The words used to describe private exchanges are still unnecessarily complicated.
It’s time to put away the overused analogies and misunderstood metaphors. The digital ad business is not a “community of living organisms living in conjunction with non-living components,” so we can drop the term “ecosystem” and other words we never needed, like “supply side” and “demand side.”
We get it.
Publishers sell ads targeted to specific audience segments, and the price they get should go up when more media buyers demand to buy those ads.
We get it.
Giving ad buyers a chance to review the targeting data of every ad impression so they can choose to bid or pass is an unbelievable innovation.
We get it.
Executing the sale of an ad to the highest bidder in real time, based on data associated with that particular impression, is stunningly complex.
A publisher who does not participate in a private exchange can appear out of touch, perhaps creating perception problems at the agencies they want to do more direct business with. For that reason alone, having their own private exchange — or, being part of a network of sites that allow publishers to select which advertisers they offer this kind of access — seems like a good idea.
What gets lost in this current complicated technology pitch to publishers is a core question: What is the likely business outcome if a publisher with its own sales team also sells ads in a private exchange?
The “Lumascapees” want publishers to believe a competitive bidding process to buy ads will ultimately increase rates. They like to tout this growth by comparing the increase in private-exchange CPMs versus open-exchange CPMs.
I would like you guys to stop doing that. You are better than that. There are more obvious and relevant comparisons that can be made to see if private exchange CPMs rise or fall year over year for publishers.
Regardless, private exchanges as they work now are not yielding higher CPMs, and they likely never will. That’s because ads bought through an exchange (of any kind) are 100% audience-driven. The value of the content where the ad appears is not considered, so value is solely determined by performance, which gets defined by the buyer. Performance expectations will always increase plan over plan, which means prices will always be forced down, especially when there is an oversupply of ad impressions for sale. Traditional publishers are not search/Google. They don’t have the term “auto insurance” to auction off.
Prices will also always drop in an exchange because they are correlated to fill rates — a.k.a., how much money the publisher actually earns. Publishers learn quickly that if they lower their floor bid price, the checks they get from private exchange advertisers get slightly bigger. More money tied to pressing a button to lower prices will always result in lower prices.
So what business outcome can publishers expect when they open or join a private exchange? A lower overall eCPM for all impressions sold (direct and programmatic combined).
Someone who knows this space better than most (and certainly me), opened my eyes when he recently said that the ad impressions a publisher sells via a private exchange should be just another line item in an overall ad buy.
That got me thinking about how selling impressions via a private exchange can work more profitably for premium publishers:
1. Only allow advertisers who buy ads direct to include ad impressions from your private exchange.
2. Instead of charging a bidded or guaranteed CPM for these private exchange ad impressions, charge nothing.
3. Then demand advertisers pay a 20% CPM premium for direct ads bought, and a 20% increase in the overall budget commitment, in order to earn this new form of “added value.”
Somewhere along the way in this evolving media business, “added value” got a bad rap. In my experience, taking the most valuable item on the media plan and making it free has always helped drive a bigger overall dollar commitment and higher CPMs for the other ads sold as part of the buy.
Using private exchange ad impressions as added value is just crazy enough to make this whole thing work.