Are National TV Ads Actually Too Cheap?

  • by , Featured Contributor, August 23, 2018

Did you know that the average price across all national TV advertising in the U.S. is less than $2.50 CPM? In fact, $2.26, to be more exact.

Are you surprised by how low that number is? Most people I talk to are, particularly if they come from the digital world. There, most branded premium video ads against broad targets sell for CPMs more in the $10 to $75 range. Most folks assume TV is at least as expensive.

Where did I get that number? I based it on Nielsen AMRLD data and Kantar ad occurrence data, in which national TV broadcast and cable networks delivered 19 trillion ad impressions over the past year. According to Magna Global’s numbers, national TV advertising generated $43 billion last year. Divide the second by the first and you get a $2.26 cost per thousand. Pretty sobering.

Over the past week, I’ve asked a dozen folks in the industry to make their best guess of the average CPM of national TV advertising in the U.S. Only two were close (almost exact, actually), but each of them run TV network sales groups. All of the others, mostly digital folks, guessed dramatically higher numbers.



Why this disconnect to reality?

Demo and dayparts confuse the macro story. Almost every time you hear or read about TV ad CPMs, it's in terms of the 18-49 demographic segment (only 40% of the TV audience) and relative to prime-time shows (perceived as the most valuable and scarcest). And, sometimes, the numbers even have the “broadcast” caveat, which typically have higher prices than their cable programmer brethren. Virtually all trade and business stories about TV advertising tend to use language like “$42 CPMs against the demo on broadcast prime,” not “all P2+ (persons two and older) and aggregated across all networks, days and dayparts.” How quickly $42 can become $2.26.

TV measurement language is very different from digital. TV people talk differently than digital people, and vice versa. Most digital folks I talk to don’t even know the term “P2+.” Enough said.

TV advertising delivers massive scale, which is underestimated by most. Once again I remind you that today, “Judge Judy delivered more audience ad minutes (more people in the U.S. watching more video ad time) in just two 30-minute episodes than all of the video on all of Google’s YouTube watched in all of America all day. Yes, all videos in all of the U.S. all day long.

What are the implications of TV advertising's $2.26 average CPM? First, to anyone who understands the power of media, and the unique power of TV's sight, sound and motion and quick sales-driving impact, you can’t come to any conclusion but that a lot of TV advertising is underpriced. 

For sure, most marketers and most TV ad buyers can’t access TV in the ways that they need to exploit it in the way that they can with digital deeper targeting datasets, ability to pick and optimize at the spot level, fast reporting and fast creative swapping, centralized cross-network optimization. However, there are initiatives under way at all of the TV companies to solve these issues.

What might the future hold? I believe that better use of data, science and software is going to enable TV companies to drive higher prices over the next three or four years' prices -- $3.50+ CPM on average is certainly reasonable -- and still deliver even better and more predictable ROI for advertisers.

That will happen, and most folks will win. TV networks will get more yield on their audience. Advertisers will get more customers. Nielsen and other data companies will sell more targeting and measurement data. And consumers will get more relevant ads.

What do you think?

39 comments about "Are National TV Ads Actually Too Cheap?".
Check to receive email when comments are posted.
  1. Jack Wakshlag from Media Strategy, Research & Analytics, August 23, 2018 at 6:14 p.m.

    Price, of course, is driven by supply and demand. So long as advertisers, particularly large advertisers, concentrate disproportionately on broadcast prime and digital versions of the same, there will be huge opportunity for others to buy what they don’t — inventory on the much less pricy cable TV. Years ago Turner’s Barry Fischer developed Media at the Millennium showing how easy it is to reduce pricing and get reasonable cost per reach points. It’s almost 20 years since then. What’s the problem?

  2. Craig Mcdaniel from Sweepstakes Today LLC, August 23, 2018 at 6:40 p.m.

    Hi Dave, Question. Is the lower prices from TV also causes the online ad prices to go down? Over the last 5 or 6 years, online ad prices have gone down a ton also even though there seem to be more ads and advertisers than ever. Thoughts?

  3. Paula Lynn from Who Else Unlimited, August 23, 2018 at 7:11 p.m.

    Is this all based on national buys rather than an accumulation of regional, spot or more specialized local buys ? Would the decrepancies be as large if it was an accumulation ? 

  4. Neil Ascher from The Midas Exchange, August 23, 2018 at 7:57 p.m.

    Dave, you do a better job representing the broadcast and cable nets then they ever have. They should pay you commission on everything they sell!

