If header bidding offers such a great boost for yield among its most vocal champions, why would a programmatically active publisher have reservations about implementing it? Well, there are a few things. But one of the top concerns making some publishers shy about header bidding is latency. Pre-bid calls are being sent and received before page content even loads, and pre-bid partners may be sending inventory information onto DSPs or other sources to get bids. That’s a lot of variables, which can make for hiccoughs—any delay in this process will likely cause page latency and bring on the user ire.
So what about solutions, ways pubs can troubleshoot on their own? There’s a two-fold response. First, thoroughly test the technology of your header bidding partners before setting them live. Take note on average load times, and be in constant communication with their representatives to ensure all is well. After all, the tech providers are making revenue as well, so it’s in their primary interest for their technology to be working swimmingly. Talk with your development team, and rely on them for direction on how to best handle this testing.
That’s also because of part two: Have safety timeouts set up for all your partners. Simply put, if your pre-bidders don’t hand you something within a flash (whatever length you decide that may be), cut ‘em off and start the auction. Determining cutoff times is something you should work out with your partners during and after testing. Communicate with your partners if you notice their pre-bids are constantly timing out—you can literally ask, “What’s the holdup?” But note that even with timeouts installed, some publishers simply aren’t willing to take any chances with latency.
As such, publishers should take a look at how the content on their pages (including images, video, CSS and the like) is executing prior to bidder callback. Contention for resources in the browser as page elements are loading can lead to bidders missing the call to the publisher ad server. Work with your development team on this issue, and ask them what they can do to help make sure page elements load smoothly and bidder callback functions correctly.
You might also need to bring product and design teams into this equation. Dev teams might be accustomed to working with product and design, and sales might also be accustomed to working with product and design—but sales is probably not accustomed to working with dev. If sales teams are in the habit of selling high-impact, heavy ad packages, and those units are slowing down your page load (which, as ops, is top of mind for you now that you’re implementing header bidding and trying to reduce any latency you can in that regard), you might need to loop sales into those conversations you’re having with developers. You’ll definitely want to talk with sales directly if they’re overselling inventory and you’re using header bidding to discover the real value of your inventory in the programmatic market.
You start with inserting a little tag onto your page, but you end up inviting all the teams in your organization to at least partly rethink how they work together and what their common goals are. That’s one possible outcome. And it’s a scenario that’s frankly prohibitive for some publishers. Getting these teams to coordinate can be very time-consuming and can prevent goals being met. Ironically, the twin facts that they have dedicated developer teams and institutional knowledge of the code on the page are often viewed as advantages in making header bidding work.
This is an excerpt from the AdMonsters Playbook: Header Bidding. Download your copy today!