In the world of programmatic advertising, first price auctions are the standard. Second price auctions used to be the go-to format, but Google officially transitioned from second to first price in September 2019 and most platforms followed suit. The move was encouraged by the changes Header Bidding brought to the ad tech dynamic as well as the overall shift toward transparency that the tech world is slowly embracing. Publishers anticipated rising CPMs in the shift to first price auctions, but that did not exactly happen.
We’ll explore the economic principles driving auction models as well as some game theory strategy behind it.
Before we deep dive, we’ll just define the terms we’re working with, so everyone is on the same page
First price auction– The price you bid is the price you pay if you win
Second price auction– If you are the highest bidder, the price you pay is one cent higher than the second highest bid
Bid shading– the strategy of lowering your bid to match historical trends and expected market value so you win, but don’t overpay
Clearing price – The winning bid amount
Reduction – the difference between the winning bid and the actual price in a second price auction
Floor Price – the lowest bid a publisher will accept, lower bids will not be considered
The Basic Economics of Programmatic Auctions
It’s very easy for publishers and advertisers to cry foul and say the power balance is off in the ad tech environment with publishers often feeling like they’re getting the short end of the stick, with lowered CPMs, more brand safety demands from advertisers and the like. It’s important though not to forget the first law of economics – supply and demand, whichever side has more, has less power. In the ad tech space, there are more publishers than there are advertisers, which shifts the power dynamic over to the advertisers. For publishers to claw back some power they need to offer something that is both not available by other publishers and is desirable to advertisers.
Niche publishers, premium publishers, and high traffic publishers can each differentiate themselves in these ways, positioning themselves to advertisers as good choices. They may even reverse the supply demand dynamic (unlikely though.)
Why Be Nice – An Economic Argument
Though the power may not be with the publishers there is still a sense of fairness expected. This is both on a human level and as well as practical. While publishers may be a dime a dozen, if advertisers don’t respect them, more publishers will leave or change the space. This can lead to shifting power to the publishers.
A prerequisite for fairness is transparency. If people can’t peek under the hood, there’s no way for them to judge accurately if what they see is what they are getting.
Regarding First Price versus Second Price Auctions: first price auctions are considered more straightforward in execution as well as in detail. What someone bids is what they pay. For the seller receiving payment there’s little room for buyers to play games with the publisher (with other bidders will be discussed further.)
Conversely, second price auctions rely on the indication of the second price, which can be manipulated due to ad techs’s opaque nature. For example, an advertiser may know that they won a second price auction and what their bid was. However, they don’t know what the second highest bid was, they only know the price they are quoted to pay. This leaves room for algorithms to submit artificial bids keeping the price of the second bid higher than the real value.
Only transparency can prevent such actions. While second price auctions are easier to manage when aiming to pay real value, the lack of transparency makes it unsuitable for large scale ad tech operations.
Which brings us to,
Game Theory and Strategy in Programmatic Auctions
Auction strategy is closely studied in game theory and the research offers unique insights.
Second Price Auctions
Starting with Second price auctions, the strategy is straightforward, and referred to as a dominant strategy. Bidders are encouraged to bid the real value of the space because if they win, they’ll pay the second price’s bid plus a penny. That second highest bid can be close or significantly lower than the winning bid, but it will always be lower, therefore whatever price they pay is something that has real market value.
That’s in theory.
But as previously explained, due to the opacity of the ad tech bidding environment, there’s no way to know. There’s also the issue of there often being multiple successive auctions of mixed types which can skew the results. For example, a high-priced winning bid in a second price auction will result in it being lowered to the next highest bid, plus a penny. When it goes to the next auction (either first or second price) with its new lower winning price, the new winning bid may be lower than the original winning bid. Due to the second price auction system, the originally higher bid loses. In this case there are two losers, the original higher bidder, and the publisher in lost revenue.
First Price Auctions
First Price auctions are not as straightforward in strategy. This is because it puts the onus of value on the bidder without adjustments based on the second highest bid. This leaves bidders concerned that they are overpaying.
To avoid this, parties bidding will shade their bid to adjust to what they think other bidders may be doing. It gets complicated because other bidders are shading their bid as well. And if they’re using the same calculus to shade their bid the odds of winning diminish.
While advertisers want good prices, they also want to achieve their marketing goals and if they fail to win auctions, they’re not helping themselves. If an advertiser experiences this, they can raise their bids, and the system balances itself out.
There are concerns though, that advertisers can simply drive down the price of inventory by collectively bidding lower and lower. However a carefully calibrated price floor by the publisher can prevent this.
Additionally, according to game theory the more bidders in an auction the less bid shading a bidder should engage in. With each additional bidder a bidder should bid closer to their true value if they want to win.
Economics in First Price and Second Price Auctions
Looking at the economics dynamics in First and Second Price Auctions also falls back on basic economic principles.
In a Second Price Auction the winning bid receives the average market value. That is of course in theory, the ad tech environment is more complicated, with many more layered and moving parts.
In a First Price Auction the winner receives the value they ascribe to the space. Different people have different values is normal and the market should reflect and bear those differences.
For example, some people think brand name products are everything, and others think it’s folly. For those who ascribe to the brand name value, they’ll be willing to spend a significant sum, while the other person may not even buy it on sale. Perception is entwined with value.
In the case of ad inventory, if a bidder values a particular type of user, they’ll be willing to pay more for it. That doesn’t make the ad space intrinsically more valuable, this value is solely bidder based. So, if an advertiser doesn’t value a space as much as another, they don’t have to bid as high, and that doesn’t mean the price will always be high. Prices will fluctuate based on the intrinsic value and the assigned value of a space. These self-corrections are the invisible hand of the market where entities naturally adjust to the real market value
Publishers play a role in indicating value by setting a price floor. Advertisers look at those floors under advisement. They can choose to abide by them or bow out of the auction. Whatever action they take triggers the next correction of the market.
What Should Advertisers Do?
Advertisers want good prices, they also need the ad space, so it’s a matter of how to be most strategic in winning auctions while keeping costs appropriate.
Advertisers should identify users and publishers they value highly and bid higher when those opportunities rise and otherwise play a lower risk game to keep media prices stable.
Another option is for advertisers to participate in a pmp/preferred deal. In a preferred deal advertisers can secure the media they need at a rate they know in advance. Next Millennium can aid advertisers in facilitating a PMP with our curated list of publishers.
What Should Publishers Do?
Publishers need to be savvy and understand dynamic floor pricing, figure out the optimal ad unit density per page and more. However they don’t have the tech capability to implement it. Therefore partnering with an ad tech partner like Next Millennium is vital. With sophisticated algorithms to predict bids, we find the perfect price floor to maximize your revenue while maintaining the value of your ad inventory.
Additionally signing onto a PMP/Preferred deal can ensure you revenue at a rate you’re satisfied with.