the more people there are, the less linear TV viewing there is?   That would be non-sensical.   There is also a correlation with electric vehicle ownership ... gosh they must spend a lot of time ut there driving.   Hang on, NYSE as well.   Oh, and the BMI.   The list is endless, and sadly meaningless.

  15. Dave Morgan from Simulmedia, August 11, 2018 at 8:28 a.m.

    John, I agree. I wriite often about poor intermingling of correlatoin as causatoin. The point of this piece was not to establish that the adoption of streaming services were causing linear TV declines, though I certainy could have done that. It just would have taken several paragraphs to lay at the data at the individual household level showing that when a household adopts a streaming service, it watchs less linear TV. Given that those streaming services frequenlyt offer exactlly the same channels and content as the linear bundle that is being viewed less, it's not a stretch that we're seeing some direct consumer replacement going on.
    The point of this column though was to focus on the clear facts on the ground that linear TV is losing viewership - and ad load - at a large rate, and that this loss of ad impressions would have consequences for the industry. The relationship with streaming services is important because whether they caused the loss or not, they are growing, and they are a potential substitute for the lost ad impressions. Except, they have little or no ad load.
    I get the point about Correlation v. Causation - we all know that the opening of umbrellas at scale isn't the cause of rain. Rather, I'm trying to ask folks to pay attention to how hard it is raining and wondering what happens when it washes away one-half of linear TV's ads.

  16. Ed Papazian from Media Dynamics Inc, August 11, 2018 at 9:17 a.m.

    Dave, while I don't agree about the rapidity or the extent of the loss of TV ad GRPs, it's going to happen to some degree anyway. We should note, however, that national advertisers who spend about 55% of their TV dollars on basic cable buys are generating about 70% of their ad GRPs via such spending. This is due to cable's much lower CPMs compared to broadcast as well as higher ad clutter ratios. How much of a hit rating-wise will cable take if OTT is compting primarily with broadcast prime dramas and sitcoms for viewers? Who knows? Also, it is not at all certain that cable, which, as I noted, accounts for the bulk of TV's GRPs, will not respond to audience losses simply by adding more commercials---especially in the non-prime hours. Stay tuned.

  17. Dave Morgan from Simulmedia replied, August 11, 2018 at 10:38 a.m.

    Very good points Ed, but I think that the issue transcends GRP's. Today, unfortunately, because of the accelerating audience fragmentation on TV and the failure of how planning/buying/measurement to respond well, GRP's mean frequency, not reach. Thus, their not as substitutable as they were a decade or two ago. Your point about cable newtorks potentialy increasing their ad load is a good one, certainly given their history of doing it in the past. However, at 18+ minutes an hour already, I'm not sure that adding more wouldn't just accelerate the move by viewers to ad-free or ad-light alternatives.

  18. John Grono from GAP Research replied, August 11, 2018 at 7:30 p.m.

    Sorry Dave, but I have to disagree with your comment that "GRP's (sic.) mean frequency, not reach."

    As we all know GRPs stands for Gross Rating Points - the summation of the Rating Point for each spot in a schedule.

    What we also know is the first formula we are taught in media planning and buying - a formula that applies to all media and not just television:

         Average Frequency = GRPs / Total Reach

    The formula is written this way as the Frequency Distribution table is derived from those two primaty metrics (using elemental data).

    So put another way:
        Total Reach = GRPs / Average Frequency

    The goal of ANY media buy (not just TV) is to have the next insertion in the schedule to produce the maximum INCREMENTAL reach per dollar.

    Ergo GRPs mean more about Reach than Frequency.  Unfortunately, some media push the line that TV is all about Frequency as it is beneficial to themselves.

    Note to advertisers - if your media agency is focussing on Frequency, get a new one.

  19. Dave Morgan from Simulmedia, August 11, 2018 at 9:16 p.m.

    John, my piont on a GRP today being all about frequency was not suggesting that the definition has changed - a GRP is still R x F. My point is driven by the reality of today's GRP's. Because of massive audience fragmentation and the failure of the industry to change how we package/buy/sell TV ads - and in spite of many, many warnings from Erwin Ephron - a GRP today in practice contains 9X more frequency than it did 20 years ago. For sure, it is not "all about" frequency, but it might as well .

  20. Ed Papazian from Media Dynamics Inc, August 12, 2018 at 6:45 a.m.

    Dave, I have to disagree with your statement that today's GRPs contain nine times more frequency than they once did. We revise the reach tables in "TV Dimensions" periodically and did so recently to reflect the fragmentation of reach caused by so many channels being available and, yes, 50 GRPs, today delivers less reach than it did in 1985 or 1990. But the differential is not even close to the amount you suggest. Also, we must remember that one can calculate reach in many time frames--a day, a week, a month, a quarter, etc. As the length of time expands the relative value of a rating point in adding reach increases, so we can't accept a single time frame---like a week--- as an indicator of reach for a particular ad campaign which usually extends over the course of several years before a new approach is contemplated or the results prove ineffective.

