CPM Pricing is Doomed

Since it is the time of year that prognosticators attempt to predict the future, let me offer this: Within five years, CPM pricing will account for less than 20% of the total market.

CPM pricing will eventually be relegated to a few hundred sites like WSJ, TechCrunch, and GQ where advertisers are willing to pay for the credibility that comes from associating with these publishing brands.

The rest of us will be forced to sell on performance pricing models.

As humans we are incredibly adept at rationalizing our own thinking. We fall in love with the things we are good at, and find reasons to hate the things we struggle with. The real reason publishers hate CTR, is becausewe are not good at driving clicks. If online display had a 20% CTR, we would all exclusively sell on CPC.

Before you fill my inbox with hate mail, realize that I am not saying that online display advertising does not work. In fact, quite the opposite is true. However, when measured by CTR, we underperform, and advertisers are voting with their wallets.



The truth is in the numbers. A study by PricewaterhouseCoopers found that from 2007 to 2008, CPM-based pricing dropped from a 45% share to a 39% share, while performance-based pricing climbed from 51% to 57%. In 2009, the numbers continue to shift in favor of performance pricing. This is a tough trend to ignore, and as a publisher, you do so at your own peril.

The Problem Started Before the First Online Ad Was Ever Displayed
Ironically, CPM pricing started to die with the rise of the CMO into the executive suite.

In the last 20 years, the CMO has come of age. The 4,000 CMO's that belong to the CMO Council represent more than $90 Billion in marketing spend. Marketing was once considered a semi-serious position that often reported to the VP of Sales or CFO, but today CMO's have an equal seat at the executive table. Along with that seat came a new set of responsibilities. They now report directly to a CEO who is going to demand accountability for every dollar spent. The recession only magnifies the problem.

CMO's are a rare breed whose average tenure is shorter than the lease on your car. Fast Company called it "The Most Dangerous Job in Business," and Business Week said that "few CMOs can live up to the sky-high expectations."   The average tenure of today's CMO is a mere 28 months.

CEOs keep their jobs by growing the business, and nobody ever appeased Wall Street by pointing to their branding impact. If you sit in the executive suite, your job is to support your CEO. As a result, they are not buying branding, they are buying customers.

It All Boils Down to Risk
The question of whether pricing should be CPM or CPC can be distilled down to one very simple question: Who should take the risk?

CPM is a metric that measures a publisher's success. CPC is a metric that starts to measure advertiser success. In a world where inventory is limited and a few people controlled access to the audience, publishers were able to sell on CPM. Unfortunately, that is no longer our reality and will not be in the future.

There are more than 3 trillion impressions each year that go unsold. As more people around the world go online and Internet usage continues to climb, this number will only increase.

When supply outstrips demand, advertisers have a plethora of options, and are able to demand CPC (or even CPA) pricing.

Luckily, this is not necessarily a bad thing.

Two Roads Diverged in a Yellow Wood
As publishers, we are faced with the biggest challenge of our collective careers, and there are only two options: convince advertisers to buy into a new set of metrics where we excel, or, figure out how to get really good at CPC.

Focusing on new metrics is a huge mistake. We have tried this approach and failed every time, because advertisers are going to buy the model that most directly correlates to driving their revenue. Trying to convince them otherwise is a waste of time and energy. Instead, we need to get great at CPC.

I Chose the Path Less Traveled By, and That Has Made All the Difference
This is my last column of 2009. I look forward to continuing the conversation after the New Year, and will focus my first column in 2010 on how publishers could make a fortune selling on CPC.

Thank you for all of your compliments, criticism, new ideas, enthusiasm, and most of all, your passion for this industry you have offered me this past year; it has made me better.

I am honored to be a part of this industry.

31 comments about "CPM Pricing is Doomed".
Check to receive email when comments are posted.
  1. Jason Krebs from Tenor/Google, December 11, 2009 at 2:15 p.m.

    This is all sad but true.

  2. Nelson Yuen from Stereotypical Mid Sized Services Corp., December 11, 2009 at 2:18 p.m.


  3. Mike Molnar from Glow, December 11, 2009 at 2:18 p.m.

    I'm curious how the CMO's "prove" on-air, magazine and outdoor to their CEOs. None can "prove" an actionable result in a quantifiable way. Does CTR prove a purchase occasion took place or only assume a consumer is "more likely" to make a purchase? Does the latter actually mean anything?

