Updated: Nov 2
So, you want to know about header bidding. Well, here’s a handy header bidding guide.
Programmatic advertising and header bidding is complicated, and a lot of the content out there about it is complicated, too. There are a ton of resources written for people at varying levels of understanding: blogs that address particular issues, white papers that may require Googling a different article to understand it all, infographics that break it down but don’t provide a holistic view. Sometimes, what you’re looking for is a definitive guide that helps you comprehend a new concept or idea.
In our header bidding guide, you’ll learn about the past, present and future of header bidding technology and how it can increase your ad revenue in simple, clear English. We’ll only use jargon when we absolutely have to, and we’ve included bolded terms (school textbook style) and a programmatic advertising glossary to help you with unfamiliar terms.
We hope our header bidding guide is a useful resource for you and your team. If it isn’t, we’ll give you your money back.
The Evolution of Programmatic Advertising
To properly understand where header bidding is now, it’s important to review how the industry used to operate.
In the days of old, digital ads were bought and sold by people. It was a time and labor intensive process where a publisher would talk to an advertiser and negotiate rates and placement. With technological advances made in the advertising field, programmatic media buying was introduced to save publishers and advertisers money and time. Programmatic ads revolutionized the online advertising world as it allowed software to do the work of humans when it came to purchasing digital advertising.
How does programmatic ad technology work?
Anytime a publisher looks to sell an ad programmatically, there are five pieces of technology involved. These components engage within the programmatic software that serves as the go-between for the publisher and the advertiser:
The publisher’s ad server (houses ad space for sale)
The ad exchange (where auctions are conducted)
Bidder (the technology that buys inventory on ad exchanges)
The advertiser’s ad server (houses the ads they’re looking to place
Opt-Out (an option for advertisers not to receive bid requests from specific publishers)
The publisher’s ad server and ad exchange fall under sell-side technology, whereas bidder, the advertiser’s ad server and opt-out fall under buy-side technology. All of these pieces work together to automate most of the digital advertising process.
Obviously, this made the digital ad buying process much easier, but publishers were still spending a lot of time managing the ads themselves. Advertisers would send their ads to publishers who were then responsible for setting up, serving, and taking down the ads on their website. Though this helped publishers stay in control of the way ads were served and website performance, it also meant dealing with massive amounts of advertiser campaigns.
The solution? Third party ad servers.
This technology is responsible for pushing ads to the publisher’s website. Now, ads are housed within a third party ad server that takes care of managing, tracking, serving and analyzing the results of ad campaigns. Not only does it streamline the process, it also provides a more efficient way to study campaign performance data.
Publishers went from dealing with advertisers one-on-one, to dealing with multiple advertisers at once through ad exchanges. As the capacity to work with more advertisers increased, a process emerged to help publishers fill available ad space. This process is known as the programmatic waterfall.
Programmatic waterfalls and daisy chains
Should the first ad exchange pass on the inventory, it goes on to the next (also called a passback), and then the next, and the next. If no one buys the ad above or at the publisher’s floor price, a house ad is served.
Sounds great in theory, right?
The problem with this model is that though every publisher has their superstar players, sometimes it’s the new kid who wins the game, but an ad exchange with lower yield averages is usually lower on the waterfall list, there are instances where they will pay more for the ad space and generate more clicks for a particular set of impressions. When this inconsistency in waterfall logic occurs, publishers are leaving money on the table.
As programmatic media buying technology continued to evolve, real-time bidding (RTB) was introduced to accelerate the ad buying process. RTB runs through ad exchanges in which the ad exchange sends impressions for sale to advertisers in their network. An auction between the advertisers determines who is willing to pay the highest price. It is a faster process where publishers can make more money and advertisers get the impressions that matter to them, but it still follows the waterfall structure where one ad exchange is contacted at a time.
The technology giant Google identified an opportunity to evolve the process for publishers and advertisers, and created two companies that are keystones in the advertising industry to this day.