  5. Craig Mcdaniel from Sweepstakes Today LLC replied, August 23, 2018 at 8:46 p.m.

    Hi Neil, I sent a request to Midas Exchange just now. If you would be so kind, could you read my message?  Thanks, Craig McDaniel

  6. Dave Morgan from Simulmedia replied, August 23, 2018 at 9:45 p.m.

    Totally agree Jack. Barry's work is some of the most impactful and revolutionary in the media busienss. It was all centrd on client value. Way ahead of its time.

  7. Dave Morgan from Simulmedia replied, August 23, 2018 at 9:45 p.m.

    Craig, without question, as metrice betwen online and TV become more comparable, online ad prices ill down. For sure.

  8. Eric Fischer from HJA Strategic Consulting, August 23, 2018 at 11:12 p.m.

    If I read correctly, the $2.26 CPM number is based on the capacity of impressions (19T), not the actually number of impressions purchased.  The actual CPM of paid media schedules would be higher.  Dave, does that 19T number include promo time, local affiliate breaks, PSAs, etc...?  If so, how does that impact the total number of impressions and subsequently the CPM?

  9. Jack Wakshlag from Media Strategy, Research & Analytics replied, August 23, 2018 at 11:50 p.m.

    There is no way to calculate a capacity of impressions. This is based on actual deliveries across all Ad supported cable and broadcast channels. 

  10. Dave Morgan from Simulmedia replied, August 24, 2018 at 5:50 a.m.

    Eric, the $2.26 CPM is based on the actual impressions delivered and the actual total revenue they generated, not the capacity.

  11. Ed Papazian from Media Dynamics Inc, August 24, 2018 at 9:07 a.m.

    Dave, whatever the real CPM numbers are, I agree with you that TV is underpriced not only nationally but locally. and especially relative to digital media. In fact, subscribers to "TV Dimensions 2018" have received a number of "Alert" report on this subject and ways that the sellers might take long overdue corrective action. So far the sellers seem incabable of taking the right steps to orchestrate an across-the-board adjustment of their CPMs---but, perhaps, they will seek out some advice from "outsiders" who think they have some answers.

  12. Neil Ascher from The Midas Exchange replied, August 24, 2018 at 9:27 a.m.

    Hello Craig,

    I did not see a message from you.  You can message me directly on LinkedIn if you like.

  13. Douglas Ferguson from College of Charleston, August 24, 2018 at 3:39 p.m.

    Thanks to my DVR, it's been quite a while since I've even seen a national TV ad.  Even in live shows, I can jump to another channel or recorded show and then come back to skip forward to almost-live.

  14. Jack Wakshlag from Media Strategy, Research & Analytics replied, August 24, 2018 at 3:48 p.m.

    Doug, this behavior is well known and has been for years.  All who understand TV measurement are aware of how people skip and avoid ads. That is why the industry today generally conducts business on the basis of viewership to minutes that contain ads. It has done so for years. It’s not perfect, but it does account for ad skipping, avoidance etc. I’m sure you know this is so. 

  15. Ed Papazian from Media Dynamics Inc, August 24, 2018 at 7:28 p.m.

    But Jack, Nielsen has no way of knowing whether its "average commercial minute viewers" are even in the room, let alone paying the slightest attention  if present. We figure that only half of the reported average comercial minute "audience" actually notes a given commercial in that minute---with variations above and below this norm depending on program type, time of day, degree of ad clutter in break, etc. All of this is covered in detail in "TV Dimensions 2018".

  16. Jack Wakshlag from Media Strategy, Research & Analytics replied, August 24, 2018 at 8:42 p.m.

    Obviously true Ed and we will never know that apart from estimates like you provide through your service. Still it is a good place to start, in my view. Adjustments can be made later. In the digital world you also don’t know whether an ad has been viewed in whole or in part, or even by a real person. Still, the prices strike me as very different because of supply (low in digital) and demand (high in digital). That’s is the only thing that will change pricing. Advertisers get smart about the value not leveraged in traditional tv video, will find lots of bargains. 

  17. John Grono from GAP Research replied, August 24, 2018 at 8:58 p.m.

    Ed, while whay you say is correct, Nielsen (here in Oz) do telephone 'co-incidental' studies.   Basically they telephone the panellist homes and ask something along the lines of "when the telephone rang which of your TV sets were on and what channel/programme was tuned and who in your family was watching".

