Before getting into the more complex stuff, let’s do a quick definition of Real-Time Bidding, or RTB – surprisingly, it is often confused with “programmatic”, however those two are not really interchangeable. “Programmatic” literally means “automatic process”, so it can qualify any automated media buying process. However, RTB happens (you guessed it) in real time, which is not the case for all programmatic solutions. RTB is an auction that happens while a page loads, usually somewhere between 200 and 500 milliseconds – to give you a frame of reference, the blink of an eye takes 400 milliseconds.
So, as the page loads, the publisher makes the impression available through an ad exchange for the buyers to bid on. Through their DSP (Demand-Side Platform), advertisers will evaluate what they know about the visiting user and the context to be able to assess the bid, depending on how relevant the user and context are to their ad. Once all bids are submitted, the highest bid wins and their ad is served.
Now those auctions can happen two ways: first or second price. Let’s take an example to illustrate this: say we have advertisers A, B and C entering the auction. A will bid 1.25$, B will bid 1.50$ and C will bid 3.00$
Here is what would happen in a first price auction: advertiser C would be the winner, as they bid the highest, and the auction is cleared at 3.00$. Simple.
However, advertiser C still paid twice the price that advertiser B was willing to pay.
In a second price auction, the winning bidder is who bid the highest, but will pay the equivalent of Advertiser B’s bid plus a small fee, usually one cent. In this case, advertiser C would still be the winner, but pay 1.50$ (Advertiser B’s bid) + 0.01$. So, in a second price auction, the same auction would be cleared at 1.51$.
However, with the advent of header bidding, the concept of second price auction is being challenged: by duplicating the same auction across multiple ad exchanges, DSPs integrated in multiple exchanges risk outbidding themselves, and therefore raising their costs. The various ways exchanges manage auctions and the potential lack of transparency also cause an uneven playing field for many buyers, hence the recent moves to first price auctions from various actors of the industry.
With RTB, and especially with second price auctions, publishers have become worried they are not getting their full revenue potential. That’s why publishers can set floors for their inventory, by deciding on the minimum price an advertiser will need to pay to win the ad space.
Hard floor vs soft floor
Again, there are two ways publishers can set that minimum price: hard or soft floor.
Simply put, a hard floor is when a publisher sets a minimum price for its inventory: this means all bids that are below that price will be discarded, even if they are one cent under. This enables publishers to get their inventory’s worth in revenue; the downside, however, is that if all bids are lower than the hard floor, the space will not be sold.
On some platforms, publisher can also set a soft floor, which is threshold for otherwise missed opportunities to monetize their ad inventory and is always set higher than the hard floor.
When both hard and soft floor are in place, one of three things will happen – let’s take our previous example to illustrate this:
In the end, through our exchange, district m supports both approaches through transparent unified auctions because they serve different purposes. First price auctions make sense when a buyer is really serious about winning a specific impression for a specific audience.
By knowing the fair market value of that impression through hard floor mechanisms, they can take all into account to bid the amount they are really willing to pay. Other buyers may want to reach a broader audience and minimize their costs by only bidding through select supply partners with full transparency about floors and fees, maximizing their ROI because of the savings second price auction often bring.