OpenX has turned the growing number of log-level data requests from buyers into a product that simplifies the process, called Bidding Intelligence Data Set (BIDS).
“Our intent was to make what used to be a wonky and difficult process into as streamlined, productized and standardized a process as possible, to make our customers’ lives easier,” said Joey Leichman, VP of buyer development at OpenX.
BIDS uses Snowflake Data Marketplace, a popular cloud tool for data warehousing, querying and sharing, in order to make the data accessible to buyers. On the buy side, The Trade Desk and Beeswax have both recently plugged into Snowflake.
OpenX populated Snowflake with highly detailed information about each individual bid. This info includes descriptive details such as the seller ID, site ID, ad unit ID, page UR and the DSP used. It also contains the brand ID, information about pricing (for example, the minimum bid to win in OpenX’s auction) and identity information (for instance, whether OpenX matched a cookie or if it used other people-based IDs, such as LiveRamp’s IdentityLink).
Brands, agencies and holding companies have been clamoring for log-level data for a few reasons. At a most basic level, many buyers want to audit their supply chain and understand what they’re buying, where and at what price. Log-level data enables the most precise verification.
Other brands and agencies are taking log-level data a step further, and using it to optimize their buying, including doing supply-path optimization. If they can use log-level data to find high-performing pockets of supply – and avoid total duds – they can route more of their programmatic spend to the best places.
“We think that the core use case for log-level data is having a complete view of our ecosystem, and the control that comes from having the complete view – whether you’re an agency or a brand,” Leichman said.
Unfortunately, the process of getting that data has been manual, costly and time consuming. Because these files are so huge, they’re expensive to store and query, and they require engineers and data scientists to process. Developers had to do custom work to access and organize the data– both things that are turnkey in OpenX BIDS setup.
“In our view, it dramatically improves on some really bad pain points,” Leichman said. “The costs for log-level data are enormous for the provider, in terms of both human resources and structural engineering costs.”
OpenX’s setup with Snowflake should make costs “infinitesimal,” Leichman said. If a client is already a Snowflake customer – and many already are – they can use their credits with the company to access and use the data. Non-Snowflake customers can either set up an account or use OpenX’s credits. OpenX will charge customers essentially the same it pays Snowflake, Leichman said.
There’s already a long list of early adopters that want to transition to accessing logs via the Snowflake setup, and OpenX has scheduled three waves of customers to start accessing the data, Leichman said.
With better data portability and ease of use compared to previous log-level data setups, Leichman said, even more brands and agencies find the juice worth the squeeze to analyze log-level data.
