ROXOT BLOG
A Publisher's Revenue Lost
to the Second-Price Auction Game
to the Second-Price Auction Game
The shortened version of the article published on adexchanger.com can be found here
Currently, publishers are discussing emerging S2S solutions, latency issues with header bidding, and other hot topics, but ignoring the biggest issues with website revenue: Why are publishers still relying on second-price auctions to sell online advertising? As a publisher, you owe it to your bottom line to reexamine this strategy.
Why an auction?
On the surface, there are two main reasons why the auction method is so ubiquitous. First and most simply is speed. The entire process must be executed as close to real time as possible. An ad should be delivered to a browser instantaneously so as not to interfere with the user experience. Secondly, the process must be as simple as possible. All market participants need to be clear on what's happening on both ends. If the process isn't clear, participants are likely to choose strategies that lose money and see diminished returns. It won't be long before they are disappointed and abandon the method. An auction's clarity and transparency are indicators that the process is fair to all its participants.
On the surface, there are two main reasons why the auction method is so ubiquitous. First and most simply is speed. The entire process must be executed as close to real time as possible. An ad should be delivered to a browser instantaneously so as not to interfere with the user experience. Secondly, the process must be as simple as possible. All market participants need to be clear on what's happening on both ends. If the process isn't clear, participants are likely to choose strategies that lose money and see diminished returns. It won't be long before they are disappointed and abandon the method. An auction's clarity and transparency are indicators that the process is fair to all its participants.
Why do most automatic auctions use the second-price model?
Traditionally, there are four types of auctions used for allocating a single item: open ascending price auction, open descending price auction, first-price auction, and second-price auction. In an open ascending auction, participants openly bid against one another, increasing the bid. In open descending, the price is lowered until a bidder is willing to pay the price. Neither of them fit the first requirement for an RTB auction - they are too slow and are controlled manually. Therefore, we are left with two alternatives: a sealed first-price auction or a sealed second-price auction. When a first-price auction is used, bidders simultaneously submit sealed, secret bids. The highest bidder pays the price they submitted. Second-price auctions are similar, except that the winning bidder pays the second-highest bid rather than his or her own bid (second-highest bid plus $0.01 in RTB environment).
Traditionally, there are four types of auctions used for allocating a single item: open ascending price auction, open descending price auction, first-price auction, and second-price auction. In an open ascending auction, participants openly bid against one another, increasing the bid. In open descending, the price is lowered until a bidder is willing to pay the price. Neither of them fit the first requirement for an RTB auction - they are too slow and are controlled manually. Therefore, we are left with two alternatives: a sealed first-price auction or a sealed second-price auction. When a first-price auction is used, bidders simultaneously submit sealed, secret bids. The highest bidder pays the price they submitted. Second-price auctions are similar, except that the winning bidder pays the second-highest bid rather than his or her own bid (second-highest bid plus $0.01 in RTB environment).
Unfortunately, auctions inherently encourage low bids. When you pay the price you bid, there's an inherent desire to bid low and save money while keeping the knowledge of what you'd actually be willing to pay secret. Even in a sealed auction environment, not knowing the other bids encourages one to be conservative; consider what others might bid, make a slightly higher bid and hopefully win out at a discount in the end.
The second-price auction, however, does uncover each bidder's real value for the item auctioned. Because the winning bidder pays the second highest bid, there's no real advantage in staying conservative. Each participant bids precisely what they value the item to be worth, while the winner still gets a discount. This advantage gave the second-price method huge favorability in paid search and became the natural go-to for RTB.
Secondly, the second-price auction's idea of truthful bidding heavily relies on the bidders' estimations of their private values - the maximum they want to pay for an auctioned item. But what if advertiser's information about their private values (max CPMs) is wrong or not exactly honest? Marketers' calculations of max bids are usually based on their estimation of Customer Acquisition Costs (CAC) which relies on their specific attribution model. Building such a model is a sophisticated task even the most savvy advertisers rack their brains over. Consequently, many advertisers use significantly simplified and inaccurate attribution models such as first- or last-click that result in flawed data.
Finally, the traditional second-price auction theory doesn't allow for auction participants to have different budgets varying in size and scope. Simply put, how are you getting the most out of your auction, if the highest-bidder has an unlimited budget, while the second-highest is fighting to keep the lights on? If a bidder with a large war chest bids unrealistically high to win the auction knowing no one else would come close, that bidder has guaranteed a win at a substantial discount - at your loss.
