Have you ever asked yourself what Header Bidding is? Or worse, have you ever been asked what Header Bidding is? Do you see companies promoting their Header Bidding solution, or Dynamic Allocation, or Cloud-Based Header Bidding solutions? If you answered yes to any of these questions but don’t fully understand what they mean, then keep reading.
To start with, let’s define who even cares about Header Bidding. Answer: Digital Publishers. Those can be websites or apps, mobile sites and desktop. ‘Publishers’ is the umbrella term for any company that has a digital presence and monetize it with ads. Be that a free mobile game to the New York Times, all of these Publishers care about Header Bidding.
Part II of that answer is that DSPs, SSPs, Exchanges, and Ad Networks all care about Header Bidding too, but publishers drive most of the demand. In short, Publishers want a Header Bidding solution, so it’s up to demand partners to provide one.
With that out of the way, let’s start to talk about what the heck Header Bidding actually is.
At its core, Header Bidding is a solution that allows advertisers to bid on digital inventory in a true auction environment. That means if three advertisers all submit a bid for the same 320x50 ad unit, whoever bids the most wins. Seems simple right? Well...kinda.
The issue at hand is that up until a few years ago, a true auction rarely happened. Instead the industry used what’s called a waterfall. A waterfall meant that a publisher would have multiple demand partners lined up in a waterfall setup. When they had an impression to serve they would send it to the first in the waterfall to place a bid. If they placed a bid that was high enough, they won, the ad served, end of story. If the first in line didn’t have an advertiser, or their bid wasn’t high enough, the request would go to the next in line, and the next, so-on, and so-on.
While this setup does help ensure that publishers fill their impressions by having multiple demand sources lined up, it also presented a couple problems.
The First: The #1 demand partner in the waterfall would have a high floor (Floor being the minimum amount they had to pay in order to buy the impression) that they had to fill at (Example $2.00 CPM). If they had demand but not at the floor price, they wouldn’t win and the request moves on. The demand partner in the #2 spot has a lower floor (Example $1.50) so they could win for a lower amount. #3 is lower than #2, and so on. So if the #6 spot in your waterfall is winning the ad, it’s winning at a very low rate.
- While Publishers would rather get something than nothing, many demand partners realized that they were filling in position 4, 5, 6 but could’ve paid position #1 rates if they had the opportunity. This was an opportunity for both publishers and demand partners.
The Second: These transactions need to happen fast, and if the first few demand partners didn’t fill, the waterfall may timeout meaning that no ad serves at all. A faster solution would ensure more fill and more revenue.
The solution to this was a true bidding experience. An environment where the waterfall is removed and all demand partners bid at at the same time. The winning bid wins.
It’s important to note here that this is generally a second-price auction. (If you don’t know what a second price auction is, click here).
So why wasn’t this done originally? There are two answers to this question.
- No one thought about it.
- No one thought it was possible. Until it was.
The first company to introduce Header Bidding was OpenX. And it wasn’t introduced as an industry changing solution, it was a hack.
OpenX found that many times they were lower in a waterfall then they wanted to be, and were filling, but not filling as many impressions as they wanted, and not at the CPM level they could afford. They might be #5 in a waterfall and required to only fill at a $0.75 CPM, but if they have first look they could’ve paid $2.00 for the same impression.
For OpenX they wanted to fill more impressions at higher rates because that meant they could spend more of their client’s money and demand larger budgets. And publishers would rather get $2.00 than $0.75 for the same inventory, so it was a win-win.
OpenX found that Google was their biggest competitor (not a surprise) so they found a way to put some code in the header of a website that would allow OpenX to see the content of the website and ad units available, before Google. If the site had content and ads they wanted to buy, they could submit a bid before Google could, and win the impression before Google even saw it.
OpenX convinced publishers to put this code on their site with the promise that it would result in more money for them. An easy pitch, and it worked. This piece of code was placed in the header portion of a website, which is where we get the term 'Header Bidding.'
Once it took off other demand partners took notice and wanted in. Eventually other companies developed technology that could compete right along with OpenX so publishers would have multiple demand partners all bidding on their inventory.
Over the past few years the technology has developed even more, but this is how it began and what it means. For more information about how it works in mobile vs. desktop check out additional What the FAQ posts and check out AdCoach Courses.
TL;DR - OpenX realized they were losing potential impressions and budget to Google so they created a hack in order to place an ad before Google got the chance. This resulted in more revenue for publishers, and bigger budgets for OpenX. This was so successful that other companies copied the idea and developed technology to compete in a true auction environment. This code was placed in the header portion of a website, which is where we get the term 'Header Bidding.'
Courses That Cover Header Bidding