  4. Gary Kreissman from Group PRM, December 11, 2009 at 2:29 p.m.

    As an advocate of performance marketing, I applaud your perspective. Your assessment is correct and it will lead to media success being driven by the consultative sale. The more publishers can provide insight on their audiences and collaborate with advertisers on offers and messages that motivate key segments of those audiences, the more dollars will flow in their direction.

  5. David Chu from Eton Corporation, December 11, 2009 at 2:38 p.m.

    Hey Mike,

    It's possible to 'quantify' TV ADs by measuring the lift or drop that you get during a campaign. It's not perfect, but it works. An example is how Coke pulled it's famous "Mean Joe Green" Ads because even though people loved the commercial, it wasn't leading to more sales.

    With CTR, you assume that not all of the clicks are real, but that a certain percentage will convert into sales, making it easy for the CMO to set the budget and prove the value of the ad to his superiors.

  6. Jonathan Mirow from BroadbandVideo, Inc., December 11, 2009 at 2:38 p.m.

    CPM, CTR, CPC, CMO, CEO, CFO, CPA....WTF? (But I agree with the gist)

  7. David Chu from Eton Corporation, December 11, 2009 at 2:40 p.m.

    Hey David K,

    Let me know what you think of this explanation.

    Advertisers like CTR because it makes it easier for them to budget and carries less risk.

    Networks and publishers like CPM because it makes it easier for them to budget and carries less risk.

    The problem is that while content is growing exponentially, advertising budgets are shrinking.

    Thus, CTR will win out.

  8. Erin Ulicki from Centro, December 11, 2009 at 2:42 p.m.

    Mike Molinar, you summarize the question I was going to raise.

    I am not saying the CPM model can be improved, but I'm very skeptical of hitching my horse to the CPC wagon just thinking about how often I click on ads (and this has been my industry for 10 years). It doesn't mean the branding for Kashi bars isn't working on me or the eBay ad that was served up today - I saw them, I didn't click on them, but I did buy 4 boxes of Kashi bars yesterday and I thought of checking out eBay later today.

    David, I'm looking forward to your CPC column in 2010.

  9. Paula Lynn from Who Else Unlimited, December 11, 2009 at 2:57 p.m.

    Ever so true. What is gained from measuring is lost in patience. For the most part ( not all, please understand), the actual results of the coordination of marketing and advertising cannot be seen for months and where it is seen best is in the sales figures. Millions and millions of people see a site, a banner, a photo and never click it on. And just because there is a click, it doesn't mean the product is purchased. I look at cars for what I will be interested in 3 years from now when I buy one off a lease. Would my impression count? Would my click count? Should it? Somebody is going to sell me a car in 3 years and what they do now will add to that decision including quality assessments.

  10. Jeff Sherman from, December 11, 2009 at 3:26 p.m.

    After analyzing the highest and lowest performers from its database of more than 170,000 online ads, the Millward Brown company determined that creative factors such as persistent branding, strong calls to action and even human faces -- and not super-targeted or high-profile ad placements -- make for better ad recall, brand awareness and purchase intent.

    It's a balance.

  11. Greg Strauss from river town newspaper group, December 11, 2009 at 3:32 p.m.

    The whole CPC concept devalues the fact that view through and dwell time exist. Hey, who needs branding? On top of that, what good is a click anyway? Step up and lets all jump on the CPC bandwagon, eh?

  12. Greg Strauss from river town newspaper group, December 11, 2009 at 3:34 p.m.

    The true gold lies in conversion don't come looking for a free impression on my viewership.....

  13. Dana Todd from SRVR LLC, December 11, 2009 at 3:37 p.m.

    This is a rather pointless discussion. Whether an advertiser is buying on performance or not, it still all backs into an eCPM on the publisher side. If they're only getting paid for clicks, it yields a very low eCPM (or a better measurement is, revenue per pageview). If the revenue doesn't let them make payroll, then how the heck do you expect quality publishers to survive?

    If I wear my ad-buyer hat, of course I want all the risk to be on the media side. Then I take no responsibility for my ad creative, communication strategy, etc. But when I put my publisher-side hat on, the CPC model so deeply devalues the properties that I build with blood, sweat and tears, that there's no way I'm going to bet my entire business model on performance only.

    Now...if I'm a splog network, and I'm cranking out vats of content with CPC ads on them, that's a different story. In that case, the eCPMs they earn (below $.50) are priced just right. 'Course, advertisers are pretty much unaware of how much banner ad click fraud occurs in these types of networks. But hey, that's a different column.

