Back in the early days of the Internet, when an advertiser wanted to buy ad space on a publisher’s website, they would call up the publisher’s sales team to make the purchase.
While the direct advertiser-publisher relationship is still around, the introduction and growth of technology has added speed, efficiency, and scale to the online media buying and selling process.
As the number of technology platforms has increased, so too has the number of processes; and one process gaining traction in the ad-tech world is header bidding.
What is Header Bidding?
Header bidding (aka pre-bidding, advance bidding, and holistic yield management) is a process that enables publishers to simultaneously collect multiple bids from a number of demand sources (not only from their ad server) on all of their ad inventory prior to a sale.
This process allows publishers to “see” which demand sources (e.g. demand-side platforms [DSPs] and ad networks connected to ad exchanges and supply-side platforms [SSPs]) are placing the bids and the monetary value of their bids (e.g. $1.05), allowing them to get the highest cost per mille (CPM) possible. As bids are received before a publisher’s ad server is called, they are able to compete with a publisher’s premium deals, meaning advertisers have a better chance of winning the impression and publishers earn more money for their inventory.
How Did Header Bidding Come About?
AdTech vendors developed header bidding to enable publishers to collect bids on their inventory from demand sources before the ad call was sent to their DoubleClick for Publishers (DFP) ad server. The reason for this was to avoid the common situation where the DFP, which is used by a large majority of publishers, favored bids from the Google ad exchange (AdX), meaning non-Google demand sources would often miss out on bidding on or purchasing the publisher’s inventory, even though they were willing to pay more for it.
Header bidding has also, somewhat accidentally, replaced waterfall auctions.
A waterfall auction is a structured process for selling inventory handled by the publisher’s ad server, which would sell the available inventory to potential buyers in a sequential fashion.
Typically, the ad server would offer the inventory to premium buyers (e.g. guaranteed buys, direct sales made between the publisher’s sales team and the advertiser, and sponsorships). If the premium buyers failed to purchase all of the available inventory, the ad server would approach the next potential buyer (e.g. an ad network) with the aim of selling the remaining available inventory, known as remnant inventory.
The ad server would continue contacting various demand sources until all of the available inventory was sold. In the event the available inventory wasn’t sold, the publisher could display in-house ads (ads promoting their own products or services).
Learn About The 12 Main Challenges And Opportunities In Programmatic & RTB
A comprehensive guide for all those who work within the online advertising industry...Download FREE Guide
How Does Header Bidding Work?
From that moment on, each time a page loads, the demand sources (e.g. ad networks and buyers connected to supply-side platforms [SSPs] and ad exchanges) will be able to bid on each impression on that page, even if some or all of the inventory had already been sold through direct buys between an advertiser and the publisher.
A Step-by-Step Explanation of How Header Bidding Works
Here’s how the header-bidding process works from the moment a page loads on a publisher’s website to the time the available inventory is sold:
Here’s what is happening in the image above:
- A user opens their web browser and types in the publisher’s URL (e.g. publisher-abc.com).
- The browser starts loading the page.
- Bids from various sources start coming in.
- The highest bidder wins and the ad is served (i.e. displayed on the page).
As you can see in the diagram, the speed at which bids are placed varies. This latency not only means some demand sources may not be able to place a bid due to being timed out, but also results in the page taking longer to load. The timeout rates vary and are different on desktop and mobile. With desktop, the timeout range is 400–800 milliseconds; with mobile, it’s 800–1,200 milliseconds.
The process described in the image above is an example of client-side header bidding, but there is also a server-side implementation.
What is Server-Side Header Bidding?
Server-side header bidding follows a similar process like the client-side implementation, but instead of sending the requests from the browser, server-side heading bidding sends the requests to different demand sources (e.g. ad networks and DSPs) via SSPs/ad exchanges from one server-side header bidding vendor, which is typically an SSP/ad exchange or ad server (Google’s DoubleClick for Publishers [DFP] now offers server-side header bidding).
The server-side implementation is much faster and less prone to timeouts and errors from the browser, but its implementation is much more complex for the header-bidding vendor.
It’s important to note that while the header-bidding process is nothing new — it’s been around for about 10 years — it really started to take off in 2014 and 2015 with the help of programmatic media buying. In fact, according to Tom Shields, the SVP of Publisher Strategy at AppNexus, around 70% of major publishers use header bidding in their programmatic sales process.
Get our whitepaper!
Share this article
FREE AdTech & MarTech Resources
Thousands of C-level executives, software engineers, marketers, and advertisers all learn about the inner workings of AdTech and MarTech with our bimonthly newsletter — and so can you! Subscribe today and get access to the latest and best articles, videos, and guides!