In the end, advertisers adopt new bidding strategies where bidding your true private value is no longer the most profitable approach. In attempt to buy the sum total of impressions needed at the lowest rates possible through substitution, advertisers programmatically vary bids from impression to impression and auction to auction. When bids unpredictably and rapidly fluctuate, huge gaps between a winning bid and a second bid may occur. Lack of accuracy in and discrepancies between CACs and max CPMs when different attribution models are applied increase gaps between the advertisers' bids even further. As a result, these gaps discount publisher inventory prices, lower their potential revenue, and jeopardize their position in the market.
Secondly, the second-price auction's idea of truthful bidding heavily relies on the bidders' estimations of their private values - the maximum they want to pay for an auctioned item. But what if advertiser's information about their private values (max CPMs) is wrong or not exactly honest? Marketers' calculations of max bids are usually based on their estimation of Customer Acquisition Costs (CAC) which relies on their specific attribution model. Building such a model is a sophisticated task even the most savvy advertisers rack their brains over. Consequently, many advertisers use significantly simplified and inaccurate attribution models such as first- or last-click that result in flawed data.
Finally, the traditional second-price auction theory doesn't allow for auction participants to have different budgets varying in size and scope. Simply put, how are you getting the most out of your auction, if the highest-bidder has an unlimited budget, while the second-highest is fighting to keep the lights on? If a bidder with a large war chest bids unrealistically high to win the auction knowing no one else would come close, that bidder has guaranteed a win at a substantial discount - at your loss.
In the end, advertisers adopt new bidding strategies where bidding your true private value is no longer the most profitable approach. In attempt to buy the sum total of impressions needed at the lowest rates possible through substitution, advertisers programmatically vary bids from impression to impression and auction to auction. When bids unpredictably and rapidly fluctuate, huge gaps between a winning bid and a second bid may occur. Lack of accuracy in and discrepancies between CACs and max CPMs when different attribution models are applied increase gaps between the advertisers' bids even further. As a result, these gaps discount publisher inventory prices, lower their potential revenue, and jeopardize their position in the market.
What is wrong with second-price auctions for programmatic advertising?
Like all good theories, the second-price auction model loses its advantages when implemented in the real world. The programmatic advertising environment adds new variables to the equation and the idealistic concept of a fair auction where buyers bid their true private values starts to crack. These inefficient auction mechanics imbalance the market, jeopardize publishers' positions in the market, and lower their potential revenue.
First of all, the logic behind truthful bidding, the main theoretic advantage of the second-price auction model, is only true if a singular, unique item is auctioned. However, impressions auctioned in programmatic advertising are not as unique as you would think - ad auctions are replayed over and over again against the same visitors, websites, inventory units, ad sizes, and more. It doesn't take long to analyze auctions won and lost to find a way to game the system.
Like all good theories, the second-price auction model loses its advantages when implemented in the real world. The programmatic advertising environment adds new variables to the equation and the idealistic concept of a fair auction where buyers bid their true private values starts to crack. These inefficient auction mechanics imbalance the market, jeopardize publishers' positions in the market, and lower their potential revenue.
First of all, the logic behind truthful bidding, the main theoretic advantage of the second-price auction model, is only true if a singular, unique item is auctioned. However, impressions auctioned in programmatic advertising are not as unique as you would think - ad auctions are replayed over and over again against the same visitors, websites, inventory units, ad sizes, and more. It doesn't take long to analyze auctions won and lost to find a way to game the system.
Conclusion
For better or worse, RTB functions on the second-price auction model with crucial limitations affecting a publisher's position in the market and their revenue. Because advertisers are able to game the system and employ bidding strategies that result in low bids and/or huge gaps between the winning and second-highest bids, the current state of the industry is proving untenable. As it evolves, where will the industry look? Will it be to first-price auctions? One unified second-price auction with first bids submitted by all demand partners? Or something else entirely, something one step ahead? As programmatic automation becomes more sophisticated, efficient, and accessible, is it possible to dynamically set the price floor for each individual impression to overcome market inefficiencies?
For better or worse, RTB functions on the second-price auction model with crucial limitations affecting a publisher's position in the market and their revenue. Because advertisers are able to game the system and employ bidding strategies that result in low bids and/or huge gaps between the winning and second-highest bids, the current state of the industry is proving untenable. As it evolves, where will the industry look? Will it be to first-price auctions? One unified second-price auction with first bids submitted by all demand partners? Or something else entirely, something one step ahead? As programmatic automation becomes more sophisticated, efficient, and accessible, is it possible to dynamically set the price floor for each individual impression to overcome market inefficiencies?