What Is Waterfall Bidding?

Waterfall bidding – also known as “waterfall mediation,” “waterfall advertising,” or an “ad waterfall” – is a programmatic method of buying and selling ad inventory. For years, it has been the standard ad monetization setup for app publishers.

The method got its name from ad requests cascading down from one ad source to the next (in what looks like a waterfall auction) until it reaches the ad network that can fill the request.
Key Takeaways
  • Waterfall mediation calls demand sources one by one, typically ranking them by highest CPM
  • Hybrid mediation is where waterfall bidding and in-app bidding are simultaneously in play
  • Due to compromised revenue yields, many publishers are moving away from waterfall bidding toward in-app bidding

How Does Waterfall Bidding Work?

What does waterfall mean in advertising? Waterfall mediation allows app publishers to sell their ad inventory to a competing pool of demand partners – also known as demand sources (DSPs) or ad networks. Publishers use these multiple partners to fill their ad space and generate revenue.

Whenever an ad impression becomes available, a publisher’s waterfall mediation solution manually calls these demand sources one by one, in a specific fixed order. This order is typically ranked based on historical data – from highest to lowest eCPMs. If the first-priority partner (the buyer) rejects the bid, the request will go to the next partner in line. 

A publisher might reject the ad request for reasons such as the network’s ad not fulfilling the requirements provided by the publisher in the request – such as floor price and geo.

diagram showing how waterfall bidding works

This process continues trickling down from ad source to ad source until one fills the impression. This is the premise of waterfall bidding and is how it got its name.

What Is Hybrid Mediation?

There has been industry pressure to move away from waterfall bidding due to its shortcomings in maximizing revenue from ad requests. 

Some publishers for this reason run an auction in parallel with waterfall bidding. This is known as hybrid mediation and is when waterfall mediation and in-app bidding are simultaneously deployed. 

The concept behind the hybrid model is to help publishers boost their revenue by integrating in-app bidding into various positions in the ad waterfall. Instead of only receiving bids through waterfall bidding, publishers can make impressions available to as many buyers as possible. This means that there is a higher chance of filling ad space and allows publishers to increase revenue generated from their ad inventory.

diagram explaining hybrid mediation in programmatic advertising

Real-Time Bidding vs Waterfall Bidding

Unlike waterfall bidding, which is based on fixed costs and historical data, the costs in in-app bidding (also known as “real-time bidding” or “header bidding”) are based on real-time data. 

This data reveals how much demand sources are bidding on ads. 

Since this bidding occurs outside of the waterfall, the auction is unbiased. Bids are placed in real time and the highest bid wins. There is no fixed sequence that prevents ads further down in the chain from being called and the publisher subsequently losing out on revenue from bidding partners willing to pay more for inventory.

In this form of mediation, publishers tend to prioritize the demand partners and ad networks they work with. This means that it’s not always the highest bid that wins but the ad network that has manually been prioritized by the waterfall solution and has the highest historical eCPMs.

Why Are More Publishers Turning Away from Waterfall Mediation?

Waterfall mediation is often problematic when it comes to publisher revenue and also ad networks wanting to pay for inventory. 

Some monetization managers are, however, reluctant to shift all their solutions to purely in-app bidding platforms. Some experts believe that a waterfall auction gives the partially false perception that monetization managers are able to remain relevant through maintaining manual waterfall placements. 

There are other reasons for this hesitancy – which is why publishers tend to opt for hybrid mediation solutions. However, it has been proven that waterfall auctions restrict a publisher’s performance for the following reasons:

  • Waterfall bidding relies on historical performance data and therefore limits revenue yield. This is because app publishers often end up underselling their inventory. For example, a lower ranking network might be willing to bid higher to get the ad inventory, but is unable to do so because a top-ranking network has won the bid. 
  • There is limited competition. With this mediation model, publishers can only call one ad buyer at a time. This makes the bidding process less transparent and open. 
  • Latency is compromised. Setting up waterfall bidding is a tedious manual task that increases the time it takes for ad inventory to reach a demand source willing to pay for inventory. In the end, publishers can miss out on efficiency and revenue. Especially, in some cases where inventory may remain unsold completely. 

Conclusion

App publishers have used waterfall bidding as an ad monetization solution in the adtech space for several years. 

However, waterfall bidding is now widely considered as an nontransparent mediation solution that restricts competition between bids and prevents app publishers from maximizing their revenue and earning premium.

FAQs

What Is Waterfall Bidding?

Waterfall bidding is a programmatic method of buying ads. For years, it has been the standard ad monetization solution, but it’s heavily criticized by many as a nontransparent mediation method that restricts competition and prevents app publishers from maximizing their revenue.

How Does Waterfall Bidding Work?

Whenever an ad impression becomes available, the publisher’s waterfall mediation solution calls demand sources one by one, ranking from historical highest to lowest eCPMs.

Why Is Waterfall Bidding Not Good?

Waterfall mediation constitutes many pitfalls regarding publisher revenue and also for ad networks wanting to pay for inventory. This is due to reliance on historical performance data. Furthermore, competition and transparency is limited, and latency is compromised.