IX Perspectives

Header Bidding: It’s All About Latency

Electric graphic lock

Header bidding is barely out of its infancy, but it has quickly proven to be a powerful programmatic technique to drive revenue and maximise yield for publishers. While there is a great deal of financial incentive for publishers to implement a header bidding solution, Ad Ops and Product teams are still getting comfortable introducing them into their ad stacks.

As with any new technology that has entered the marketplace, there will always be a learning curve and a certain amount of testing that is required. Publishers have to balance the competing demands of improving user experience, maximising revenue, and supporting their own direct sales teams, all while implementing new technology and minimising the friction to go live. It’s a tall order and any tech that makes that more difficult isn’t going to last long.

The key to a successful transition to the header has as much to do with implementation as it does the specific solution itself.

Publishers have a responsibility – and a strong interest – to provide readers with a high-quality experience. Different types of content, device, advertising, and latency control and measurement mechanisms will all have a direct impact on user experience optimisation. So how do you get the right implementation? It’s a question of balance.

“The Index solution was an easy choice once we examined it,” said Erik Requidan, Vice President of Sales and Programmatic Strategy at Intermarkets, Inc.. “It offered the same value prop with similar latency as remarketing solutions, but exchange-level demand.”

Impact on Latency

One of the great advantages of initiating bids from the header is the way that it can directly improve latency by front-loading decisions. Publishers should refine their implementation to take advantage of this.  The process starts with the acquisition of bids and creative before the ad server auction. Should a header partner win the auction, the ads can then be displayed as quickly as possible without having to make secondary requests as is the case with conventional tags and certain suboptimal header solutions.

In order to optimise for low latency, publishers should be loading all header partners asynchronously, as this ensures that pages load as quickly as possible. The last thing a publisher wants is for a latency issue on ads to block the loading of page content while waiting for a bid response. Tag management, wrapper and mediation solutions are the common models used by publishers to reduce latency by setting a timeout limit to all header bidding tags, which will cap latency levels to where they are acceptable. Any cases where excessive latency occurs due to header bidding is the direct cause of deficiencies of the wrapper implementation.

Modernising Direct Sales

Another advantage of header bidding is its ability to deconstruct the traditional waterfall, which is designed in such a way that it places a higher value to certain demand sources, regardless of the CPM that they are willing to pay. The traditional waterfall setup can leave money on the table for publishers.

Header bidding allows publishers to compare demand from both programmatic and direct sources equally, ensuring that those programmatic buyers who are willing to pay higher prices for very specific, high-value users can reach them, and publishers can increase their yield.

While that is great news for publishers, a common misconception is that header bidding will cannibalize a publisher’s direct sales efforts. A properly implemented header bidding solution will allow publishers to bring large guaranteed budgets back to direct sales via header fulfilment channel, instead of having those impressions find their way via indirect channels like the open market at only a fraction of the spend. These programmatic direct campaigns are monitored similar to conventionally booked direct campaigns to ensure delivery commitments and pacing are fulfilled.

“In the industry, we’ve all been conditioned to assume that any increase in revenue has to be cannibalizing the take from somewhere else,” said Scott Mulqueen, Director, Programmatic and Audience Strategy at About. “Header doesn’t work like that — when appropriately configured, it’s all incremental. We’ve gone from a zero-sum game to a positive-sum with a much greater degree of control to manage the potential upside.”

Because header bidding is a fairly technical advancement, publishers and their Ad Ops teams should experiment and work with partners that understand the potential issues that can arise. Header bidding was second only to ad blocking in the publisher’s zeitgeist this year, so it’s important to look beyond the buzz to learn how it will work best for your individual needs.

With the current popularity of ad blockers, user experience is a number one priority for publishers. While header bidding can alleviate many of the challenges that publishers face with the traditional waterfall setup; if not implemented correctly, it can also create new ones. By adding just the right number of header tags, and executing header bids asynchronously with latency control, publishers can ensure that their page load times remain top speed and optimised to drive revenue.

Technical Addendum

The browser controls the queuing of async events. When installing a wrapper solution we fire all the tags immediately and based on what is already in the browsers async queue, and how long the different requests take to process, the various bidders will return with demand. This happens in parallel with no bidder given preferential priority in async. Alternatively, with sync, it is serial and makes things take longer thus we advocate for that practice to no longer be utilised.

As published on LinkedIn https://www.linkedin.com/pulse/header-bidding-all-latency-jourdain-alexander-casale?trk=pulse_spock-articles

Leave a Reply

Your email address will not be published. Required fields are marked *