    Just sayin' - if you look around at what performance pricing has wrought on the digital content world, we've achieved: crap content, teeth whitening ads, and a collapsed news publisher market. If a publisher's only recourse is to start working the numbers and churning out as much made-for-clicks content as they can, we're putting ourselves on a dangerous course for everyone involved.

    Another rant: Lead-gen companies drive a lot of CPC business, but their own business models are very cyclical - they'll pull out as soon as they're not able to sell their leads in a vertical. So if publishers put all their eggs into these baskets, the volatility alone can jeopardize their livelihoods.

  14. Dana Todd from SRVR LLC, December 11, 2009 at 3:42 p.m.

    By the by...according to Mary Meeker (Morgan Stanley), the effective CPM of Search in 2008 was $31eCPM. Banners? $1.34.

    Have fun with those numbers, willya?

  15. David Koretz from Adventive, Inc., December 11, 2009 at 4:26 p.m.


    It doesn't *have to* yield a low eCPM. It does because publishers today lack the technology to display the ad that will generate them the highest eCPM.

    Long-term, publishers can't just change to a performance-based model (although they will be forced to), they are also going to have to change their ad management, placement, optimization engines.

    More on that in January.

  16. David Koretz from Adventive, Inc., December 11, 2009 at 4:28 p.m.

    @David Chu, I think your definition is right on. You summarized my column in far fewer words than I could have :)


  17. David Koretz from Adventive, Inc., December 11, 2009 at 4:29 p.m.


    CPC does not have to devalue branding.

    If you sell on a performance-based pricing model, then over time the market will increase the price to account for the branding impact.

    That's the great part about selling on an auction model. The market drives up pricing for you, and all variables are ultimately accounted for.

  18. Paula Lynn from Who Else Unlimited, December 11, 2009 at 7:01 p.m.

    David, you may want to check out an article about auctions in the NYT last week. It's the auctioneers that make out, not so much for the buyers.

  19. Mike Handy from Nostrum Inc, December 11, 2009 at 8:45 p.m.

    Personally I hope CPM doesn't go away totally. That said I think you are correct and the matrix of the future will be a different animal. There are many studies that demonstrate internet advertising doesn't drive click sales. The data suggests that internet advertising has a substantive effect on on in store brand awareness.

  20. Roy Dahl from RoDa, December 12, 2009 at 5:15 a.m.

    For Communicators Rule#1 should be to always tell the readers the meaning of their abbreviations. This can be done in the first sentence where the abbreviation is put. Like "CxxxPyyyyMzzz (CPM) pricing will account for..." Then it will be much easier for foreign readers to understand what you are writing about.

    In My Humble Opinion (IMHO) only.


  21. David Koretz from Adventive, Inc., December 12, 2009 at 7:58 p.m.

    @Roy, good point. I will make a point of this for future articles.

    I was using industry-standard language, but I didn't account for foreign readers.

    For reference:

    CPM is cost per thousand impressions.
    CPC is cost per click
    CTR is click-through rate.

    Thanks for the feedback.


  22. Guillaume Wolf from Alenty, December 14, 2009 at 5:01 a.m.

    Hi David,
    Great post here. I totally agree with you CPM isn't a fair metric. Ads are different and don't have the same impact according to the format used, to hosting page, to the placement and to the efficiency of targeting.

    However, I wouldn't agree on keeping the CPC as the relevant metric to sell on-line advertisement. Like you wrote: "CPC is a metric that starts to measure advertiser success." Publishers shouldn't accept to be paid less because the advertisers aren't attractive enough.

    I would say that a new metric has to be set up. Something that emphasizes the impact of an ad regardless to the advertiser. The duration per impression could be great.
    A French TV channel (Canal +) decided to sell its on-line ads the same way as on TV, with a pricing depending on time of exposure. This great move allowed them to increase impressively their revenues.

    You're right that implementing a new metric is quite impossible. But the duration isn't a new metric. It's just an adaptation from what already exists in other media.

  23. Nelson Yuen from Stereotypical Mid Sized Services Corp., December 14, 2009 at 11 a.m.

    I have to make one more comment - trying to address a couple of points.

    To David Chu on CTR.

    I think thing to take away from CTR is that there is now a measurable step to engagement. If you look at CTR as a business process, at the very least you know if the audience paid attention - as opposed to traditional media where you couldn't identify if an ad was engaging to the end user. So before we couldn't measure the channel surfing in between TV commercials, now we know whether the person is clicking on the ad right?

