Dynamic Price Floors Perpetuate An Ad Stack Cold War

willdohertyThe Sell Sider” is a column written by the sell side of the digital media community.

Today’s column is written by Will Doherty, senior director of business development at Index Exchange, a division of Casale Media Inc.

The jig is up. And it’s been up for a long while. It’s time to move past dynamic floors.

They simply have no place in the current programmatic ecosystem. They are shortsighted and deny publishers the opportunity to work closely with their clients – many of whom are gearing up for a programmatic-only future. You’re not simply short-changing a bidder by deploying dynamic floors, you’re hindering your ability to support accurate price discovery for buyers. If they can’t price it, they won’t buy it. More importantly, any effort to obscure the value or credibility of any given impression – to represent it as something it’s not – is deceptive.

Dynamic floors, or “soft floors,” allow publishers to create artificial market density in order to drive up the cost of a given impression. In any open exchange, buyers only expect to pay slightly more than the next highest bid. If a buyer bids $2 and the next highest bid is $1.50, most auction mechanics would sell that impression to the buyer at $1.51. When a dynamic floor is introduced into this equation, the SSP will create “phantom demand” that would price the bid at $5 as if that demand actually existed. That figure now represents the high end of the auction. But since it doesn’t exist, the SSP will sell that impression to the buyer at $2, since it is the highest bid below the dynamic floor of $5. This is an artificial price increase of nearly 33% for the buyer.

Doesn’t feel right, does it? We have turned an extremely sophisticated system that uses exponentially great numbers to determine price, develop efficient attribution schemes and respond in blindingly fast speed into something like a silent auction at the local Elks Club. It’s misguided at best. And, at worst, it’s dishonest since buyers are being led to believe that another buyer determined the price they paid, not an algorithm.

Buyers know this is happening. Many have developed features that will sniff out these floors and then bid down impressions until the real price is discovered. Many will strategically pull budget entirely, since their bidders are now tasked with sniffing out malfeasance instead of looking for efficiencies for their clients. Everyone loses in this scenario. Time and resources that could have been better spent on developing more impactful creative, a true deal ID marketplace and fighting actual fraud were spent correcting artificial obfuscation in the market.

We have to get beyond the ad tech “Cold War” and stop trying to outgun one another. It’s mutually assured obstruction.

In order for the programmatic marketplace to deliver the value both publishers and advertisers want, both will have to move away from the “best line of defense” mindset. This is especially true now of publishers. Marketers, independently of what happens in the exchanges, are moving closer to publishers on their own. Both sides want more direct, transparent and meaningful relationships, which is expressly the type of exchange programmatic infrastructure was built to facilitate. The flexibility in pricing and closing deals should be an asset embraced by both sides, using strategies that are advantageous to both and are agreed upon beforehand. That flexibility shouldn’t be used as a weapon. It’s fine for publishers to set floors for their deals, the lowest bid price they’ll accept or to choose who they will or won’t sell to – as long as it’s consistent and in the open.

As the digital marketplace progresses, publishers will need to start closing major deals with entities that manage buys around massive amounts of media. And, to do that, they’re going to need programmatic. Publishers can’t keep using dynamic floors to maximize revenue in the short run as it erodes the buyer/seller relationship when buyers find out that publishers are using them. Smart buyers know how to overcome this and their algorithms are bidding down CPMs until they hit the real floor – in effect training the dynamic floor algorithm to go lower.

Once a publisher’s inventory is trained by a bidding strategy, it becomes exponentially harder to win high CPM bids. You have trained the machines to look for an artificial efficiency. A good SSP can identify bid strategies by leveraging the entire swath of demand in the market to build a consensus pricing model. Artificial tools only perpetuate a misalignment. They correct nothing, and only hurt a publisher’s ability to extract the real value of their audience, which is often higher than any floor that could be set.

There’s no reason to perpetuate a “cold war” when we can use the same infrastructure to create customized deals that nurture lasting business relationships. Publishers can use the data in programmatic to unlock the real value of their inventory by gaining insights into how that inventory is accessed and valued. By using that data to ensure the inventory’s value matches the price they’re getting for it, publishers can create pricing strategies that give them a fairer price while reflecting what the buy side wants to pay. But the first step is getting rid of dynamic floors.

Follow Index Exchange (@indexexchange) and AdExchanger (@adexchanger) on Twitter.

1 Comment

  1. Jim Metzler

    There are differences between dynamic floors and soft floors. Soft floors are false floors -- they don't really exist. If nobody bids above the soft floor, either a second lower floor is used or the auction becomes a first price auction.

    Dynamic floors are set dynamically for each auction depending on the characteristics of the impression, time of day, browser/OS being used, user frequency, the relative demand for that user's impressions and perhaps other factors. Dynamic floors can be soft floors but they can also be hard floors -- if nobody bids above the hard floor then the impression goes to a house campaign.

    While dynamic floors can cause auctions to take on first price characteristics if a user is in the market for a car and many bidders know this information, standard second price dynamics probably play a bigger role in setting the clearing price. The real differentiator is whether a floor failure leads to a zero-value house ad or turns the auction into a defacto first price auction.

    Reply

Add a comment

XHTML: You can use these tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>