Publishers, Let’s Talk Programmatic CPMs

joannaoconnelrevised“Marketer’s Note” is a weekly column informing marketers about the rapidly evolving, digital marketing technology ecosystem. It is written by Joanna O’Connell, Director of Research, AdExchanger Research.  

I just had a very interesting conversation with a reporter who’s trying to understand why so many publishers link “programmatic” to declines in ad revenue and/or profitability in their quarterly earnings statements. I have strong suspicions on what might be happening, but I don’t yet have a firm answer on what’s really happening, at least not yet. 

Certainly, publishers will indicate that the volume of impressions they sell programmatically outstrips the revenue associated with those impressions (at least that’s what I found in my State of Programmatic Media report). But that isn’t a surprising statistic, nor is it a new phenomenon.  Certainly, this was the case when publishers relied on ad networks for indirect inventory sales.

So why is programmatic seen as the culprit for declining ad dollars? My suspicion is that many, many factors are at play.  Here are a few:

–       Ad networks sent nice, fat checks every month, without fail.  Publishers achieved consistent monthly revenue through those relationships, even if they weren’t optimized, or well understood, by some publishers.  In the programmatic world, there isn’t yet a clear replacement to the networks – i.e. a single entity that consistently buys up gobs of inventory, month after month.   I do anticipate, over time, this may change – with large advertisers like P&G investing mammoth sums of money in programmatic and agency trading desks increasingly striking holding company-wide deals the flow of consistent money will continue to grow.

–       Media buyers were happy to buy a lot of ROS, but aren’t so much anymore. The upside for buyers of audience targeting and impression level buying – hallmarks of programmatic – is control and choice.  But publishers don’t realize the same benefits, or at least in the same way.  Today, programmatic can mean a lot of (sometimes unintentionally) small buys for very specific impressions – laborious to sell and manage from a human capital standpoint (certainly more, I’d imagine, than days of yesteryear where publishers could contract for giant IO’s of ROS impressions). I believe, though, as standards continue to be developed and global yield management tools continue to be developed and are adopted, the human burden of managing many, many complex programs will start to lessen, and those programs – as a whole – will add up to strong overall revenue for publishers.

–       Programmatic necessitates new skills, which both publishers and buyers are still developing. We’re all still new at this, in spite of the fact that “programmatic,” in one form or another, has been around for several years.  Case in point: we’d finally started to figure out exchange-based buying and selling and along came programmatic direct and all its cousins – a whole new set of programmatic transaction types was born and we’re all trying to figure out how best to make it work*.  We’re still in a period of massive disruption – buyers and sellers are being asked to behave in totally new ways, and making attitudinal and behavioral transitions takes time for people.   So of course things are messy and suboptimal.

BUT, these are my suspicions only, as my experiences with publishers on this subject are many but anecdotal.  And I am sure this is not a complete list of culprits – just some starting thoughts. I’d like to hear from you, publishers, directly or via comments on the site – what is YOUR market reality with respect to programmatic and CPMs?


* I think we should acknowledge that RMX had the 1:1 relationship model very early – it just happened that the two parties were often networks.

Follow Joanna O’Connell (@joannaoconnell ) and AdExchanger (@adexchanger) on Twitter. 

Enjoying this content?

Sign up to be an AdExchanger Member today and get unlimited access to articles like this, plus proprietary data and research, conference discounts, on-demand access to event content, and more!

Join Today!


  1. Patrick Landi

    You hit the nail right on the head with point #2. As ROS shifts to programmatic we need to find a better way to scale our buys. As you said, we’re getting a lot of small buys for very specific impressions, which doesn’t make up for what’s lost in traditional ROS buys. But I think/hope our tech platform partners will catch up with scalable programmatic guaranteed solutions to re-coop that revenue.

  2. Jeremy

    Most advertisers don’t know how to measure ad effectiveness. A “Mom” who has seen 150 impressions in the last 30 minutes is valued the same as a “Mom” who just logged in and saw her first ad of the day. The latter is worth much more because she is more likely to be influenced by the ad unit but advertisers have forgotten this.