  21. Dave Morgan from Simulmedia replied, August 12, 2018 at 12:12 p.m.

    Ed, you are correct that the "average" R/F balance in GRP's hasn't changed as dramatically over the past 20 years, though it the F component is up a lot on average. However, if you de-average the Reach and Frequency at a person/impression level is where you see the massive increaase in Frequency as a component of the GRP. A theatrical release 15-20 years ago might have been supported with $15-20 million on TV for 2-3 weeks before the premier. It would have reached 90% of total audiences on TV (and similar portion of frequent movie-goers) 10-12 times each on average over those 2- 3 weeks, with the heaviest quintile of TV viewers getting 15-20 ads each and the lightest quitile getting 5. In that scenario. Today, since theatrical release TV buyers are super focused on primetime, 18-49 demo and shows with the highest average rating point, which is opposite to where fragmenting audiences now view TV, they will spend $50-60 million on national TV in 2-3 weeks. They will reach 80% of total TV audience (and similar portion of heavy moview goers), but the lightest quintile of TV viewers will get no ads over that time, the next lightest will get 1 ad, and the heaviest quintile will get 50+, and the next heaviest at 20. The average Reach and Frequency balance will still look OK - total reach of 80% with an average of close to 10, but more than 80% of the total GRP's will have been delivered to viewers who had already seen the ad 10 times in the preceeding two weeks. That is where the massive increase in frequency has happened - the crabgrass that Erwin warned us against.
    You don't have to look at different timeslots to come up with these numbers.
    Key takeaway - we need to stop planning, buying and measuring TV ads by "averaging" methods that don't work in today's fragmented TV ad world. We have tons of data at the person/impression level, whether in directl measured at the set-top box or Smart TV, or in the Nielsen AMRLD. That is where the truth of today's GRP is told.

  22. Ed Papazian from Media Dynamics Inc, August 12, 2018 at 1:30 p.m.

    Dave, currently, the heaviest quintile of primetime viewers accounts for roughly 45-50% of the viewing while the lightest does only 2-3%. If we take your example of a primetime movie buy across the total population you are talking about an average reach of 80% with a 10 frequency---or 800 GRPs. If we see howthis probably looks by primetime quintile, the GRPs for the heaviest quintile should come in at about 1800 which would generate a 95%+ reach; in contrast, the GRPs for the lightest quintile would be a mere 100 generating a reach of about 50%. While this represents a higher disparity between the quintile extremes than existed twenty years ago, it is not a gigantic shift. More important, the difference between the quintiles is not nearly as great as you postulate. It's more like 15-20-to-one as opposed to 50-to-one.Advertisers have been coping with this kind of thing ever since measurements of TV's audience alerted them to the matter.

    I do agree with you, however, that the way that corporate upfront buys are divvied up among the brands often forces them to accept ad positions derived from low CPM but often off-target shows that were bundled together by the sellers and time buyers and that much greater care needs to be taken in this regard---but that assumes that the brands are released  at least to some degree from the shackles of corporat time buying.

  23. Dave Morgan from Simulmedia replied, August 12, 2018 at 7:56 p.m.

    Ed, you are assuming that the theatrical campaigns' frequency delivery correlates with the amount of total viewing by overall TV viewing quintiles. It does not. Almost no campaign I have looked at does. Current bulk buying methods dramatically oversaturate a portion of total viewers at an even more skewed precentage than the rest of thier viewing time cohort. Who are they? They disproportionally watch broadcast prime and highly rated shows. There was day (many years ago) when most/all viewers watched like that. Now, only a subset do.
    Under any circumstance, you I agree on one of the biggest barriers to smarter, better delivered campaigns. We need to unshackle brands from bulk corporate buys. Bulk buying (and allotment) only makes a difficult situation totally untenable.

  24. Ed Papazian from Media Dynamics Inc, August 13, 2018 at 3:37 a.m.

    Dave, one final point on the math. If you are correct that the heviest movie schedule quintile out views the lightest by a margin of 50-to-1 among those reached by the campaign this would eat up almost all of the GRPs. Based on a total reach of 80%, the heavy viewing quintile would comprise 16% of the target group but would receive almost all of the movie GRPs. In other words, the lightest viewing quintile as well as the next two quintiles would get no GRPs, hence no reach. Your total reach couldn't be 80% it would be more like 30%.

  25. Dave Morgan from Simulmedia replied, August 13, 2018 at 8:19 a.m.

    Ed, you are directionally correct. The heaviest viewing quintile doesn't eat up all of the GRP's, but it gets a masssive amoiunt of them. And, yes, the actual reach on some of their campiagns now falls pretty low and doesn't always get to 80% ... though not down to 30%, but frequently 50%.

  26. Ed Papazian from Media Dynamics Inc, August 13, 2018 at 8:50 a.m.


Next story loading loading..