Free network mediation in mobile has become a standard for publishers. But is there really anything “free” in technology? There’s always a cost or some type of value that is given away to a vendor that offers something for free. With free mediation and ad serving, the true value of publisher inventory is never actually realized.
The conflict of interest exists because ad exchanges and ad networks operate on a rev-share business model, taking a cut of publisher revenue when their demand fills an impression. This means they are incentivized to always fill impressions with their demand sources, if possible.
Offering an ad exchange for your partner CPMs provides an inside look on how much others are willing to pay for the inventory. Rather than offer their strongest bid, the mediation platform can win the auction at a discount, only one penny higher than the next highest paying demand source.
As we’ve seen with header bidding on web publishers, allowing all demand sources and ad exchanges to compete with the same footing on the playing field can yield up to 30% to 70% more than the current revenue metrics.
Why Bidding Into Mediation Is Only Temporary
Because mediation layers do not account for demand partners willing to pay a higher price in real time for a given impression, demand platforms not in the primary ad server or mediation position are jockeying for higher spots in the waterfall, instructing publishers to set up hundreds of line items signifying different price points.
Just like header bidding on the web, mobile apps are now evaluating client-side technologies that allow an auction in the app before the ad server runs its final auction. This will allow for more demand sources to compete for the ad impression, forcing the ad server to either fill with the highest paying demand source, or with a bid that’s higher.
Unfortunately, this fix is only temporary and introduces a bigger infrastructure issue for the ad technology industry. This setup creates operational and trafficking overhead, only solving part of the overall problem, yet lacking a holistic vision of demand competition.
Machine Learning Tech For Mobile Apps
A true competition that allows networks and real-time bidders to compete must be done in a real-time unified auction. This is an opportunity for the ad ecosystem to get ad tech right with mobile. A unified auction reduces the latency required for numerous asynchronous auctions and mediation layers.
But some SDK network partners (i.e. Facebook, AdMob, Flurry, etc.) do not expose their bids in real time. That is where machine-learning technology plays a role in ad tech.
With objective predictive technology, a smart eCPM for non-RTB demand partners completes the approach toward a fair, real-time unified auction. If done well, this technology essentially replaces the manual need to update the static waterfall. It allows demand partners to compete in real time and dissolves the overhead required to run hundreds of line items in the ad server.
Applying machine learning to mediation offers the most holistic approach to increasing programmatic revenue. It not only forces every demand partner to bid their true value, but gives everyone a equal chance to win the auction with the same amount of data. For publishers to do this successfully, they must reclaim the ownership and privacy rights to their proprietary data, like sold inventory, fill rate, and CPMs.
Owning Your Data
Imagine you were selling a house, and there was a potential buyer that knew what all the other offers were before the deadline. Because he had this intelligence, he would only offer one penny more than the highest offer to guarantee he got the house. But as the seller, why would you give out that information for free? This is analogous to what your ad server has access to when using their free mediation tools.
Free mediation solutions have saturated the market and have gained momentum because they have access to all of your data, getting an advantage over all of your other demand partners. They know how much everyone else is willing to pay for each impression, so they strategically only bid one cent more to win the auction.
For publishers to own a unified auction, they need to take control back into their own hands and have full ownership over their data.
Publishers need an unbiased system of record. Having exclusive access to your own data allows you to run a real auction in which every demand partner offers their best and highest bid in real time, ultimately creating a level economy for your inventory.
Ad servers that do not take a cut of your revenue, but partner with you and offer the best machine-learning technology for your business, offer a true unified auction between all demand sources and gives mobile app publishers the upper hand to drive more revenue in real time.