    So the point I'm trying to make is that we've moved PAST mere benchmarking (as in your example of the Mean Joe Green Coke campaign) and graduated to predictive analytics. At the very least, we're real time no?

    To Dana and Expanding on David's response:

    Can you please, at the very least, gravitate away from making assumptions about what "quality content" is then? Cause, I really don't care if you put your "blood, sweat, and tears" into something that isn't engaging to me. (Get the example?)

    Just because you think your content is worth X amount of dollars doesn't mean it is. And if you can't come up with a better method to measure the worth of your content other than a performance based model then give me a good reason why I should pay for your content.

    I'm not siding with either. But ragging on the author and the performance based trend doesn't solve the issue that (from an Econ standpoint) your content was never worth what you thought it was. (Don't make me draw the inefficiency triangle or opportunity cost matrix from Econ 3001.)

  24. Nelson Yuen from Stereotypical Mid Sized Services Corp., December 14, 2009 at 11:01 a.m.


  25. Fraser E from Opinions expressed herein are solely my own, December 14, 2009 at 5:48 p.m.

    Nelson, "real time" is very over-rated, especially when the data it's giving you is such an inaccurate representation of the ad's effect. We don't base our assessment of our overall personal health on the fact that we're currently having a pollen-induced sneezing fit, and we shouldn't base our assessment of campaign effectiveness on the fact 0.1% of our audience clicked on one ad and 0.11% clicked on the other. That's still 99.89% of our audience who didn't click, but who may have still taken an effect away from the exposure that's positive to our business, which will need to be measured...over time.

  26. Courtland Smith from Adify, December 14, 2009 at 7:56 p.m.

    Agree with Dana. The unwanted externality of going with a CPM pricing model is there is a glut of impressions, in part caused by sites designed to maximize impressions.

    If CPC becomes the dominant pricing model, we'll see ads designed to be more intrusive, more obnoxious, more misleading...just to drive CTR up. We've all already seen ads like this...think of the ad types that float over the content of a page, and when you click the "x" close button, it takes you to the advertiser's site...

    Koretz, I'd be interested in hearing some elaboration on why attempting to find a metric to sell ad space that aligns publisher and advertiser interest is such a bad idea other than "we've tried before, and failed". I've discussed alternatives to CPM pricing a couple months ago, for anyone who's interested:

  27. David Koretz from Adventive, Inc., December 15, 2009 at 12:07 a.m.


    I wasn't saying that we shouldn't explore new *pricing* models, I was talking about how we measure the value of the ads.

    The weird metrics publishers create to measure brand lift and other vague concepts are too hard for advertisers to correlate to revenue.

    Ultimately advertisers are going to pay on the model that is the least risky.

  28. Nelson Yuen from Stereotypical Mid Sized Services Corp., December 15, 2009 at 11:23 a.m.

    Hey Fraser;

    I'm not sure if you and I are talking about the same thing when it comes to "real-time." I was really just being sarcastic.

    But in response to your example - and I'm not necessarily disagreeing or agreeing with you...

    But consider this. In the same way you can't DISREGARD the visibility of an advertising campaign based on "clicks" you can also NOT count the value of the add as relevant to the ENTIRE audience it reached right? So how would we solve OUR example? Maybe a click is worth MORE than we thought based on the probability of attacking certain points along the buying cycle. Maybe we should DEVALUE the reach of the campaign due to the inability to measure relevancy of the ad? Get what I'm trying to say?

    The solution is technology. Cheers.

  29. Mike Einstein from the Brothers Einstein, December 16, 2009 at 10:14 a.m.

    If my CPM for a banner schedule is $2, and my CTR is .1%, my cost per visitor is $2. How much is the impression worth? About $.002, according to the guys setting the price. In other words, it's virtually worthless.

    I work with a company that delivers targeted, qualified visitors to any advertiser destination for only about dime each. No fuss, no muss, no waste.

    You want accountability? Contact me for the details.

  30. Dj Das from, February 10, 2010 at 2:02 p.m.

    Yes, I think that the current CPM model is doomed and this model needs to evolve to better match the needs of the advertiser - those who demand performance metrics along with developing brand awareness (thats the best a CPM campaign can do, I think!)

    In fact, I wrote a blog around the same lines here

  31. Dj Das from, February 10, 2010 at 2:05 p.m.

    Yes, I think that the current CPM model is doomed and this model needs to evolve to better match the needs of the advertiser - those who demand performance metrics along with developing brand awareness (thats the best a CPM campaign can do, I think!)

    In fact, I wrote a blog around the same lines here

Next story loading loading..