“The Sell Sider” is a column written by the sell side of the digital media community.
Today’s column is written by Rotem Shaul, CEO at Primis.
Since the coronavirus pandemic started, the online advertising ecosystem has experienced two major trends: higher traffic and lower budgets.
More supply and less demand has forced many publishers to lower floor prices. However, many do not realize that lowering floor prices may actually increase their invalid traffic (IVT) rates and lower the viewability rates that the supply-side platforms (SSPs) measure, harming their overall performance.
This is happening because there is a discrepancy between the tools used by SSPs and demand-side platforms (DSPs) to measure IVT and viewability.
The tools
Pre-bid scanning is used for prevention. Buyer uses third-party providers to only buy impressions that were verified as viewable and IVT-free. To do that, buyers scan every received ad call and only buy those that are safe.
Buyers use post-bid scanning for reviewal purposes. Buyers use a third-party provider to scan traffic after purchase. Bad traffic may get into the campaign, but the buyer is able to monitor and remove bad sellers and domains to keep the IVT and viewability rates down. To run post-bid verification efficiently, buyers scan approximately 20% of the traffic, as opposed to 100% with pre-bid.
The discrepancy
Most SSPs use post-bid scanning to understand the quality of the traffic publishers are sending them. For buyers such as DSPs and brands, there is usually a mixture of pre-bid scanning, post-bid scanning or even no scanning at all. Verification is expensive and, the higher the campaign’s CPM, it is more likely that buyers can afford to use verification.
Here is how this discrepancy occurs. Say a publisher has 5% IVT rates. It sends the traffic to the SSPs with a $5 CPM floor, and the SSPs send the traffic to two buyers:
• Buyer A, which is willing to bid $10 CPM, uses pre-bid IVT detection. It bids on 95% of inventory, so every clean impression is then sold to Buyer A.
• Buyer B, which is only willing to bid $0.2 CPM, does not bid because the floor is too high.
• Post-bid scanning shows the SSP and publisher a 95% fill rate for an average price of $10. The IVT rates will appear to be zero, because the SSP only scanned the sold impressions for IVT.
In another scenario with the same publisher, buyers and SSP, the publisher has reduced its price floors to increase its fill rate:
• Buyer A, which is willing to bid $10 CPM, uses pre-bid IVT detection. It bids on the 95% of the inventory that is clean and wins every clean impression.
• Buyer B bids against all inventory once the price floors were reduced. When the traffic is clean, it loses to Buyer A, which had a better CPM. When the traffic is not clean, Buyer B wins the auction, buying the remaining 5%, or all impressions that Buyer A did not bid on because of pre-bid IVT detection.
• The SSP and publisher now see a 100% fill rate for an average price of $9.51. However, the IVT rates increased to 5%, because now all of the impressions – both the bad and the good – were sold.
• Buyer B scanned the traffic post-bid and now complains to the SSP that it sees 100% IVT.
The main takeaway here is that we have a huge shift due to the changing of the floors. The SSP used to have a publisher with 0% IVT rates. But once the publisher reduced its price floors, the IVT jumped to 5%. On top of that, the SSP is now getting calls from Buyer B about 100% IVT rates backed by MRC reports.
Even more distressing is that the revenue upside was so small. The publisher used to do 95% fill on $10 CPM. It now does 100% fill on $9.51 CPM, an increase in revenue of just 0.1%, made possible by a disproportionately huge jump in IVT.
It’s important to acknowledge here that this publisher is not bad and isn’t trying to engage in any fraudulent activity. This is just a regular publisher that never knew it had 5% IVT. The buyer breakdown, heavily affected by floor prices, affected the way the market perceives the publisher, when measured post-impression.
Floors affect how SSPs and other companies that use post-impression measurement see traffic. Pre-bid prevention and post-impression measurement are just two different methods that are bound to create some discrepancy.
Usually, when everything is stable, this isn’t an issue. I am sharing this now because I assume that publishers and SSPs are reducing floors dramatically due to COVID-19. Some will probably see an increase in IVT and not understand why. Many won’t connect the dots between reducing floors and IVT, and they may end up blaming the MRC or another innocent party.
Follow Primis (@PrimisLtd) and AdExchanger (@adexchanger) on Twitter.
