Never Buy On CPM Alone

About a month ago, Shelby Bonnie, the CEO of Whiskey Media, articulated a fantastic argument for "killing  the CPM" on TechCrunch. I couldn't agree more with Bonnie' s ideas about discontinuing the use of CPM as the primary metric for online advertising, as I have written in my previous posts "Online Branding And The Definition Of Insanity" and "Engagements Will Be The Internet's 30-Second Spot." The next logical step in the conversation is: If not CPM, then what?


Online publishers get short-changed for delivering impressions to brands that can impact people's perceptions and purchase behavior, but do not result in an action. Advertisers get short-changed if publishers deliver impressions that don't impact brand perception or purchase decision making, which, in a world of LIMITLESS impressions, will be more often the case than not.



I think there's a balance between the two arguments that will allow advertisers and publishers to exchange value fairly. I think that balance comes from looking at the value of impressions as an output, rather than an input. If advertisers paid for engagement, it would be easier to factor in the value of impressions based on the rate of engagement. For example: If Web site A delivered an engagement or click-on rate  for 1% of impressions, and Web site B delivered an engagement on 2% of impressions, given the same creative execution on both sites (this is very important), then Web site B is delivering higher value impressions.

If a market existed where advertisers could see this, they should be willing to pay more for engagements on Web site B. Why? Because the advertiser can assume that those impression are attracting a greater share of attention (or the audience is a better fit for their offering). In effect, the value of the publishers' impressions will be priced into the rate they can charge per engagement/click. (For more on my definition of engage, read "I'm Sold On Engagement.")

It's still very possible, though, that in the short run, publishers will not get full value for impressions -- because advertising creative does not get people to click/engage, while still delivering value to the advertiser -- but this is less of a problem than one might think. Especially when you consider that falling CPM prices across the industry are probably already hurting online publishers, even those that could deliver quality consumer attention to advertisers. The key would be for publishers to innovate on the delivery of quality attention to advertiser messages, while an ad unit that makes sense for consumers in an online experience would be created. Add to this  the idea that engagements add more value than simply "message delivery" -- which is all impressions can do, be they online or on television -- and advertisers and agencies will be motivated to create advertising that induces engagement, so that they can drive conversations with consumers and social media pass-along (more impressions, but this time peer-to-peer impressions).

This is by no means a solution, but a potential step in the right direction. Publisher who can say they drive higher engagement rates should get higher CPMs, and advertisers who get more people to engage with and share their content as a percentage of the impressions they received in a campaign will likely make a bigger impact on sales and branding objectives. For now, it seems that since there isn't a standard, there's a lot of opportunity for publishers and advertisers who are ahead of the curve on setting objectives that matter for major impact in the online and social media world.

Thoughts? Feelings?  Then drop me a line below and watch as the conversation plays out on Twitter:

12 comments about "Never Buy On CPM Alone".
Check to receive email when comments are posted.
  1. Richard Monihan, October 27, 2009 at 12:59 p.m.

    I agree with the premise, but not necessarily the solution. There are 2 flaws to clicks as a form of "engagement".

    First is the question of long term value - a click is a momentary action locked in time. If a publisher is only valued based on the clicks it can provide, then they are only momentarily valuable and any long term value they've provided (a consumer's visualization that if this site approves the content, it may be good, therefore worth revisiting later) is lost. This arrangement is best depicted by an auto repair shop advertising on an auto site. Clicks start out high, then taper off as the usual visitors see the ad, but know they already visit that auto repair shop and like their quality service. Over time, the auto repair shop loses interest in the auto site, ends its relationship due to "lack of engagement", and visitors to the site may think "hmm....I wonder what happened?". Sure, most will stay, but if ANOTHER auto repair shop may be lost.

    A second problem is in the ad itself. I have had advertisers put up HORRIBLE ads and complain about the click rate. Is it my fault, as a publisher, that someone doesn't know how to get their message across? Should I be penalized for their flaws? Or should I take the leap and tell them "your baby is ugly" and get them to change it? Or perhaps they will listen to my advice to rotate and refresh their ads, that the one banner they are running is old and tired?

    Sadly, the second problem is the biggest one. I cannot call an ad agency and tell them how to do their business better. All I can do is offer contextual or content placements to hopefully "offset" or "improve" their positioning and engagement due to their own poor judgement. To say this is rare would be misguided. In my current role, I have not seen it much, but at previous publishing outlets I have seen this behavior frequently and scratched my head, realizing there is nothing I can do about it.

