“The Sell Sider” is written by members of the media community and contains fresh ideas on the digital revolution in media.
Today’s column is written by Ian Trider, VP of RTB platform operations at Centro.
SupplyChain Object, along with its companion Sellers.json, are critical tools for enhancing transparency in the programmatic ecosystem. They enable ad platforms and buyers to reduce fraud exposure risk and improve supply path optimization. It’s been over a year since their release, and adoption is now widespread. The data they produce provides deep insight about the state of the ecosystem.
More specifically, SupplyChain allows bid requests to contain information about the sequence of businesses that an impression passes through. Put another way, it documents the chain of who is receiving money for an impression. Its data enhances what DSPs would learn from the sellers.json files that function as directories of businesses from whom advertising systems (exchanges, SSPs, networks, etc.) receive supply.
SupplyChain and sellers.json work together to clearly show which businesses and advertising systems the impressions pass through on their way to a DSP. Because every major exchange and SSP now publishes a sellers.json file (450 distinct advertising systems have a sellers.json file in place), the next level of ad tech ecosystem transparency requires larger adoption of SupplyChain. SupplyChain provides a lot more context in the common behaviors of sellers.
Excessive resale and duplication
SupplyChain data helps illuminate the extent of reselling and duplication in the ecosystem. In my company’s dataset, we observe 142 intermediaries whose supply is resold by a dozen or more other intermediaries. In many cases, this supply is then further resold by other intermediaries. We see some SupplyChain data that shows three, four and sometimes even five hops from the publisher to the DSP.
There are valid cases for intermediaries in the ecosystem. For example, there are businesses that specialize in providing small publishers access to major exchanges, which creates value. However, other cases, especially with long supply chains, introduce fraud risk and waste. Inventory that a DSP thinks it has blocked leaks back in through indirect paths. Further, the DSP’s customers pay more ad tech tax for inventory that could have been purchased through existing direct integrations.
The data also reveals which advertising systems predominantly have direct publisher relationships or simply resell other intermediaries. We observe that there are at least 38 advertising systems where the majority of their supply relationships are other intermediaries – and several who appear to have few or no direct publisher relationships at all!
Lies, errors, and omissions
The information in Sellers.json and SupplyChain needs to be more complete and more accurate.
The challenges of sellers.json are well known. It’s common to see numerous confidential records in sellers.json files. There are also cases where the seller type or domain declarations in sellers.json seem incorrect, and it’s difficult to parse which are honest errors or intentional misrepresentation. And data quality is essential to use SupplyChain effectively.
Because of these challenges, SupplyChain data is not yet as complete or accurate as it should be. Some intermediaries are not supplying the data yet, even when it is supported by exchanges. Some bid requests are marked as having a complete SupplyChain even when they don’t. And there are other cases of missing or omitted hops in the supply path.
We expect these problems, because the data is self-declared. However, together these two specifications make it more difficult for bad actors to present a cohesive fiction. They are part of a complete toolset for inventory quality practitioners and are not magic cure-alls. My company audits and bans supply from those who lie to us. Consumers of the data should expect they need to cross-reference with other data available to them. The IAB Tech Lab FAQ gives examples of how to do so.
We encourage all DSPs to require adoption of both sellers.json and SupplyChain data, and refuse to transact with parties who do not provide it or provide untruthful information. We also encourage exchanges to require intermediary partners to have the specs in place, with the data withstanding scrutiny. Further, exchanges could make use of the data themselves to ensure compliance with their own policies.
The continued adoption of these specs is important for a clean and well-lit ecosystem. Although these are not silver bullet solutions, sellers.json and SupplyChain give us far more insight than ever before.