Changing auction dynamics mean advertisers may be overpaying for media, writes Brian Fitzpatrick. Is your programmatic ad tech up for the challenge?
Earlier this year, in this same media outlet, Emma Newman of Pubmatic made a case for why publishers need to understand and start paying attention to auction dynamics. The shift from second-price to first-price auctions was having serious implications for a publisher’s revenue, stack configuration, and monetisation strategies.
There’s certainly no doubt that publishers should understand auction mechanics more intimately, but I’d argue that it’s even more critical for agencies and the brands they represent to have a deeper knowledge of the issue and be prepared to get under the hood of the tech partners they work with to grasp the factors at play. It’s their money after all, and the shifting of auction dynamics means they risk over-paying for media if the buying platforms they use aren’t yet ready for the change.
Second price auctions
Programmatic trading has historically operated on a second-price auction model in which the second-highest bid determines how much the impression is sold for, i.e. how much the winning bidder will pay. For example, if the highest bid is £5 and the next is £3, the winning advertiser pays £3.01 (think eBay or Google AdWords).
Theoretically, this is a more economically efficient marketplace. It allows the buyer to express the maximum he or she is willing to pay for a piece of inventory, but often win the auction at a lower price, thereby creating ‘buyer happiness’.
First price auctions
Today, we’re increasingly observing a move to first-price auctions, driven largely by the adoption of header bidding and the need for supply side platforms (SSPs) to be more competitive in a holistic auction environment that has them competing against all other demand sources in the header simultaneously, including direct-sold campaigns.
Using the example above, in a first-price auction the winning advertiser will now pay the full £5. There are arguments to be made that this increases transparency, so long as the type of auction being run is passed on to the buyer, as advertisers now know exactly what they’ll pay should they win.
It’s not hard to see why this shift in auction mechanics also requires changes to an agency’s bidding strategy. A buyer accustomed to the second-price environment might bid ridiculously high to make sure he wins the impressions needed for a specific client’s campaign, knowing he’ll never pay the full price.
This same tactic in a first-price environment would have the buyer paying significantly more than he’d like or expect to pay for that same inventory, which leads to buyer’s remorse and a resultant shift of budgets to other supply sources or paths.
It’s tempting to assume that programmatic ad buying technology (namely demand side platforms or DSPs) would do this automatically.
But here’s the rub – most DSPs were built and fine-tuned to work in a second-price environment. They process massive volumes of data to make split-millisecond decisions about how much to bid on specific pieces of inventory based on an advertiser’s individual campaign settings, targeting parameters, pacing status, performance requirements and other factors to ensure they win the impressions they need at the optimal price for the client.
This same intelligence and processing horsepower breaks down when bidding in a first-price auction; and supporting this transition requires a tremendous amount of development work on the part of the DSPs.
To compound the situation, SSPs often deploy a variety of tactics to extract greater yield for publishers (and raise the winning bid price for buyers in the process). In many cases, an agency and its DSP may not even know what type of auction is being run. This makes the likelihood of overpaying even greater and the need for automated auction detection and bid intelligence even more critical for brands.
So what should agencies and brands be doing to offset this risk?
1) Evaluate the level of transparency they have into the auction mechanics of the supply partners from whom they are buying. If the answer is none, consider alternative supply paths.
2) Develop differentiated bid strategies for second vs first-price auction environments. Bid shading (i.e. setting the maximum bid slightly lower than the true value one is willing to pay) is a common tactic to ensure some level of payoff and avoid overpaying in first-price environments. This comes with risks though including potential lower win rates, which may need to be considered when setting up campaigns.
3) Ask DSP buying partners what automated tools they have to sniff out different auction types, and what technology they have to properly adapt bid strategies in real time to accommodate those different auction mechanics. Again, if the answer is none, consider migrating spend to other systems or bringing on other partners who can provide this intelligence.
While the benefits of programmatic are well cited, the shift has also introduced a host of new challenges and unknowns to both advertisers and agencies. As the saying goes, it’s turned advertising’s Mad Men into Math Men.
Today’s programmatic buyer needs sophisticated algorithms, tools and technology up to the challenge of this constantly evolving buying ecosystem to ensure they can still deliver on the promise of finding the right inventory at the optimal price.