Programmatic Media Buying 101: Bid Shading And Bid Caching Defined

Programmatic auctions are experiencing a big shift in the way inventory is priced. Media buyers who have been using second price auction strategies are finding that more and more SSPs are moving to first price auctions.  In the second quarter of 2018 a report by Getintent found that 43.3% of impressions were sold through first-price auctions, where the highest bid wins an impression. That figure was much higher than the previous study done in December 2017, when just 5.8% of the impressions analyzed were sold in a first price auction. With more money shifting to first price auctions, the programmatic advertising industry, both on the buy and sell side, had to come up with strategies and tactics to mitigate the effect this change has had on the way they buy and sell inventory.

Bid Shading

Bid shading is one of the tactics, which is relatively new to DSPs, that has been developed as a compromise between first- and second-price auctions.  Buyers are unhappy about paying higher prices than they were used to in second price auctions. Bid shading reduces the price for the advertiser so rather than pay the full first price, they get charged and average price between the first and second price auction.  The price is usually calculated by the SSP or DSP from bid history information which includes typical bid win rates for certain websites.

For right now bid shading is mostly used by the supply-side platforms (SSPs) and is free to the user.  But as first price auctions become the norm for programmatic media buying, more Demand Side Platforms (DSPs) are adopting this tactic out of necessity.

Another, less accepted method of dealing with the shift in auction dynamics has been developed by Ad Exchanges and is called bid caching.

Bid Caching

Why has bid caching all of a sudden become an important ad-tech term to be familiar with over the past month?  This is due to the fact that media buyers using the Global ad marketplace Index Exchange did not know that this tactic was being used for over a year until they published this blog post.

“We didn’t think it was an issue with buyers. We were so surprised. We thought this was an industry practice,” said Drew Bradstock, SVP of product at Index Exchange.

So What is Bid Caching and Why Do Media Buyers Not Like It?

It’s when a lost bid on a programmatic auction is used to fill a subsequent auction, where the impression characteristics do not necessarily match up.  For example, if the buyer bids on a particular publisher’s homepage to appear in the morning and loses that bid, the exchange would roll the bid into another auction for the next piece of content that the consumer views. So instead of running on a homepage, the ad would most likely end up on an article page. For the media buyer, this means that they are not  getting what they originally paid for.

So what are the other negative aspects of bid caching for DSP buying?

  1. Overpaying for inventory: advertisers are willing to pay higher prices for the first page in a user session, so if that’s not what they are getting then they have overpaid for the ad
  2. Messing up frequency caps and pacing: delays between a bid and ad server could increase the likelihood that a DSP found the user elsewhere
  3. Brand safety concerns: ad is served on the same domain but not on the same page

What’s Next for Programmatic Media Buyers?

Ad tech professionals have expressed their frustrations with yet another example of the lack of transparency in the industry.  Just when we think we are finally making strides to improve the negativity surrounding programmatic advertising, a story like this pops up and stirs up controversy all over again.  The negative press that has resulted from Index Exchange’s lack in public disclosure of their tactics has hopefully served as a warning to others in this space to own up to any changes in their platforms that would have a drastic effect on the way inventory is bought and/ or sold.  In the meantime, it’s important for media buyers to know all the tactics and associated terms with the move to first price auctions and how their DSP partners are adapting to the changes being made.  Talk to us if you have any doubts or questions about what your next steps should be.

Programmatic Media Buying 101: The Difference Between First and Second Price Auctions in RTB

If you are buying advertising programmatically then you are most likely using either a first or second price auction bidding process.  Most recently there has been more talk of moving towards first price auctions because of the popularity of header bidding.  DSPs (Demand Side Platforms) have traditionally been set up to use second-price auctions and for most DSPs adapting and changing strategies to first price auctions is expensive because they have to invest in technology that will specifically adapt to the rules of every auction and allow them to bid effectively.

So what are the differences between the two types of auctions?  And why should media buyers care which one gets used?

