WTF is bid stacking?

WTF is bid stacking?
Bid stacking — also described as “multi-bidding” — is a technique used by an exchange to inflate its chance of winning a bid for an impression by placing multiple bids on the same piece of inventory. That makes its chances of winning that impression higher and so increases its match rates. And the better the match rates, the better the payout for the vendor, because they charge a publisher a fee for every impression won.

So what problems does that create?
Exchanges engaging in bid stacking are artificially inflating demand to drive prices up in the auction as demand-side platforms could be bidding multiple times on inventory without realizing, said a publishing executive who asked to remain anonymous. That can mean that demand duplicates numerous times, but the supply amount stays the same. It’s an example of auction gaming to increase the chances of more money flowing through respective exchanges, according to sources.

How common is this?
Like many things ad tech, this isn’t clear. The opacity of ad auctions means buyers and sellers don’t see log-level data that would show the transactional data. “It’s always been an issue, but buyers and publishers have become more educated over the years and are now beginning to ask questions of vendors as we want a fair marketplace,” said a publishing executive.

Is it an issue for the buy side?
Potentially yes, if it has inflated the true cost of the inventory. If a bid has been multiplied, it’s essentially harder to buy so the buyer may increase their rates to ensure the win the auction.

How can publishers get visibility on when it’s happening?
The best way to get visibility is through log-level data, (also referred to as log-file data) which means data from the bid stream. Publishers can request to see log-level data. So far, the number of vendors willing to hand it over are few and far between though. Some publishers worry that agency partners believe they are the ones pulling strings, or know what’s happening in the auction when they don’t, so it creates trust issues.

OK, so why is it important to know about this?
It’s another example of the auction-dynamic bias that inflates a vendor’s revenue in a way that’s unfair to its competitors, and that overcharges the publisher.