Roxot blog
Avoiding Header Bidding Loopholes
with Bid Adjustments
with Bid Adjustments
The ad server or header bidding wrapper decides who won the auction, but in the end, demand partners decide what to pay publishers. Whatever the reason for a discrepancy is - publishers should minimize the effects on revenue they are paid.
With the introduction and growing adoption of header bidding, the rules of the ad tech game are rapidly changing. Publishers who have adopted this method of integrating demand partners are reporting higher yields as a result of increased competition for their inventory. Transparency in ad auction mechanics - what lies at the core of popular open source header bidding wrappers - has empowered publishers to make data-driven decisions and strengthen their position in the market. However, being an immature technology with complex implementation, header bidding has uncovered loopholes that can be used - intentionally or inadvertently - against publisher's interests.
Besides following technical requirements when writing an adapter for a wrapper (such as prebid.js), demand partners do not have to adapt their bid responses to the format publishers are eager to work with. With bids in different formats, prebid or ad server auctions might be conducted incorrectly - lower bids would win impressions and publishers would lose revenue. While converting bids from one currency to another is more or less straightforward, reflecting networks' take rates, differences between gross and net bids, and other discrepancies is tricky. Demand partners might be vague in explaining what bids they are sending to your header; they might change the bid format without notifying you, or increase their take rate.
As the competition grows more and more fierce with every new adapter introduced, lack of unification in header bidding may be used to win more impressions without reducing take rates. Manipulating ad auctions is as simple as demand partners using gross bids instead of net.
If making a call for standardization and unification in header bidding is to be a long-term effort, the first step is as simple as utilizing ongoing maintenance and bid adjustments to increase publisher's revenue right now. Managing the discrepancy between demand partner, ad server, and client-side data is the most effective way to make data-driven bid adjustments and directly influence publishers revenue. Besides reducing bids that are known to be gross, publishers should adjust the bids of all partners who have high revenue discrepancy with ad server and client-side data even if these partners use net bids. This requires frequent data collection from multiple sources, hours of analytical work, and manual config updates. Despite resource investments, bid adjusting is an optimization practice that positively influences your bottom line. Putting aside the possibility that demand partners can intentionally manipulate ad auction results, ongoing bid adjustment can help publishers minimize negative effects of the rejected due to viewability or fraud issues impressions on their revenue. As a result, header bidding auctions become more efficient as bids that make you the most money win more often.
The ad server or header bidding wrapper decides who won the auction, but in the end, demand partners decide what to pay publishers. Technical issues, fraud traffic, view-ability, intentional bid manipulation - whatever the reason for a discrepancy is - publishers should minimize the effects on revenue they are paid. With header bidding data being more accessible and reliable, publishers can now adjust bids with a greater level of confidence and significantly improve their bottom line.
Besides following technical requirements when writing an adapter for a wrapper (such as prebid.js), demand partners do not have to adapt their bid responses to the format publishers are eager to work with. With bids in different formats, prebid or ad server auctions might be conducted incorrectly - lower bids would win impressions and publishers would lose revenue. While converting bids from one currency to another is more or less straightforward, reflecting networks' take rates, differences between gross and net bids, and other discrepancies is tricky. Demand partners might be vague in explaining what bids they are sending to your header; they might change the bid format without notifying you, or increase their take rate.
As the competition grows more and more fierce with every new adapter introduced, lack of unification in header bidding may be used to win more impressions without reducing take rates. Manipulating ad auctions is as simple as demand partners using gross bids instead of net.
If making a call for standardization and unification in header bidding is to be a long-term effort, the first step is as simple as utilizing ongoing maintenance and bid adjustments to increase publisher's revenue right now. Managing the discrepancy between demand partner, ad server, and client-side data is the most effective way to make data-driven bid adjustments and directly influence publishers revenue. Besides reducing bids that are known to be gross, publishers should adjust the bids of all partners who have high revenue discrepancy with ad server and client-side data even if these partners use net bids. This requires frequent data collection from multiple sources, hours of analytical work, and manual config updates. Despite resource investments, bid adjusting is an optimization practice that positively influences your bottom line. Putting aside the possibility that demand partners can intentionally manipulate ad auction results, ongoing bid adjustment can help publishers minimize negative effects of the rejected due to viewability or fraud issues impressions on their revenue. As a result, header bidding auctions become more efficient as bids that make you the most money win more often.
The ad server or header bidding wrapper decides who won the auction, but in the end, demand partners decide what to pay publishers. Technical issues, fraud traffic, view-ability, intentional bid manipulation - whatever the reason for a discrepancy is - publishers should minimize the effects on revenue they are paid. With header bidding data being more accessible and reliable, publishers can now adjust bids with a greater level of confidence and significantly improve their bottom line.