    Unfortunately, for now, the CPM is the one gauge we must live by. It is a standard that allows us to measure visibility via impressions, but it can be modified to capture SOME LEVEL of engagement. What it WILL ALSO capture is also SOME LEVEL of long term brand value on both the part of advertiser AND publisher.

  2. Mark McLaughlin, October 27, 2009 at 12:59 p.m.

    When a consumer engages inside an ad unit, it is a win for the consumer (the ad is relevant and interesting), the publisher (they make money and the consumer does not leave their site), and the advertiser (who delivers a powerful combination of branded content and engagement). So, great ads create value for everyone and lousy ads are bad for everyone. If you make "cost per engagement" part of the fee structure, you penalize great ads with higher prices and reward lousy ads that don't generate any interactivity. Our biggest challenge to growing display advertising revenue and commanding premium CPMs is mediocre and downright horrible creative. Charging for engagements is the kind of self-destructive idea that plagues the online advertising industry.

  3. Joelle Kaufman from BloomReach, October 27, 2009 at 1:08 p.m.

    Relying on clicks to measure anything is akin to measuring the effectiveness of the ads on NYC Taxicabs by the number of people who jump in and ask the cabbie to take them to the store advertised on the roof. Ludicrous.

    The better the content, the more time consumers spend with that content and the better the advertising has to be to get any attention. In fact, last week's news was about the Dynamic Logic study of almost 200,000 online ads showing that compelling and contextual creative is more predictive of performance than targeting of any sort.

    The metrics we use to price and deliver advertising online are still very immature. CPM is a proxy for time - and a poor one. The industry doesn't yet have a good one. Clicks don't estimate much of anything well.

    Measuring brand lift and long term (say, three months) purchase lift is a meaningful measure of advertising effectiveness. Brand lift can be used for real-time optimization. We've commented extensively on Shelby Bonnie's editorial at

  4. Paul Debraccio from Interevco, October 27, 2009 at 1:10 p.m.

    Nice thoughts and nice ideas but nothing new here. This conversation has been going on since the 90s and no one seems to have found a workable alternative. Perhaps an industry wide Task Force can do the research and come up with a real solution. The ARF and the IAB maybe?

  5. David Shor from Prove, October 27, 2009 at 1:23 p.m.

    Post-impressions visits + clicks = engagement. There's plenty of technology around that lets companies know how effective their ads are at triggering interest in the brand to visit a website.

    The conundrum is that traditional brand advertising combined with ad blindness is a losing game. You are almost forced to buy interstitials or page takeovers if you're just trying to put impressions out there. Good digital advertising should at least ask users to take an action of some sort.

    Unfortunately, CPMs have not dropped appropriately to reflect ad blindness. On major sites they're sustained by media buyers who unfortunately accept vendor quoted prices rather than push for 50% reductions.

    As an industry we're a LONG way off from proper creative executions that are tracked properly and which have been bought at the right price to meet the analytics objectives. The sad thing is, the performance marketing folks (my firm is one that straddles both sides of the fence) are the main folks who get the whole picture, and yet most CPA/performance deals are for scam-type companies and offers. Quillion is attempting to change this by bringing major brands into a performance environment and using all of our media attribution techniques to place great buys and optimize quickly.

  6. R.J. Lewis from e-Healthcare Solutions, LLC, October 27, 2009 at 1:27 p.m.

    Joe, What about the flip side of the coin? Advertisers (manufacturers and agencies) who provide more engaging offers and creatives that are more relevant to a publishers audience should get lower rates? They are adding value to the equation - not peddling junk.

    The problem with the media model is one of blurring lines of responsibility. "Engagement" is only a part audience. Much bigger factors impacting engagement include the quality of the offer and the product, the creativity of the ad units and message, etc... who "owns" engagement?

    If publishers charged more for engagement advertisers would circulated non-engaging heavily branded messages to get "free" branding (look no further than the CPA market).

    I disagree that we are entering "...a world of limitless impressions...". With continued audience fragmentation, the quality of the audience matters now more than ever. It's harder and harder to reach the RIGHT audience. Rates on reaching the right audience are going UP not down. The Wanamaker "50%" of waste is slowly being eliminated, and the savings are going in both directions, some to the client in the way of reduced spend and some to the publishers than can reach the RIGHT 50% in the way of increased rates.

