Header Bidding vs. Waterfall: What’s the Difference?

Header Bidding : A technical term from the field of online marketing and eCommerce.

With digital advertising, standards and ad tech evolve over time. While there can be significant differences between how programmatic auctions are run and how impressions are awarded to the highest bidder, most advances are favoring publishers these days.

Programmatic advertising has created easy ways for publishers to sell more advertising automatically without the high expense of direct selling. Publishers can also use programmatic auctions to fill space that might otherwise have gone unsold.

This article will detail and compare the two programmatic bidding options, the waterfall model and header bidding, including how they work and the pros and cons of each strategy in the programmatic ecosystem.

What is Header Bidding?

Header bidding is a programmatic advertising auction strategy that sends bid requests to multiple demand sources simultaneously. Publishers place some JavaScript code in the header code on a webpage (called a header bidding wrapper) to trigger real-time bidding across different ad exchanges. The ad unit is awarded to the highest bid programmatically. Learn more about the differences between real-time bidding and header bidding.

Also called advanced bidding or pre-bidding, header bidding allows for greater competition among advertisers, which benefits publishers by creating higher cost per thousand (CPM) rates for available ad units.

How Does Header Bidding Work?

Publishers can connect their websites to supply-side platforms (SSPs) and include information about visitors which in turn inform advertisers about the type of audience that is available.

Each time a visitor views a webpage that includes ad inventory, as the page loads, it triggers an ad call. The code in the header requests bids from several ad exchanges, which connect with demand-side platforms (DSPs) to solicit bids. If the ad inventory publishers are offering matches up with advertiser criteria for targeting in the demand source platform, bids are proffered and the highest bid wins the auction.

All of this happens in the blink of an eye, literally milliseconds, in the background as multiple demand sources are queries and ads are served.

Publishers can also incorporate several header bidding networks on their sites but each additional one can introduce some latency which may slow page load times and impact the user experience. The best header bidding solutions will include best practices to avoid slowdowns, such as asynchronous loading or “lazy loading.”

Client-Side vs. Server-Side Header Bidding

There are two different ways that header bidding can be implemented:

  • Client-side header bidding
  • Server-side header bidding

Client-side header bidding executes an ad auction within the user’s browser. Server-side header bidding products the auction on external servers rather than a publisher’s ad server. Server-side header bidding generally improves page load speed while client-side bidding offers more control and accurate data. It’s important that users understand how these technologies work and the key differences.

Client-sideServer-side
LatencyBecause auctions are initiated in the browser, it can impact page load and introduce latency, especially if users aren’t using a high-quality, high-speed connection.Auctions are initiated on servers that are designed to run auctions, which allows for faster processing. While there can be latency due to multiple demand sources, it’s minimal.
IntegrationRunning a client-side tech stack requires more upfront investment and resources if you do it yourself. Some vendors do provide managed client-side wrappers.Vendors provide the ad servers to handle ad requests, meaning publishers can implement the code and let ad tech handle the rest.
LimitationsThe number of bidders is capped by the browser, which has browser call limits. At the same time, each additional bidder adds latency, affecting page speed.Only a single call is placed to the auction server, which means publishers can add additional bidders without adding latency.
TransparencySince publishers control the auction code, they have complete control and visibility into the programmatic ecosystem, including bid and log-level data.There is less transparency in the bidding process as only a single ad call is sent and a winning bid is received. Line-item details are not shared.
RevenueCookie matching rates are higher. Advertisers can more accurately ID users because they do not need to be synced with an additional server. This can lead to higher bids.Cookie matching rates are lower as different SSPs and DSPs may use different methods of identification, increasing misidentification. As a result, revenue per impression may be lower.

Many publishers start using client-side header bidding, but struggle with page loading times and have to maintain the system. While the yield may be slightly better, the impact on the user experience has most publishers now favoring server-side header bidding.

If you are considering a header bidding partner, here are the 10 best header bidding network partners. One name that’s not on the list is Google AdSense. While perhaps the most user-friendly way for ad ops to start monetizing websites, it limits bids to a single demand partner (Google). In other words, it does not support header bidding across multiple demand sources, limiting ad revenue optimization.

What is a Programmatic Waterfall Model?

While header bidding passes add requests to multiple demand partners at the same time,  the programmatic waterfall model works sequentially. Called daisy-chaining, the bidding process follows a path from one ad network to the next until all of the impressions are sold. If no bidders are willing to pay the publisher’s inventory floor prices, ad units are then offered to the next demand source.