    This data is then compared back to the peoplemeter data that had already been (or was currently being) captured and sent back overnight.

    I don't have any recent data, but from memory in the first decade of this century it averaged in the low 90% range but with some demos (younger ones) being in the mid 80s.

    I wonder if Nielsen US conducts such on-going studies - worth asking?

    And by the way, maybe TV is not underpriced but digital is over-priced?   Replies on the back of an envelope please.

  18. Ed Papazian from Media Dynamics Inc, August 25, 2018 at 12:17 a.m.

    John, Nielsen has conducted such tests in the past, though I'm not aware of any that dealt with commercial minutes. My issue with the latter, if conducted, is that the respondent in the telephone coincidental is almost certain to answer for program content not commercials---unless the latter are specifically cited as the subject ( "Was anyone watching the commercials that were just on?" ) which would very likely generate a much lower and possibly a false finding. The typical peoplemeter result for commercial minutes tells us that almost all program viewers---over 90%---also watch every commercial minute, which flies in the face of what the various observational studies as well as what viewers, themselves say.

  19. Guy Powell from ProRelevant, August 27, 2018 at 1:16 p.m.


    Great article.  I like the comparison, although it would seem that with digital I can target generally better and then als with retargeting can target based on behavior.

    Nevertheless, very interesting how TV can target the mass market much less expensively.


  20. Jack Wakshlag from Media Strategy, Research & Analytics replied, August 27, 2018 at 2:33 p.m.

    In theory digital can target better. In reality, when you look under the covers at the data used for targeting the “emperor has no clothes.”  See recent article by Neumann et al. that confirms what should be well known. That the data used for targeting is so bad you might as well use a monkey to choose. More importantly the cost of targeting — the targeting tax — paid to all the middle men simply is too large to justify the cost. There is less waste in not targeting than what is paid to these folks. 

  21. Jack Wakshlag from Media Strategy, Research & Analytics replied, August 27, 2018 at 4:02 p.m.

    Here is the reference for the article showing how poor the data used for targeting is:

    Neumann, Nico and Tucker, Catherine E. and Whitfield, Timothy, How Effective Is Black-box Digital Consumer Profiling and Audience Delivery?: Evidence from Field Studies (June 25, 2018). Available at SSRN: or

  22. John Grono from GAP Research replied, August 27, 2018 at 5:25 p.m.

    Jack, thanks for referencing Nico's work.   He is one of group of Australian academics that are belling the cat (Ritson, Nelson-Field, Sharp).

    Guy, I suggest you get hold of Prof. Byron Sharps' "How Brands Grow".  I must say that when I first read it I was shaking my head and going "Nah...".   So I immediately read it again and realised how perspicacious and challenging his thinking is.

    In a nutshell, if you know who your customers are then digital is a very direct medium to use to sell product.   If however, you want your brand to be big and grow further, then in the main traditional media and especially television remain king.   One marketer's "wastage" is the savvy marketer's "customer and brand growth".

  23. Ed Papazian from Media Dynamics Inc, August 27, 2018 at 6:14 p.m.

    Good discussion, guys. One additional clarification. When we say that TV excells at "reach" this is sometimes taken to mean that TV reaches so many non targets that this overcomes its failure to target those the advertiser values the most. Actually, TV reaches just about  everybody, including the advertiser's prime prospects---plus many secondary targets and might-be buyers. Also, TV gives each advertiser the same chance---a 100% chance---to garner actual commercial viewing every time an ad is placed. This is not even close to being the case for many digital campaigns---no matter how well they seem to targeted. You can't sell a product if your ads are blocked, never get to the users' screens, go to bots or are only on-screen for a few seconds.

  24. Jack Wakshlag from Media Strategy, Research & Analytics, August 27, 2018 at 6:42 p.m.

    I am familiar with Sharp’s work and that of Andrew Ehrenberg whose line of work Sharp has brilliantly extended. It was my pleasure to work with and publish with Andrew, and currently serve as Chair of Ehrenberg-Bass Institute’s North American Board.   

  25. Jerome Samson from 3.14 RMG, August 28, 2018 at midnight

    But wait, Dave - what's the average CPM for online impressions these days? If we're talking about all ad impressions on national television, then shouldn't we look at all ad impressions online? Not just for 'branded premium video ads' or even YouTube pre-rolls, but good ole display ads as well. Otherwise you're pitting the whole 'tonnage' of television against the cream of the crop online, and that's not really a fair fight.