    Furthermore, most performance campaigns continue to reward cookie bombing approaches with discredited attribution models like “last touch”. These models drag down the CPMs by rewarding the lowest common denominator.

    • Joanna

      I agree, and it’s super unfortunate, as it ultimately benefits no one but the media seller, most likely an intermediary – it’s not good for consumers, marketers or publishers to participate is what is ultimately a “gamed” system. I spend a lot of time preaching about things like multi touch attribution and frequency management – and I have to say, in general the marketers I talk to know what I am talking about and aspire to get there, even if they’re not yet. Maybe I’m naive but I think it’s only a matter of time before those kinds of practices – better measurement, more effective media management as a result – are the norm.

  3. JIm Spanfeller

    Two thoughts here. First ad networks were and are bad for premium publishers. They have gotten a tad better as they have been held to higher standards by both the demand and supply side but their core business model is around arbitrage and that is of course at odds with either end of their supply chain.

    Second and more to your question, the biggest issue with programmatic at this point is transparency. Or rather lack thereof…There are a lot of layers to this thought and one of the most important is that with a lack of transparency (seller not really knowing what they are buying) many bad things can and do happen. Bots, pixalated ads, ad collisions and in general non viewable inventory. The standard response to these issues from the buying community is that these issues are “priced into” the buys. This serves to substantially lower the cost-per basis of programmatic buys as the folks with good inventory have to basically pay for the black hat bad inventory scams. And as a result harms the overall ecosystem. Programmatic can and should grow into a very viable part of the ad stack…but as there is a distorted perception of what is what it due to the overwhelming lack of transparency it will be hampered. This is one of the reason why private exchanges have become a growing option.

    • Joanna

      It’s a really interesting point, Jim, worth much more discussion. Transparency means so many things: site/placement transparency, tech/partner fee transparency, price transparency, bid landscape transparency, etc. etc., each of which could take a column in and of itself. And I like your point about good pubs essentially “paying for” the bad styff – I’ve heard that a few times recently – good pubs suffer because bad actors pull down prices and buyers can’t differentiate the good from the bad – it’s no wonder publishers are looking to create direct deals. I am not entirely ready to give up on the open exchange model – I just like the idea of it and always have. But certainly we can all see the writing on the wall – both big buyers and premium sellers are pursuing direct programmatic deals in their many forms.

  4. Programmatic isn’t the culprit, after all the word programmatic is quite broad to summarise as ‘the thing’ that caused erosion to CPM rates.

    I am assuming you are referring to Real Time Bidding in the Exchange, if yes then the reason is simple economics really, the market has overwhelming supply of inventory over demand – so things naturally get cheaper.

    Since we live in a dynamic world, buyers traditionally bought inventory directly with publishers are shifting more spend budget to Real Time Bidding. First is cheap, secondly the process is fast and efficient, thirdly technology allows buyers to overlay data to relevant audience, on top of this buyers have the luxury of real time reporting and optimization options… the list goes on… in the end is better control, less wastage and improved ROI (if done correctly) for buyers.

    Publishers should learn from diamond industry if they are trying to revert back to the good old days with high CPMs. Scarcity and quality diamonds will get top $$$.

    My 2 cents!

    • Joanna

      I guess what I am looking for is real proof one way or another, Al! I hear programmatic getting blamed – as I mentioned, for example, in earnings calls, and it strikes me as really problematic to assign so much blame there.

  5. If publishes allowed advertisers/ trading desks to buy high impact ad formats they would see an increase in their revenue. There wouldn’t be a need for as many ads on a page, all the advertising would be 100% visible and there would be less wastage. I believe trading desks understand that driving down the cost of an impression has had a negative in at least two ways.

    Publishers have been forced to put more ads on their site to try and overcome the falling CPM’s. Users are seeing more and more ads and are therefore becoming even more “banner blind” .

    Why not allow Trading Desks to utilise the data that is available out there, bid on the user that matched your profile, and show them an ad that they want to interact with. Surely this will make everybody a little happier? The publisher receives higher CPM’s, the advertiser receives higher CTR’s and the users experience is less cluttered?