While the waterfall method works, it’s an inefficient way to conduct programmatic auctions. It can also significantly limit publisher revenue as floor prices decrease each time ad units are offered to different demand sources in sequences.

How Does the Programmatic Waterfall Work?

The waterfall bidding process is sometimes referred to as real-time bidding (RTB). After any direct deals publishers offer are filled, the remaining ad space is offered to RTB auctions one at a time until ads are filled.

The waterfall method offers two significant advantages to publishers. First, since auctions continue until the ad unit is filled, improving fill rates. However, it can also leave money on the table. If the first bidder meets the publisher price floor for CPMs, the bid is accepted. There may be other bidders willing to pay higher rates further down the line.

Publishers may also opt for the waterfall method because they can eliminate selling ad space entirely, selling all of their ad units to anyone willing to meet the minimum price thresholds.

Header Bidding vs. Waterfall: The Ultimate Comparison

Each method has its pros and cons, so we need to discuss the advantages and disadvantages of header bidding vs. waterfall. First, however, we should also talk about how both methods allow flexibility for publishers.

In both cases, publishers can choose which ad inventory to make available programmatically and how. This means publishers can still do direct deals by working with advertisers or ad agencies. Generally, this requires manual negotiations but results in higher overall rates for premium inventory.

Publishers can also opt to sell programmatic guaranteed or set up private marketplaces (PMPs) where select advertisers can bid. In both cases, deal terms and prices are set up in advance rather than using a real-time bidding process.

As you are considering your alternatives, you should also have an understanding of the differences between ad exchanges and ad networks.

Advantages and Disadvantages of a Waterfall Auction

Let’s start with the waterfall auction model.

Pros

  • A waterfall ad RTB auction allows publishers to sell off their remnant inventory.
  • Easy process to set up because it only requires pasting in a passback tag from one SSP or ad network.
  • Use a standard ad tag, which is ideal for in-app advertising.
  • Can produce high revenue opportunities by maximizing fill rates and reducing unsold inventory.
  • For first-time advertisers, there is often less competition.

Cons

  • Produces a lower yield for publishers because ad networks and SSPs are offered bids in sequence.
  • CPMs are sold at publisher price floor minimums.
  • As impressions are passed from one ad network (a passback), CPM price floors decrease until ad units are filled, resulting in less revenue.
  • Page latency can be a problem, especially if price floors are set too high, as ad impressions are offered to different platforms sequentially.
  • Page latency can also occur when demand partners or SSPs do not respond quickly enough, resulting in timeouts.
  • The sequential selling process has a high impression discrepancy, unlike header bidding, which is a single auction. 

Pros & Cons of Header Bidding Auctions

Now, let’s turn our attention to header bidding auctions.

Pros

  • Allows exposure to more advertisers. Because bids are solicited from multiple ad exchanges, ad networks, and SSPs, each platform has its own advertiser base.
  • Because bids are solicited simultaneously, competition increases, so there’s a significantly better chance to match with impressions at higher rates.
  • Users will see more impressions, meaning high bidders will reach a frequency cap, so exposing the inventory to more buyers as there’s an increase in demand of choice. 
  • Publishers realize better yield management and an increase in revenue due to multiple people bidding on the inventories. This also eliminates the passback process during an auction.
  • Reduces latency due to the simultaneous bidding process.

Cons

  • Does not offer as much transparency as third-party servers, meaning sometimes third parties prioritize their own advertisers.
  • Some bids can be duplicated due to advertisers often using several DSPs, while publishers might have dozens of SSPs.
  • You will need some sort of technical experience to set them up. Header bidding wrappers will need to be created and implemented, but not everyone will know how.
  • While reducing pass-backs that are common in the waterfall method, there can be latency, especially when using client-side header bidding.

Final Thoughts

As you can see, there are advantages and disadvantages to the waterfall method and header bidding. The waterfall method is easier to set up and process for beginners, but publishers that want to maximize their yield and achieve consistently higher CPMs favor header bidding technology.

Newor Media offers publishers the opportunity to maximize their revenue using header bidding without having to deal with all of the backend ad tech. Newor Media will evaluate your website and provide the header wrappers you need. Once they are installed on your site, everything else will happen automatically. Using real-time auctions and header bidding, the publisher’s inventory is optimized to make sure winning bids are processed quickly and ads are served.

Newor Media can also help you determine the right formats and ad placements, providing all of the ad tech you need.

If you are interested in seeing how much revenue you could be generating right now from your website, try the free Earnings Calculator to see for yourself. To learn more about how Newor Media can help you improve site monetization, contact us or apply today.