1. Auction structure : Waterfall versus Unified auction
For each ad request, mediation technology sends ad requests to Ad networks connected to the SDK, one after the other, until one responds based on the order. The mediation platform will put the ad network that performs the most on the top of the waterfall. In most cases, the eCPM is the main criterion to order ad networks. (waterfall principle). The waterfall is optimized, which means that the ranking changes as often as the theoretical CPMs are updated. If the top ad network is unable to fill the ad request, the mediation platform tries the next ad network until it fills the request. As each ad network is called one after the other, no ad will render until an ad network actually takes the impression.
As a result, this could lead to huge latency and bad user experience. And when one ad network takes the impression, it will stop the waterfall, even if another ad network would have had an higher price to offer.
With in-app bidding technology, every ad network is called simultaneously. He will send back its decision (impression and price) to the adserver. All demand sources have a fair look at each ad requests, as they are all called at the same time. This results to a significant latency reduction.
2. Pricing methodology : Historical CPM versus Real-time CPM
For mediation, the publisher will set its waterfall and the order of its ad networks based on historical CPM. The publisher is responsible for ensuring this theoric CPM is as accurate as possible. For example, he could update them on a daily basis based on previous day data. Unfortunately, the historical CPM is just a hypothetical CPM that doesn’t accurately reflect the real bid they will answer.
In-app bidding demand sources have the opportunity to bid in real-time for each ad request for a given placement and a given user. This real-time mecanism will ensure a fair competition between all demand sources. This usually leads to an increase in revenue. The highest bid wins.
3. Major customer benefit : Fill rate versus higher CPM
The main objective of publishers implementing mediation is to maximize their fill rate by adding many ad networks in their waterfall. The more ad networks they plug in, the more chance they have to get a winning bid and fill their inventory. But unfortunately, mediation requires to call each SSP after the other until one ad network fills the ad. This could lead to multiple calls and increased latency.
Above all, one of the biggest benefits of in-app bidding is the maximization of the publisher’s revenue. With in-app bidding, you directly and simultaneously have the information of who is bidding and at which price. It’s in real-time and all demand sources compete in a 100% fair competition. So, you get higher eCPM and the highest bid wins.
But how does in-app bidding work?
In terms of integration, you will have to integrate several Bidder SDKs, such as Amazon, Facebook… that will call them for each inventory opportunity.
#1– For each inventory opportunity, Bidder SDK will trigger all their demand sources and run a real-time competition.
#2– The SDK bidder will pass the highest price to the primary ad server SDK through an adapter.
#3– The primary ad server SDK makes the full competition between all other channels (Direct, Deals, programmatic guaranteed, Open auction).
#4– The winner bidder SDK will be render the ad with the highest eCPM among in-app Header Bidding, direct campaigns and RTB.
At Smart, our experience with our clients shows that enabling a real-time and on real-price competition between all demand sources increases yield for the publisher. In the end, in-app bidding offers the same benefits of desktop client-side header bidding : a fairer and more transparent auction leading to higher yield, better monetization and improved user experience.