  26. Dave Morgan from Simulmedia replied, August 28, 2018 at 5:27 a.m.

    Jerome, "all ad impressions" on TV are premium video (certainly by the digital definition), so comparing TV to premium digital video is certainly the closest to an apples-to-apples comparison, even though the average viewable digital video ad is 6-10 seconds while on TV 30 second is the norm, with 15 seonds and 60 second ad quie common as wel. Comparing TV to digitial banners make no sense, since most web pages display at least six banners at once and TV only plays one ad at a time - no comparison of the relative share of voice delivered.

  27. Ed Papazian from Media Dynamics Inc, August 28, 2018 at 7:10 a.m.

    Absolutely right, Dave. In fact, if you compare CPMs between TV and digital video for "15s" using 100% viewability for both---which TV automatically delivers----TV is usually much cheaper than digital.

  28. Jerome Samson from 3.14 RMG replied, August 28, 2018 at 11:52 a.m.

    Hi Ed, could you share with us a CPM number for 100% viewable 15s digital video ads, and what that number is based on? I'm being over dramatic with my note that we should compare to online display ads, I'll admit! But I don't agree that all TV ads are equivalent to branded premium digital video ads. Unless I'm misunderstanding what you mean by branded premium video ads, Dave. To me, those digital ads have an inherent 'premium' targeted component that we're not accounting for on the TV side if we're looking at the whole thing.

  29. Ed Papazian from Media Dynamics Inc, August 28, 2018 at 12:16 p.m.

    Sure, Jerome. This is from our annual, "Intermedia Dimensions 2018". We don't do teens and children but focus on adults. Taking the typical TV branding advertiser buy for digital video, counting mobile, tablets and desktop, the CPM is about $9.75---obviously with considerable variation around the norm. This is a "raw" CPM estimate. If one adjusts for 100% ad viewability, you deduct 45-50% for ads not on screen plus another 25% for ads not played through their entirety and the result is a CPM of around $40. If the buyer is buying "premium" or "targeted" placements as opposed to "impressions" the CPM would probably double. By way of comparison, the corresponding figure for TV, counting all dayparts and network types is about $7.50 for "15s". As regards sources, these are our own independent estimates based on many referrence points and, as I noted, there will always be variations around the norms for both TV and digital.

  30. Jerome Samson from 3.14 RMG replied, August 28, 2018 at 1:55 p.m.

    Thanks, Ed! Really helpful. I know there are viewability and bot considerations of course, but once those are properly addressed they won't really translate into higher CPM, so I think your $7.50 and $9.75 estimates are a good basis for comparison. I'm in agreement with the gist of Dave's article, but the $2.26 to '$10 to $75' comparison seemed a bit of a stretch to me.

  31. Ed Papazian from Media Dynamics Inc, August 28, 2018 at 2:48 p.m.

    Jerome, I'm not sure that I get why the application of adjustments to account for lack of ad viewability, fraud, etc. doesn't raise the effective CPM for digital video in comparison to TV CPMs. If I'm charged, say $10 per thousand for so many digital video impressions but I deliver only 25% of them yet still collect the full amount specified for the buy, haven't I raised my CPM?

  32. Jerome Samson from 3.14 RMG replied, August 28, 2018 at 3:53 p.m.

    It raises the effective CPM, for sure, but not the contracted CPM: if you're paying $10 for a thousand impressions, and you find out that 500 of those are invalid (for one reason or another), you've indeed effectively paid a $20 CPM. It doesn't mean however that it's fair value for the buy: you've been overcharged. Imagine now that the publisher corrects the problem and fully delivers on the contracted 1,000 impressions: you won't all of a sudden be okay with paying $20 for them, will you? In other words, when the 'fraudulent' impressions disappear, the effective CPM will go down accordingly to match their value (the $10 contracted rate). It's all a bit of semantics of course and we're on the same page - I'm not saying that there isn't considerable waste in digital buys, but the fact that digital publishers are getting away with it doesn't mean that their impressions are worth $20.

  33. Craig Mcdaniel from Sweepstakes Today LLC, August 28, 2018 at 4:45 p.m.

    Dave, as you know, I don't sell TV advertising but have be following you for other reasons. You wrote maybe a year ago or longer about "trading desk" in TV ad sales. It appears to me that selling/trading in the TV industry is changing again especially with Google knocking on the door. I am in the camp that I hate automatic and programmatic trading and would love to see person to person trading come back to the net ad sales. The whole person to person sales seems to be a part of this discussion I want to hear from you to see what your thoughts might be now.

