ConnectAd Chronicles


Why there is no such thing as a ‘full ad stack’

Let's be honest: approaching publishers and making them aware of us is part of our daily business. We love making new contacts, getting into conversations and exchanging ideas. In doing so, we always meet publishers at eye level, strive for a partnership-based relationship and take them and their feedback more than just seriously.

And yet there is this one thing that regularly makes us throw our hands up in horror. Specifically, it's a statement that we simply can't hear anymore. We are talking about the little sentence "Sorry, but our ad stack is already full". Because that's not only synonymous with "A cooperation with you is out of the question focus", it's also an excuse that we can't accept. Why? Because an ad stack is never really full – or at least it shouldn’t be considering that this great thing called header bidding has been around for quite some time now.

The big ‘header bidding vs waterfall’ discussion 

More specifically header bidding made it into the AdTech world back in 2014, went viral in 2015 and replaced the good old waterfall that has been an industry standard for many years. Now we have the year 2023 and ‘header bidding vs waterfall’ still is one of the most searched terms in AdTech. So today we are going to explain what a waterfall auction is, how header bidding works and what the differences are. Spoiler: The biggest difference between waterfall and header bidding probably is that the ad stack is never full for those who rely on the latter. … But: first things first!

Programmatic waterfall: A trip into the past  

Programmatic waterfall, also known as daisy-chaining or waterfall tags, is a technique for publishers to sell remnant ad inventory. In this process they pass their inventory from ad network to ad network in descending order of importance until all impressions are (hopefully) sold.

Inventory is passed from exchange to exchange 

The exchanges are usually ranked based on the average historic performance they have produced for the publisher. This means that ad networks with the best record of eCPM, fill rate, latency, etc. will beat the top. They will get first chance on further impressions from the same publisher. Unsold inventory is then passed on from one ad network to another in sequence until it is sold. Sounds easy? It is! Because of that and because this solution helps publishers to make sure that they aren’t left with any unsold ad space, the programmatic waterfall has been the method of choice for a long time. Nevertheless it is far from perfect.  

Waterfall auctions have some drawbacks

Why that though? First and foremost because of the process of “passbacks”. If publisher accept bids from various ad networks in turn, there is a lack of competition for ad impressions. And therefore the publishers receive a lower overall yield. Once the impressions are passed onto the ad network, CPM price floors decrease. Without giving everyone access to the auction from the very beginning, potential high-paying advertisers are cut out. This prevents them from placing higher bids for a certain piece of inventory that might have been of extreme value for them and ultimately reduces publishers' revenues as the prices end up being lower automatically. Another disadvantage of waterfall auctions is that the process takes very long. This results in page latency. As the ad call goes from ad network to ad network, page loading times can be affected, which can cause a negative user experience.

To summarize: Waterfalling has not been particularly efficient and so header bidding was invented out of necessity.

Header Bidding: Born to democratize the programmatic world 

It’s nothing new that necessity is the mother of invention, and header bidding is a pretty good example for that. So, when publishers and advertisers were losing revenue, were fed up with big players’ monopolies and in strong need for more efficiency, transparency and flexibility, header bidding was born. 

Fighting against almighty Google

One of, if not the largest company with monopoly position in the programmatic game back in the waterfall times was Google. So you could say that header bidding was invented especially to make life difficult for Google and the game a more equal playing field. To be more precise, it was about competing with Google'sAdX, which was able to preview auctions and win impressions they valued by bidding $ 0.01 more than the winning price. However with header bidding popping up this practice came to an abrupt end. 

Simultaneous bidding in real time for the win

All well and good, but you still wonder what the heck header bidding is and how it works? Let us tell you! Header bidding is an advanced programmatic technology where publishers offer their inventory to multiple ad exchanges at the same time before making calls to their adservers. By allowing multiple demand sources to bid on the same ad inventory atthe same time, publishers increase their yield and make more money.

… but back to the ‘header bidding vs waterfall’

You probably already guessed it – in our opinion, the disadvantages of waterfall outweigh header bidding. Yes, they are so to speak obvious. But to put it in a nutshell, the following reasons speak against waterfall auctions and for header bidding:


- The Waterfall contradicts the programmatic idea of fairness and that the highest bid should win – bidders are connected in sequence instead of simultaneously.

- The Waterfall produces a lower yield for publishers because ad networks and SSPs are offered bids in sequence.

- CPMs are sold at publisher price floor minimums in waterfall auctions.

- As impressions are passed from one ad network to another, CPM price floors decrease until ad units are filled, resulting in less revenue.

- Page latency can be a problem when waterfalling, 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 in the waterfall auction has a high impression discrepancy, unlike header bidding, which is a single auction. 

And as if this list of disadvantages alone isn't reason enough for any publisher to rely on header bidding or PreBid, it is also the case that publishers with a PreBid setup can maintain an almost infinite number of partnerships with bidders.

Long story short: A ‘full ad stack’ is a really bad excuse

As with every good story, we have now come full circle. And we are happy to repeat the thesis we set out at the beginning: An ad stack is never full – or at least it shouldn’t be.

It may be true that you reach your limits quite quickly with Waterfall, but with header bidding you can onboard many SSPs and thus increase your yield significantly! The most common way to use header bidding as a publisher is PreBid Client Side. If more than 10-15 bidders are installed, latency issues can also arise here, since processing power is "borrowed" from the user's end device. Publishers who want to integrate additional SSPs can than fall back on PreBidServer Side. Depending on the technical resources on the publisher side, there are hardly any limits to the number of bidders.

Whether the ad stack is full or not is therefore only a question of the setup. If your ad stack is indeed full, you should urgently rethink your setup. If you are not yet using PreBid, it is better to start today than tomorrow. Come on guys! It's 2023 – and good old waterfalling has had its day.

And please believe us when we say: You are not only doing us a favor, but primarily yourself. Because: Different SSPs have different strengths and PreBid helps you to continuously test which SSP can convince with performance. Don't stop testing after a while just because you feel you are well positioned. A well filled ad stack is always useful. You will feel the uplift. We promise!

How ConnectAd fits int here? Test us! We’re here to surprise with great results! Let’s talk.