First Price Auctions

The programmatic buying model where if your bid wins, you pay exactly what you bid. This type of auction maximizes revenue potential for the seller.

In the first price auction model the bidders pay exactly what they bid. This type of auction can lead to unnaturally high prices because buyers are forced to guess how much their competition will bid.  This auction mechanism gives publishers the highest eCPMs for their inventory but can lead to the advertisers overpaying which can then lead to a lower demand for that publisher’s inventory.

The first-price auction allows both buyers and sellers to see the actual cost of the impression and the fees taken by the SSP/ad exchange will at least be known. The winning price is exactly what the advertiser agreed on, but there is a risk of overpaying for impressions.

The workings of the first-price auctions make sense economically only when the buyer knows the fair market value of the impressions they are bidding on, and understands the mechanics of hard- and soft- price floor mechanisms. The Price Floor, is the minimum price a publisher will accept for its inventory, which technically means they will ignore all bids below that price. This can turn a second-price auction into a type of first-price auction.

Second Price Auctions

The programmatic buying model where if your bid wins, you pay $0.01 above the second highest bid in the auction. In this type of auction, it is in your best interest to bid the highest amount you are willing to pay to win that impression, knowing that you will most likely end up paying less than that amount.
The second price auction is preferable to first price auctions for advertisers because it gives the winner a chance to pay a little less for the ad impression than their original submitted offer — instead of paying the full price, the winning bidder pays the price offered by the second-highest bidder, plus a bit more, usually $0.01. The final and winning price of the impression is known as the clearing price.

So What About Header Bidding?

Header bidding has become a popular type of first price auction where publishers place a piece of code on their webpage headers that allows a limited number of advertisers to bid on inventory outside of their primary ad server. This lets advertisers compete for premium or reserved inventory before or instead of the second-price auction.

Header bidding creates an auction prior to the final auction in a publisher’s ad server. Because of that inefficiency, SSPs (Supply Side Platforms) who run a fair second-price auction in the header, will have less competitive bids for that final auction, and find themselves with low win rates.  Being less competitive in the auction has terrible implications for SSPs as more competitive bids from header bidding can steal their market share.

Media Buyers Are Asking for Transparency in the Bidding Process

Programmatic ad buying exchanges have a mostly obscure bidding process, making it unclear for the buyers whether they are dealing with first or second price auction. If you want complete transparency, then first-price auction seems to be the better option (there are no floor mechanisms or hidden fees), but it offers few real benefits for the advertiser. Truthful bidding in this model (i.e. bidding the real value of the impression, which means if an impression has a value for you of $1.00, you should also bid $1.00) is not only more challenging but it’s also more expensive. A transparent first-price auction will squeeze the margins of the many ad tech players in the middle, and deliver more actual working media to the publisher. But if programmatic media buyers think they are still playing according to second-price auction rules, they will end up overpaying for inventory. Advertisers don’t like the feeling that they are being manipulated into bidding higher than they need to, which is exactly why DSPs use algorithms to predict the price floors and bid accordingly.

Many programmatic traders are left in the dark when it comes to the setup of the auctions they are bidding in. Since media buyers can only audit the vendors they are working with directly on the demand side, they have no way to verify if other programmatic platforms in the ad supply chain are altering their auction structures to make more margin. Which means a buyer might think they are buying based on second price auction but really be in a first price auction. That can get expensive, since the bid strategies are drastically different.

The industry will likely be in a transition period for much of 2018 as DSPs adjust their algorithms to allow for some Bid Shading to minimize the chance of overpaying.  It’s important for media buyers to clarify the auction type (first or second) whenever negotiating a deal and floor price with a publisher. To combat price increases, some buyers have already started Bid Shading, or reducing bid prices. But that strategy comes with risks because buyers can lose out on inventory they want if they submit too low a bid.  So until trust or transparency in auction type and fee structure is available in the open exchange, some media buyers will either try to work with adjusted algorithms or push towards more private exchange tactics so that they can trust the contracts and pricing models.

Like what you see? Join the 500+ clients that have partnered with Digilant.