Google’s DoubleClick solution
Google’s DoubleClick subsidiary is known for two major properties: DoubleClick for Publishers (DFP) and DoubleClick Ad Exchange (AdX). Publishers who join Google’s DFP program gain access to a massive amount of advertisers who participate in the other branch of the DoubleClick suite: AdX. DFP provides publishers, especially smaller publishers who can’t access larger ad exchanges, with more potential buyers which in turn increases revenue opportunities.
Sounds like a great thing, right?
Not everyone was a fan of Google’s DoubleClick program. Independent exchanges who refused to join Google’s AdX network suffered in the shadow of the technology giant as more publishers (big and small) signed up for DFP. Without an existing partnership with a publisher, advertisers outside of AdX were bumped even further down the waterfall despite the possibility that those advertisers would spend more money on inventory than a direct sold campaign or AdX.
Not to mention, independent ad exchanges started to attach the word “monopoly” to Google’s DoubleClick program.
According to a DigiDay article, “Google had an informational advantage to buying the best impressions, and the informational advantage came from the fact that they own the ad server.” Since Google housed all the information, they could see what other ad exchanges were willing to bid first and then bid a penny more to beat all of the competition.
Pretty soon, people started to realize that Google had an unfair advantage. Independent exchanges knew there was no competing with Google. As the smaller kids on the playground, they had to find a way to get out the door before Google was dismissed from class or they’d never get their chance on the tire swing.
Their shortcut? Header Bidding.
Header Bidding: Flattening Waterfalls and Stomping on Daisy Chains
Rival exchanges made deals with publishers looking for ways to further maximize their inventory yield on top of direct sold campaigns and Google’s DFP. These exchanges placed code into the headers of publisher websites that allowed them to see ad inventory and the characteristics of the customer on the other end before Google had a chance to nab it. Essentially, they went from the end of the line to being a part of the circle when it came to seeing and bidding for desirable impressions.
In response to complaints that they were monopolizing ad tech services, Google made the decision to broaden their dynamic allocation to include rival exchanges in 2016. Everyone would see available inventory and bid in real-time accordingly.
Despite this, header continues to grow in popularity and implementation for publishers seeking to maximize yield.
How does header bidding work?
Within header bidding, publishers have two ways of securing demand: Real-time bidding (RTB) demand and managed demand. RTB header bidding consists of a publisher’s ad server contacting an ad exchange who then sends bid requests to third party advertisers. Though this often results in a higher fill rate, it also causes a slower response time. With managed demand header bidding, the ad exchange manages all advertiser campaigns directly, which results in a faster response time. Often, publishers use a combination of both to maximize CPM and fill rate.
Some say that header bidding flattens the waterfall, but this doesn’t mean publishers have abandoned the programmatic waterfall entirely. Many still use waterfall logic and run header bidding on a percentage of their total inventory. Publishers often allow direct sold campaigns first access to impressions, and then open it up to all other monetization channels afterwards.
The benefits of using header bidding
Although Header Bidding mostly benefits publishers, buyers and users also benefit from this system of ad management, too. For buyers, Header Bidding evens the playing field because everyone (regardless of size) bids at the same time. For users, it means ads that are better targeted to their interests and less latency.
Under the traditional waterfall structure, publishers had to run down a list of buyers which resulted in longer page load times (also known as latency). In the world of high-speed internet and microwaves, if there’s one thing people hate it’s waiting. If a page took too long to load because the publisher’s ad server was desperately running down a list of buyers, the user may exit out of the page before the process is complete (which means money and opportunity lost for the publisher.)
If header bidding is so great, why isn’t everyone using it?
Header Bidding Implementation
The secret sauce that makes header bidding work is ad server configuration. This includes a series of scripts and tags that are installed on a publisher’s website that records advertiser bids on the publisher’s ad server.
Header bidding doesn’t really work if you only have one partner. In order to optimize inventory performance, publishers need to work with at least 3-4 header bidding partners. The thing is, each of these partners require their own ad server tag to identify who they are, what they are willing to bid, and the creative assets (the ad) that will be served if they win the bid.
