Measuring bidder's incremental value with Prebid Analytics
If your new header bidding partner generates $18 000 in revenue, how much would your total revenue increase by?
Most of the publishers can't answer this and other similar questions because:
  • Every bidder you add comes at the expense of others;
  • Some or most of the revenue generated by a bidder you want to remove will be distributed among other bidders;
  • Without measuring the gap between winning bids and second-highest bids you can't measure your wrapper's bid density.
The lack of understanding how your header auction functions diminishes your ability to control its performance. When you lose control, your revenue depends on many external factors but not your work. Moreover, decisions dictated by incomplete revenue data may hurt your business and bottom line.
A quick example:
Judging by the revenue spread across bidders on, it's obvious that there are two clear leaders and two outsiders. It seems that if we remove any of the underperforming SSPs, a publisher wouldn't lose much revenue. However, the deeper understanding of's header auction may uncover the game-changing details.
Turns out the top-performing SSP often wins because it outbids competitors by a couple of cents. If it stopped working for whatever reason, the revenue would be distributed across other bidders and there wouldn't be a drop in the performance.
On the other hand, the worst-performing SSP quickly responds to almost every ad call/request but always bids around $0.80. Despite winning rarely, it gets impressions for ad requests other bidders do not respond to, which means that it generates unique revenue that wouldn't be compensated by other SSPs in the ad stack.
The variety of ad performance stats allows publishers to get a deeper understanding of their ad auctions and get control over their business performance.
How to measure bidder's incremental revenue in Roxot Prebid Analytics
We introduce a new widget on the Site Dashboard that allows you to:
  • Evaluate bidder's contribution to your total revenue
  • Measure how much revenue a bidder actually adds to your bottom line
  • Forecast a possible revenue drop if you remove a certain bidder from your header
  • Understand the level of competition and bid density in your prebid.js
The Incremental Revenue metric is site-specific for every bidder meaning that it may differ from site to site.
Roxot Prebid Analytics collects both the winning bid and the second-highest bid for every header bidding auction. This way the system can recreate an auction assuming the winning bid didn't participate.
Incremental Revenue
The gap between the winning bid and the second-highest bid represents the winning bidder's incremental revenue. For example, when the Incremental Value widget indicates that a bidder's incremental revenue equals $2000, the sum of the differences between the bidder's winning bids and the second-highest bids is $2000. As a result, you assume that you would have lost no more than $2000 in total revenue if the bidder had been removed and hadn't participated in these auctions.
Unaffected Revenue
As we mentioned above, you would lose less revenue then total revenue bidder generated if you decide to remove it from your header. Some of the yield will be distributed across other SSPs in your ad stack. The Unaffected Revenue metric represents exactly the revenue that would have been generated by other bidders if a bidder had been removed.
On the visual explanation above SSP1 wins the auction with the $1 bid. Hence, the revenue from this auction is $1 (in reality its $1/1000 as bids are in CPM but we don't convert the bid price for the convenience of this example.) In case SSP1 hadn't participated in the auction, the second-highest bid of $0.8 would have won and the publisher would have gotten $0.8 in revenue. Thus, $0.8 is Unaffected Revenue and $0.2 ($1 - $0.8) is the incremental value SSP1 generated.
Incremental value analysis in action
When evaluating bidder's incremental value, consider the following:
  • Total bidder's contribution to your total revenue
  • Amount of bidder's total revenue that is incremental
  • Percentage of bidder's incremental revenue from total bidder's revenue
  • Percentage of bidder's incremental revenue from total revenue generated by ALL bidders
Less than 10% of gross header bidding revenue & Less than 50% incremental revenue
As we mentioned above, you would lose less revenue then total revenue bidder generated if you decide to remove it from your header. Some of the yield will be distributed across other SSPs in your ad stack. The Unaffected Revenue metric represents exactly the revenue that would have been generated by other bidders if a bidder had been removed.
Even though publishers should pursue the even revenue spread across their bidders, it's common to have 2-3 partners generating more than 80% of revenue and a long tail of smaller SSPs generating the rest. However, the low-performing bidders complicate your payments management, increase website latency, and may negatively affect the total performance of your prebid.js.
Reviewing & optimizing your mix of SSPs is an effective exercise that helps find the balance between the quantity and quality. Now, the new Incremental Value widget will help you confidently identify bidders that you can painlessly remove or swap with other SSPs.
On the example above there are two low performing bidders - SSP 5 & SSP 8. They generate 2% and 1.6% of total revenue respectively. When we consider only total revenue, it seems correct to remove SSP 8 as it contributes less to the bottom line. However, only 43% of SSP 5's revenue is incremental, which means we would lose approx. 0.8% of total revenue if remove SSP 5 from the header. In turn, SSP 8 generates 56% of incremental revenue, 0.9% of total revenue on the site. As a result, it makes more sense to remove SSP 5 than SSP 8.
Despite being comparatively low, the numbers show that incremental revenue and its relation to the total revenue should be the primary metrics for in-depth bidder analysis. The more a bidder contributes to your bottom line, the more important the incremental value analysis gets.
More than 30% of gross header bidding revenue & More than 60% incremental revenue
When your bidder's incremental revenue is significantly higher than 50% of its total revenue, you can be sure it connects you to the unique demand and plays an important role in your monetization strategy. But high incremental revenue should make you cautious - what if a bidder simply don't have enough competition in the header? The situation reminds how waterfall ad stack works: a demand partner has exclusive rights to buy publisher's inventory, dictates the prices, and increases its own margins. That might be especially true if multiple bidders demonstrate incremental revenue higher than 70%.
On the example above all four SSPs have >90% Incremental Revenue. The low bid rate (how often a bidder responds to an ad request with a bid) and decent Fill Rate (how often there's at least one bid in the auction) indicate that all 4 bidders rarely compete for impressions simultaneously. Thus, the gap between the winning bid and the second-highest bid (if there is one) is huge in every auction. This setup needs immediate attention - a publisher should consider adding more top-tier partners to make the existing SSPs work harder for each impression.
Understanding bidder's incremental value is a complex task as many interconnected variables are involved. The Incremental Value widget is an evaluation of how much revenue a bidder generates on top of other bidders offer in your header bidding auctions. The widget doesn't take into consideration the bidders' influence on latency and functionality of prebid.js.
Getting started
Request a demo and one of our Publisher Relationship Managers will give you access to Prebid Analytics and walk you through the product.
Made on