    • Joanna

      Right, definitely, I hear you Paul. And I think those things are starting to happen – but it’s new and not at scale at the moment. So partly I think this is a time horizon thing – it will take time for things to normalize to a place everyone can live with. It might not mean pubs get the same CPMs for ROS programs they used to, but maybe they can sell those impressions more efficiently and concentrate on the bigger, custom programs that command the really high dollars and necessitate the really expensive sales people!

  6. Audience targeted inventory is more valuable than ROS and should be priced accordingly.

    Within the programmatic marketplace, buyers are cherry picking their target audiences through the DSP without transparency to the seller. They are able to purchase premium inventory at significantly discounted rates, which are comparable to the CPMs of the Bulk ROS Network.

    Publishers need to understand their inventory, and adjust their floor rates based on the value of the audience they are selling, even if this results in a lower fill rate through the exchange. The alternative will result in the “race to the bottom”.

  7. John K

    Nobody likes to address the real issue in these forums and that is Marketers have pinched agencies so bad on pricing that they have been forced into this race to the bottom scenarios. All anyone ever wants to do is blame programmatic and ad networks for falling CPM’s. Let’s start at the brand level where the true race to the bottom starts

    • Joanna

      John K – if it makes you feel any better, I do in fact talk to marketers about this very issue all the time. Cheapness and efficiency are not synonymous – squeezing your partners for margin, or forcing them to negotiate lower CPMs does NOT necessarily translate into more value extracted. But this would mean a huge paradigm shift in how the whole ecosystem behaves – I believe it’s possible, at least to some degree.

  8. Anonymous

    I agree that an attitudinal shift is necessary before programmatic grows beyond the industry hype. I’ve heard from buyers that they believe programmatic – premium priority private programmatic (not open auction) – should be only slightly more expensive than remnant which says two things to me: 1) that buyers still wrongly equate all programmatic with RTB/Open Auction/Remnant and 2) that the idea of paying a premium to get a very specific user at a very specific time and place is still secondary to getting that impression at the lowest cost possible. If it is a case of publisher inventory being overpriced as it is, then that’s an industry wide issue but if my sales team is selling the same impression directly with ease, then where is my incentive to offer programmatic buyers that same access but at a fraction of the price? I understand if I’m prioritizing my programmatic access along with remnant but if you truly try to build a PMP and all you hear from buyers at events is they are looking for PMP Deals at remnant prices, then I have yet to see the value to publishers in building premium programmatic.

    As for the transparency issue, I agree it is a huge drag on the industry and immeasurably harmful to premium, transparent pubs. Why has the industry not held the vendors to task on cleaning up this issue once and for all? Most of the middlemen here have done very well financially while seemingly doing nothing about cleaning up their exchanges except well-placed PR spots. When I worked for a network, I was able to verify every URL/partner with inventory we monetized. I know it’s not that simple at scale but personally think the exchange providers have failed miserably in their efforts to date. In fact, there is some evidence that the exchanges have made things less transparent to pubs while demanding more transparency from them in order to combat an issue they are largely ignoring – except, that is, for the hollow sound bites and shameless plugs about inventory quality at conferences and places like Adexchanger

  9. Chip Schenck


    Excellent observations and some interesting conclusions. In point #1, you talk about the fact that ATD’s are trying to strike agency wide deals, which looks and sounds like ‘ad network 2.0’. Given that what we have seen from the early initiatives of ATD’s and their move from many:1 deals to more 1:1 deals (which have many more mutual benefits to client and publisher), this regression to large many:one or many:many deals leaves publishers and clients open to the same ad net problems of lack of transparency, value imbalance and commoditization, where the benefit of the deal goes to the middleman, not the buyer and the seller. This may be one of the reasons (in addition to relevance, in-house skills development, etc) why we are seeing many more digital agencies taking on the task of planning and buying programmatic themselves, and clients doing many more DSP-direct deals. There is so much more value an ATD can provide its network rather than bulk deals and appearing to be a 2.0 ad net – would hate to see that value lost at such a critical stage of the industry development.

    • Joanna

      Love these thoughts, Chip – thank you so much for sharing them!