Header Bidding: The Good, The Bad, And The Ugly

Header Bidding: The Good, The Bad, And The Ugly

Whether you think it’s a fad, a “hack,” the new standard, or the latest shiny object, header bidding has had a significant and disruptive impact on the advertising technology ecosystem. It may only be a matter of time before Luma Partners adds header bidding wrappers as a new box to their (in)famous landscapes.

The promise of header bidding with multiple exchanges has yielded positive results for advertisers and publishers but it has come with a cost that, over time, might be too much to bear. Whether it survives the fray, or evolves into something new, header bidding has changed the game forever.

Header Bidding: The Good

In the movie starring Clint Eastwood, The Good guy is not necessarily altruistic in nature, but he’s really good at capitalizing on opportunities.

Header Bidding brings those opportunities by providing premium inventory into the programmatic marketplace. No longer  are the best impressions locked away in the tower of publisher ad servers. They are now accessible via  the myriad of sophisticated Seller-Side Platform (SSP), Exchange and Demand-Side Platform (DSP) technologies. This gives sellers more articulate controls over the rules of engagement for every transaction.

Retargeting, made easier by RTB, can now be applied at all inventory priority levels, improving yield for commerce sites. Elusive audiences could be more readily captured by private marketplaces and the open auction, inviting new advertisers to test, refine, and commit to new deals with new partners.

Meanwhile, demand side systems are given the opportunity to bring more premium buying contracts to their platforms. This could be why some publishers started seeing higher revenue with header bidding in play. During Advertising Week in New York, one publisher cited header bidding as being responsible for 50% lift in CPMs. These types of statistics have been echoed by several others in the industry. While this might not be the panacea that saves the online newspaper, it certainly helps keep a few more lights on.

Header Bidding: The Bad

Unfortunately, it’s not all roses and rainbows in this brave new world. Like Angel Eyes, The Bad guy, ad-tech providers would be wise to watch their flanks, because header bidding, like many new technologies, can come with the cost of adding new features to existing systems, building brand new systems, standing up additional infrastructure, or giving the end user a little more frustration online.

Exchanges and SSPs, having grown up in a world where every impression request had the intention of yielding a rendered ad, are now waking up to a reality where render rates range from 30%, all the way down to single digits. The hardware and bandwidth costs still closely align with the volume requests, but transaction fees only come into play when the ad is rendered to the page. Higher CPMs, and the corresponding fees, are doing their part to fend off the increased costs, but it’s a challenging balance that must be closely monitored.

DSPs have a similar problem in that the number of requests from each publisher is increasing quickly as the full ad stack is made available programmatically, regardless of the actual availability of each impression for purchase. This is compounded further by publishers using several exchanges at once, which brings redundant bid requests for a single ad slot to the DSP. It happens when multiple exchanges, vying for the same impression, reach out to the same set of partners for bids. With systems designed for the massive transaction load of unique requests from the established world of RTB, DSPs don’t have the mechanisms to identify or filter identical requests for the impression.

Publishers are not spared from the shake-up of header bidding. In their effort to cash in on the added competition, many are partnering with multiple exchanges. This has the subtle effect of increasing the overhead of their operations. Each Exchange and SSP has a different set of  contracts, controls, and features. The learning curve can be steep if operations teams have to take them on all at once. And when a problem occurs through one of the partners, tracking it down can be an order of magnitude more difficult with each additional exchange added.

Finally, the end user has to take it on the chin with subtle hits to browser performance. The user’s experience, already hampered by inefficient blocks of 3rd party content, pixel drops, share buttons and videos, is further worsened by several parallel content requests to multiple exchanges for each ad spot on the page. Browsers generally have fewer than twenty swim lanes for downloading content. Adding another 15 or 20 asset requests per ad slot, header bidding is undoubtedly giving them the beachball.

Header Bidding: The Ugly

Tuco had the unfortunate and unscrupulous role as The Ugly in the Western. He makes a good analog for the short-circuited systems that bring life to header bidding.

DSPs, in an effort to streamline their operations and reign in their infrastructure costs may selectively suppress bid requests from certain sites through specific SSPs and exchanges. Industry insiders suggest that some may even remove a subset of exchanges completely. This tactic would tear at the heart of the harmony brought about by OpenRTB. No longer would supply and demand be ubiquitously connected through the exchanges. Buyers and sellers would have to return to the days of picking and choosing their partners based on what was available where. It’s a step backwards to the days of partnering with a few ad networks to monetize publisher inventory.

Auction mechanics have taken a strange twist. Most exchanges and SSPs leverage something akin to a second price auction when determining the clearing price of the rendered ad. Header bidding takes the clearing prices from each exchange and essentially runs another auction using first-price mechanics. With no set of rules governing the exchanges, some appear to have become simple bid pass-thrus, sending the bid into the header, rather than the second priced clearing value.

DSPs have also suggested that some exchanges are making multiple bid requests for the same impression. These exchanges may be trying to increase bid density to eek out a few dollars more for publishers in the face of stiff competition. Some have labeled this behavior “cheating,” but without governance or rules for what happens in the browser, there’s little incentive for publishers to dish out any punishments. This practice impacts DSP budget allocations across the exchanges, as well as causing trouble with media planning as some publishers appear to have as much as 5X the amount of inventory coming from abusive exchanges.

DSPs and exchanges are already suffering from increased operating costs will all the redundancy and lower fill rates. Dollars are not yet moving quickly enough into the space to make up for those costs. In the end, smaller companies may not have deep enough pockets to stay in the game until the budgets shift and the technology shakes out.

Server side header bidding (SSHB), the latest advent against the tech, holds the promise to clean up some of the mess. Moving the exchange requests out of the browser and onto a server will improve performance for the end user. It could also create opportunities for the exchanges to take a margin on traffic that they are not directly monetizing. They might, however, do this by adding an additional toll to each transaction and intermediaries will have to adjust their fee structures to cover the costs. With exchanges being the most likely players to host SSHB, the solution relies on these industry adversaries to trust one another, a difficult challenge in a game where everyone is holding their cards close to the chest.

In the movie version of header bidding, everyone wants to be Eastwood’s Good guy. He rides away with half the loot. But with any new advertising technology solution, there will always be winners and losers. Assuming the ecosystem can come together and solve enough of the challenges presented by header bidding, certainly the publishers and advertisers stand to win. As different attempts evolve to solve the problems header bidding introduces, ad-tech providers could end up in something akin to the film’s wild stand-off in the cemetery.

Save

Leave a Reply