In Digiday’s most recent addition to its “WTF is [fill in the blank]?” series, the issue of waterfalling was tackled.
It was described as just one solution among the “dark art of yield optimization,” and was also coined as “one of the most dependable.”
As dependable as it might be, it’s definitely sub-optimal.
Waterfalling is the act of offering your impressions to one demand partner at a time in sequential auctions.
Waterfalling quickly becomes a vicious cycle. When you give one demand source first look at your best impressions, of course they’re going to bid highest and take the best ones and pass on the rest. When the next demand partner is given access only to those unbought by the first, their average bid and fill will almost always be way lower than the first position partner. And the performance gets worse and worse for each subsequent partner in the waterfall. Waterfalling is a terrible way to measure the effectiveness of demand partners because of the take-turn method of the bidding.
Wouldn’t it be great if you could get those demand players in the same room at the same time bidding on the same impressions? You know, like auctions work in real life?
There are a few methods available to achieve this outcome now, but publishers will only see actual max yield for their impressions when there is true competition for each and every impression.
Waterfalling, as dependable as it is, will never achieve that. They should stick to nature.
Read More: WTF is Waterfalling? (Digiday)