Furthermore, a publisher has to put together line items (more codes and scripts) that regulates pricing for particular ad space. The publisher’s ad server connects advertiser tag information to line item targeting to figure out who the highest bidder is. There are two ways to configure line items: exact granularity and price buckets.
Price Buckets and Header Bidding Granularity
Line items tell a publisher’s ad server how much money a bid is worth and are structured as key-values.
Exact granularity, also known as penny granularity, can capture bid amounts without rounding down. For example, if a bidder says they’ll pay $2.78 for an impression, the publisher will receive $2.78 because their line items are built to handle bids at a penny level. However, this requires a lot of work. Each penny requires its own line item, so if a publisher wants to build line items from $0-$10 to capture bids, 1,000 line items would need to be built.
Price buckets make line item setup easier because price ranges are utilized vs. penny granularity. Publishers can group bids in dollar, 10 cent, or nickel increments. In this case, a bid of $2.07 or $2.01 would be rounded down to $2.00. This option can leave money on the table, but it also means creating fewer line items.
Header Bidding Wrappers
When header bidding was just getting started, publishers had to manage all of their partners manually in terms of adding them to their header bidding ecosystem.
Header bidding wrappers are a powerful organizational tool that allow publishers to add/remove partners and collect data in a central location. It provides rules and structure for the programmatic auctions that occur and keeps things from getting crazy. Without a header bidding wrapper, publishers are looking at dealing with tons of code and a difficult set up to add in all of their advertising partners.
Currently, there are two types of header bidding wrappers: managed and unmanaged. Unmanaged header bidding wrappers simply provide the script (or code) needed to start using header bidding on a site. Prebid.js is a well-known open source script, but it does not offer much support in deploying, configuring, or managing header bidding. Managed header bidding wrappers, also called containers, are typically backed by a company that helps a publisher get header bidding setup and provides some form of reporting (like PubWise!) When you’re evaluating a managed header bidding wrapper to use, you’ll want to see if they provide robust analytics, how much of the setup process they manage, and how they choose the ad networks they work with.
Fortunately, the immense value that header bidding presents outweighs the difficult start up, and there are companies who specialize in Header Bidding implementation and reporting. At PubWise, we’ve completely automated header bidding setup so publishers never have to touch a line item.
Header Bidding Reporting
Revenue- This metric tells you how much money you’re making at a holistic level across all partners. However, it’s equally important to keep track of your partners on an individual level, too.
Bids won – Bids won shows you the total number of requests won. It also tells you which bids won, which partners are consistently winning.
Bid Discrepancies– Discrepancies in bidding styles among partners can result in publishers missing out on money.
Bid requests – The number of requests made to each vendor every time an ad is displayed. It also shows you what percentage of ads each vendor is taking.
Participation rate- Participation rate is the percentage that any one demand partner bids on inventory. As a publisher, you want this number to be high.
Won eCPMs by ad unit and line item – Won eCPMs represents a completed bid when a vendor pays for a specific impression — essentially showing how much you made from a specific impression by the thousands.
3rd party tracking
Demand partner variance – This specific metric keeps track of the differences between what the demand partner says they bid and what your own system says.
Latency reporting – On one hand you need to identify if demand partners are bidding late. On the other hand you need to identify if there is money left on the table just outside the current latency settings.
For more information about what to track and why it’s important, check out our post: Header Bidding Reporting – What You Need to be Tracking.
The Future of Header Bidding
There is no doubt that header bidding will continue to grow in popularity in the years to come. Even today, it’s a constantly evolving landscape that affords new opportunities for revenue growth and insight.
In addition to making things easier, we also make sure that things are done right. We don’t believe in hidden bidding or manipulated revenue shares, because we know the damage it can cause to a publisher’s revenue. There’s no preferential treatment towards certain ad networks, and our software makes sure the highest bid wins. We’re committed to fair auctions and complete transparency because, as publishers, it’s important to us.
So, whether you’re thinking about taking the plunge into header bidding or you’re desperately looking for a way to make sense of the tangled mess, we got your back. If you have any questions about our header bidding guide or header bidding in general, get in contact with us on Twitter @pubwiseio or email firstname.lastname@example.org.