First-Price Auction – What does it mean for publishers?

RTB; The Evolution of Programmatic Advertising

The advent of RTB (Real-Time Bidding) in the late 2000’s changed the way digital publishers monetized their ad space forever. Auctions were introduced, enabled by technology, to bid on an impression-by-impression basis. Thus, replacing sequential CPM paid directly for ad space or fixed fee, sponsorship-focused strategies replicating print advertising. These auctions rely on a bidding system but unlike real-life auctions, where the highest bid wins after a series of bids, here, all bids happen at the same time. Traditionally, with a real-time auction, the highest bidder would, of course, win, however, only paying what the second-highest bid was.

Within milliseconds, a user would visit a web page and before the content loads, the publisher will send out a request for bids to partners such as SSPs, DSPs, Ad Exchanges, and Ad Networks. Each buyer would then return with a bid in real-time, the winner gets to show their advert.

This evolution thrust pricing strategies and auction dynamics to the fore. Publishers wanted to achieve the highest price possible and generate the most value from their impressions. Advertisers wanted to pay a fair price for inventory at scale. Two pricing strategies have dominated the conversation.

First-Price vs Second-Price Auctions

First-price auction vs Second-price auction

Essentially, the difference is down to two things; the importance of the reserve price (soft floor) and the hard floor and the rate paid by the buyer. 

In a first-price auction, multiple buyers submit a bid on an impression all at the same time, the highest bid price wins, and that clearing price is paid if it beats the hard floor rate. Buyer A bids $4, Buyer B bids $3.50, and Buyer C bids $3, then Buyer A wins, and Publisher receives $4.

In a second-price auction, while the highest bid also wins, the second-highest bid is taken as the clearing price and value of the impression. This time, bids below the reserve price cannot win. What the buyer pays in this instance is the second-highest bid plus $0.01. So if Buyer A bids $4, Buyer B bids $3.50, and Buyer C bids $3, then Buyer A wins and Publisher receives $3.51.

For a long time, second-price auctions were the norm; designed to offer the fairest price, providing the most accurate picture of inventory value. This is because the rate ultimately clears at an average position between the lowest and highest bids.

What are Auction Dynamics?

Now that we understand the differences between the auction types, we need to look at how buyers and sellers use them. Many digital publishers use a combination of direct sales to advertisers, agencies and networks, exchange, SSP or DSP integrations, and Header Bidding technologies. The complexity comes when you increase granularity and introduce multiple tiers of competition. Multi-level auctions can occur resulting in volume over value and buyers winning bids that were ultimately not the highest value.

With so many interested parties, the rules of engagement become more creative. Some outcomes may favor buyers, where the winning bid may not be the highest bid. Sellers, in turn, can make use of soft and hard floors. What happens as a result is a hybrid of first and second-price auctions, aiming to reduce the difference between the bid price and the clearing price.

Additionally, multilevel and undisclosed auctions have given rise to the perennial problem in the ad-tech industry, of different players taking undisclosed fees; ad-tax. Exposing hidden costs will not be solved completely by a first-price auction, but it does make the practice harder to hide.

What all this does is increase the distrust between sellers and buyers as transparency of bid values and ultimately inventory value are not clear. When a second-price auction is part of a sequence of auctions, then what is the true value? Buyers want to be able to know what they are buying and not get involved in games involving price. Floor optimizations enable publishers to maximize their CPM and page yield forcing the buyers to pay closer to their maximum bid capability. But at what cost? 

Why did Google move to a First-Price Auction?

In May 2019, Google made the announcement to move its Ad Manager/DFP and AdX to a first-priced auction over the course of the rest of 2019. A big change in a direction having led the standardization of a second-price auction within the digital advertising ecosystem.

Moving to a unified bidding system and the first-priced auction ensured Google no longer got involved in multilevel auctions and the complexity of auction dynamics. The aim was to level the playing field and simplify the programmatic ecosystem. Creating a more transparent and fair environment for buyers to compete equally and sellers to achieve fair value. 

Ultimately, Google AdX was competing on an uneven playing field against other technologies, such as header bidding, exchanges, and networks on a publisher page. While these other technologies utilized a first-price auction, Google AdX would return with the results of their second-price auction. This meant Google AdX second-price auction was competing against a first-price auction that had essentially set the floor.  The move to a first-price auction ensured Google no longer lost out to those using sequential auctions and specifically header bidding. 

What does this mean for Digital Publishers?

On the face of it, moving to a first-price auction implies clearing rates and CPM will go up. While Google insists on simplicity, fairness, and transparency are brought by the move to a first-price auction, the reality is less certain for digital publishers. 
One thing that can be guaranteed, is that buyers will now be forced to alter their bidding strategies. Where once their bid would clear at $0.01 above the second-highest bid, now, it would clear somewhere above that. In the earlier example, the same bid in either auction, led to Buyer A spending $0.49 more on the same impression in a first-price auction.

Buyers will, in the long run, need to be a little bit more conservative and lower bids to find that sweet spot. Especially as buyers focus on impression value (how much they can pay to win) may turn to impression cost (how little they can pay to win). This should all level out in time as buyers and sellers figure out best practices. 

What should Digital Publishers do?

Digital publishers will need to better understand and then justify the value of their inventory if they want buyers to bid at favorable rates. The market is moving to a more supply-focused state, so positioning price points and audience correctly will ensure buyers are comfortable with a bidding strategy that clears at face value. 

Numerous SSPs, DSPs, and Exchanges are looking at how they can automate the bidding strategy. Only by bidding, winning and losing, do you get an understanding of quality, value, and cost. 

Digital Publishers should focus on the following, to give them the best possible chance of success in a first-price auction world:

  1. MORE on understanding your audience
  2. MORE on your inventory quality
  3. LESS on auction dynamics and gaming the system

Ultimately buyers want transparency and good quality inventory at a fair price. Solve the first two and you won’t need to worry about the third.

What does the future hold?

While nothing is certain, we do know that OpenRTB Protocol provides the specific fields that contain auction type, so buyers and sellers should be clear what they are engaging in, ensuring transparency. 

Ideally, we should want to see all exchanges operating on a first-price model and closing on a second-price model. In practice, this would mean a publisher would have multiple SSP partners, operating a first-price auction with multiple buyers. The clearing price to the publisher should follow once all winning bids have been submitted and the SSPs provide their first-price results. The publisher can then select the winning bid based on a second-price auction. 

In this scenario, SSP 1 competes with SSP 2. SSP 1 has Buyer A bidding at $4 and Buyer B at $3.50. SSP 2 has Buyer C bidding at $3 and Buyer D at $5. SSP 1 returns with Buyer A at $4 and SSP 2 returns with Buyer D at $5. The Publisher would see a clearing price of $4.01 and the winner is SSP 2 and Buyer D. 

The difficulty is having all parties operating on the same model, in the meantime, digital publishers should focus on what they can control; understanding their audience and providing quality inventory.