    Craig McDaniel
    Mr. Sweepy

  34. Ed Papazian from Media Dynamics Inc, August 28, 2018 at 4:52 p.m.

    Jerome, I agree that if the shortfall in actually delivered and viewable impressions is made up for without charging the advertiser more then the original CPM is valid. But it now seems that a large percentage of digital media buys do not see such adjustments or "make goods" , so the adjustment in such cases is merited.

  35. Dave Morgan from Simulmedia replied, August 29, 2018 at 8:46 a.m.

    Craig, I do think that we will see more automated trading on TV soon for sure, but human tradking still represents 99+% of all TV ad transactions, so I wouldn't worry too much about human trading disappearing anytime soon. I suspect that, at best, automated trading will represent 50% of TV trqnsactoins in 5 years.

  36. Ed Papazian from Media Dynamics Inc, August 29, 2018 at 10:57 a.m.

    Craig, I agree with Dave about the failure of automated TV time buying to make inroads on a national basis. At present, all of the small number of sales made under this promotional banner by some of the networks are single seller transactions---- tightly controlled by the sellers--- where the buyer has no alternative but to take what is ultimately offered---or "walk". Also, humans make all of the decisions. The same thing is true for local---or "spot"--TV but here there is more movement, though I doubt that we are going to see anything that approaches "automated"  buying and selling at scale for a long, long time. Unlike digital media where the number of "GRPs" is infinite, this is not so in TV.  The sellers have no incentive to give away their premium and most other content at bargain basement rates. So virtually all of the spot TV buys that are placed via "programmatic" platforms are neither automated---humans are still in charge--- nor do they offer buyers the capability to evaluate all of the competing options as many stations still do not participate or do so only on a limited basis. Not to worry, the machines are not replacing buyers and sellers---not in TV at any rate.

  37. Jack Wakshlag from Media Strategy, Research & Analytics replied, August 29, 2018 at 12:07 p.m.

    Important to note that spots that go unsold still generate revenue. Channels fill their full ad loads and do not “go dark” if an avail goes unsold. For decades (remember Ron Popeil), Television has had a robust and pretty lucrative Direct Response business. Spots not sold by humans can be filled with ads like these with decent revenues. That is one reason TV hasn’t had to jump on the “programmatic” bandwagon. 

  38. C D, September 4, 2018 at 3:39 p.m.

    The average CPM for National TV is not $2.26. Im pretty confident that agencies could give you a more reasonable estimate (there are major companies whose sole business is based on reporting this data). Frankly, I’m concerned about the two network execs who guessed close to that number.

    Has anyone checked the math here? 18 trillion impressions and $43 billion in sales?

    I’m guessing the impressions were calculated by average impressions multiplied by commercial seconds, right? That would give you a $2.26 CPM...but advertisers are not paying by the second. You need to equivalize against a 30.

    Furthermore, the Judge Judy stat doesn't appear to be close to accurate. The average Judge Judy telecast did just under 9.8 million average minute impressions. Episodes are 30 minutes long. So, two episodes will deliver 587.8 million impressions (total impressions, ads would be less).

    YouTube, per comScore, averages around 9.1 billion minutes per day. Again, not all ads...but this stat doesn't pass the sniff test either.

  39. Pravin Chandiramani from Simulmedia, September 5, 2018 at 7:59 p.m.

    Hi CD@TWWL

    C D@TWWL
    Great Question.

    Firstly, to clarify our sources. We sourced the impression data from two Nielsen products i.e Nielsen Ad Intel and Nielsen All Minute Respondent Level Data (also called AMRLD). The total US TV Ad spend number ($43B) is from Magna Global - see  link

    The impression number is the Nielsen P2+ impression count of all TV spots (from Nielsen Ad Intel) across all nationally rated TV networks. The impression count for each spot represents the total number of people (as estimated by the Nielsen panel) who watched the entire spot for entire 30 seconds (we equivalized to a base of 30 seconds). We did not perform the calulation as you indicated i.e. by taking "average impressions multiplied by commercial seconds" for the exact same reasons you have stated.

    Lastly, companies such as Kantar, SMI, SQAD etc. report on such information by limit it to ad spend by network, by brand or some other attribute. However, to my knowelge none of these companies reconcile ad spend back ad impressions and CPMs.

    More than happy to answer additional questions.

Next story loading loading..