What Is Header Bidding? A Complete Guide for Publishers
Header bidding is one of the most significant changes to how publishers sell programmatic inventory, and it's now a standard part of most serious monetization setups. It's also one of the most frequently searched, and frequently misunderstood, terms in ad tech. This guide explains what header bidding is, how it works, the difference between its two main setups, and what it does and doesn't solve for publishers.
The Simple Definition
Header bidding is a technique that lets a publisher offer an ad impression to multiple demand sources at the same time, before the ad server makes its final decision. Each source submits a bid, and the highest bid then competes for the slot. The name comes from the fact that the code originally ran in the header section of a web page, so it could start the bidding process as early as possible.
The Problem It Was Built to Solve
Before header bidding, most programmatic inventory was sold through a waterfall. A publisher ranked its demand partners in a fixed order, and an impression was offered to the first partner. Only if that partner passed, or its bid didn't clear the floor, did the impression move to the next partner, and so on down the list.
This had two structural weaknesses:
- Lower ranked partners rarely got a fair chance. A partner further down the list might have been willing to pay the most for a particular impression, but it was never asked if someone above it had already filled the slot at a lower price.
- Rankings were based on historical averages, not live demand. The order reflected what partners had paid in the past, not what they would pay for this specific impression right now.
Header bidding addressed both by asking everyone at once and letting the actual bids decide.
How Header Bidding Works, Step by Step
Step 1: The page begins to load. The header bidding code, called a wrapper, runs as the page starts rendering.
Step 2: Bid requests go out simultaneously. The wrapper sends requests to all connected demand partners at the same moment.
Step 3: Partners respond with bids. Each partner evaluates the impression and returns a bid, within a time limit set by the publisher.
Step 4: The wrapper collects the bids. After the timeout, the wrapper gathers every bid it received and identifies the highest one.
Step 5: The result goes to the ad server. The winning header bid is passed to the publisher's ad server, where it competes against any direct sold campaigns and other eligible demand according to the publisher's rules.
Step 6: The winning ad is displayed. Whichever source wins in the ad server delivers the creative to the slot.
Client Side vs Server Side Header Bidding
There are two main ways to run the auction.
Client side header bidding runs the auction in the visitor's browser. The browser sends requests to each partner directly. This gives partners good access to browser level data, which can help with user matching and pricing, but each additional partner adds work for the browser and can slow the page down.
Server side header bidding moves the auction onto a server. The browser sends a single request, and the server handles the calls to all the demand partners. This reduces the load on the visitor's browser and generally helps page speed, but it can limit data matching between partners, which some buyers rely on to bid well.
Many publishers use a hybrid, running a few of their most important partners client side and the rest server side, to balance revenue and performance.
What Prebid Is
Prebid is an open source header bidding framework widely used across the industry. Rather than building a wrapper from scratch, many publishers use Prebid to connect to demand partners through standardized adapters, with the benefit of an open, community maintained approach and no single company controlling the auction logic.
Benefits of Header Bidding
- More competition for each impression. Multiple sources bid at the same time, which generally pushes effective prices up.
- Fairer access for demand partners. Partners aren't locked behind a fixed rank order.
- Better transparency. Publishers can see which partners bid and at what levels.
- Higher yield on primary inventory. For high demand impressions, simultaneous competition typically beats sequential calling.
Drawbacks and Limits
- Page speed cost. Each partner and each call adds latency, especially in client side setups, so timeouts and partner counts need careful tuning.
- Setup and maintenance effort. Wrappers, timeouts, partner adapters, and floors all need ongoing management, which usually takes technical resources.
- Diminishing returns from too many partners. Adding partners that overlap heavily with existing ones adds complexity without much extra competition.
- Duplicate paths to the same inventory. Running many overlapping integrations can create the kind of cluttered supply path that buyers increasingly filter out.
Best Practices
- Set timeouts based on real data rather than defaults, balancing capturing slower bids against page speed.
- Choose partners for genuine demand coverage of your traffic, not just to increase the count.
- Review which partners actually win impressions and remove the ones that don't.
- Keep your ads.txt file current as partners change.
- Monitor page speed alongside revenue after any change.
What Header Bidding Doesn't Solve
Header bidding improves competition for the inventory that has bidders. It does not create demand where none exists. When every connected partner passes on an impression, or no bid clears the floor, that impression still goes unfilled. Historically that leftover inventory has fallen back to an older static waterfall or gone unsold. This is why many publishers pair header bidding on their primary inventory with a dedicated real time layer for the remnant impressions it leaves behind, rather than assuming header bidding alone covers everything.
Frequently Asked Questions
Does header bidding replace the waterfall completely?
For primary inventory it largely replaces it, but many setups still keep some fallback layer for impressions header bidding doesn't fill.
Is header bidding worth it for small publishers?
It depends on traffic and available technical resources. Smaller publishers can benefit, but the setup and maintenance effort is real, and some start with simpler configurations before investing in a full wrapper.
Will header bidding slow my site down?
It can if it's configured carelessly. Reasonable timeouts, a sensible number of partners, and server side or hybrid setups all help keep page speed under control.
Can I use header bidding alongside Google Ad Manager?
Yes. Header bid results are typically passed into the ad server, where they compete with other demand under the publisher's rules.
Key Takeaways
- Header bidding lets publishers offer an impression to many demand sources at once, replacing the sequential, rank ordered waterfall for primary inventory.
- Client side setups run in the browser, server side setups run on a server, and many publishers use a hybrid to balance revenue and page speed.
- Benefits include stronger competition and transparency, while drawbacks include added latency and ongoing management effort.
- Header bidding doesn't fill impressions nobody wants to bid on, so remnant inventory still needs its own dedicated approach.