  7. Paula Lynn from Who Else Unlimited, October 27, 2009 at 2:01 p.m.

    The discussion has been going on since the '90's, the 1890's, that is. 1 + 1 = 3. Share x HUT = rating point is real. Instinct and common sense has not even been touched here. Why does a great creative spot fail to deliver when a mediocre one crams the doorstep? Awards do not = profit. There is no one magic pill, not a one take panacea for success. Do the same thing next month and the water ebbs and flows as desired. This month was a drought. You can love an ugly baby.

  8. William Nann, October 27, 2009 at 2:08 p.m.

    Rich media = front end engagement metrics & reporting. Front end engagement metrics & reporting + post click analytic data = media buyer nirvana!

  9. Thomas Kurz from EFP, October 27, 2009 at 2:11 p.m.

    I continue to disagree with underlying premise to the argument that “killing the CPM” is a good idea for online publishers and advertisers. As long as mankind has been selling things, marketing has always been about exposure and generating awareness and interest in a product and or service. I find this notion that measuring exposure should be devalued and completely replaced by some metric that measures how people engage with a campaign to be problematic on several fronts. For one thing, I do not believe that most people, as a rule of thumb, are genuinely interested in engaging with advertisers when they are in the process of consuming content, whether it is online or not. Valuing engagement as suggested here requires a change in the mindset of how people consume content. Also, publishers are content producers and not ad agencies. This idea that publishers should innovate on messaging is putting the onus on the wrong party.

    I understand that there are people out there that are trying to game the system. But just as publishers have wrongly inflated traffic/pageview statistics, there will also be ways to lift engagement stats. And in the end these engagement numbers will still fail to properly recognize the value of exposure that is generated when one matches excellent creative in a timely manner with content destinations whose audiences are most likely to be receptive to nonintrusive messaging.

    I am not suggesting that there isn’t some value to be assigned to engagement advertising. But I am suggesting that captivating messaging and well placed exposure cannot be properly valued without having some measurement that accounts for exposure.

  10. Kim Lloyd from Bright Hub, October 28, 2009 at 9:33 a.m.

    I couldn't agree with Thomas Kurz more.

    If advertisers continue to overuse networks to save cost and expand reach, then how can a single publisher possibly be accountable for engagement?

    The advertising world would be a perfect place if every consumer acted in a linear fashion on a single ad exposure.

    Please let me know when this world of consumer robots takes hold.

  11. Joe Marchese, October 28, 2009 at 1:11 p.m.

    Lots of great thoughts everyone.

    @Richard Diminishing engagement over time is a potential issue, but that should be a constant across multiple sites. As to the second point, I thought I addressed that in the column pretty clearly. Poor creative will perform poorly on multiple sites, if your site performs comparatively better, then you are providing better impressions, or audience.

    @Mark like charging for clicks penalizes advertisers for writing great copy and contextual adwords ads? If the engagement is valuable, you should want to pay for it. Publishers should optimize to ads that perform for them and drive revenue.

    @joelle I am all for measuring brand lift. Getting to a standard and a way to identify the unit responsible.

    @thomas I agree that impressions are valuable. I am just looking at a way to signal to advertisers "how valuable" the impressions are. I think publishers should get paid for every impression that delivers a message for a brand, there just has to be a way to differentiate impressions. I have an idea that might make this clear. I am going to flesh it out and publish it next week in the column.

  12. John Dietz, October 28, 2009 at 7:17 p.m.

    Wow, I love the NYC Taxicab example from Joelle in the comments.

    I don't think the industry is going to move away from impressions or CPM anytime soon, there just isn't a viable alternative, and clicks or engagement still miss the majority of the impact of branding display ads. The real problem is that currently all ad impressions are treated equally even though we have the capability to look at them in more detail.

    Right now all impressions are counted equally, whether they show up above the fold or below the fold, whether they are visible for 5 seconds or 30 seconds, whether they are with relevant contexts or contrary contexts, etc. There's a lot more we can do to look at the qualitative aspects of those ad impressions to boil them down to marketing objectives of exposing the right kinds of users to the message in the right context.

Next story loading loading..