Who needs another supply-side platform for streaming TV advertising?
No one, according to Michael Bassik. Which begs the question: Why did he decide to launch one? The company, called Starling, launched last month with roughly 20 employees after about a year in stealth mode.
“Our vision is ambitious,” Bassik told AdExchanger.
Starling sees an opportunity to address one of the biggest pain points facing marketers who want to buy streaming TV advertising: the lack of show-level data. Advertisers and agencies have been begging publishers and their partners for years to get the same content transparency they have for traditional TV ad buys.
Slowly but surely, media sellers, including Spectrum Reach and Amazon, have recently started rolling out more program-level data.
Starling’s pitch is to turn those program-level signals into something buyers can use to reach audiences across platforms at scale.
Pairing show titles with audience identifiers for targeting is like the “holy grail” for buyers, Bassik said.
Whatcha watching?
To get a hold of the show-level data that buyers want so badly, Starling works directly with connected TV publishers, including AMC, TelevisaUnivision, Paramount (sorry, Skydance) and Telly. Those signals include the content surrounding ads, genre, app and show title. Starling also has access to some audience and identity data from publishers it works with as well as from third-party audience data providers (which Starling declined to name).
Starling then creates custom segments that can help buyers reach relevant audiences based on what they’re watching.
For example, if an advertiser wants to reach viewers who watch specific morning news shows, Starling can aggregate that inventory and make it available to buy in private marketplace (PMP) deals, Bassik said, adding that PMPs are important for scale. To help with scale, Starling also uses an audience modeling methodology that expands a brand’s target audience to include similar viewers.
From the publisher point of view, deals backed by content signals typically see higher demand and better ad prices, Bassik added.
Riddle me this
Sounds like a no-brainer. But providing show-level data is far easier said than done.
Publishers differ in how much and what type of viewing data they’re willing to share. Some see viewing data as a competitive advantage, and then there’s also the privacy question.
Many publishers are reluctant to share viewership data because of the Video Privacy Protection Act (VPPA), a Blockbuster-era law passed in 1988 that prohibits video content providers from disclosing details about video consumption linked to an individual without consent. Nowadays, media companies interpret that law to mean they shouldn’t include show-level data alongside user IDs in a programmatic bid request.
“The VPPA has convinced certain publishers, especially larger media companies, that they should not be sharing content-level signals,” Bassik said. Although, that’s not the case for all publishers. Some platforms provide show-level information using precautionary measures, such as data obfuscation, modeling and/or aggregation.
So, how does Starling guarantee that its methodology doesn’t violate the VPPA? According to Bassik, Starling works with third-party providers for data modeling to ensure that data is used in a privacy-friendly manner. It also works closely with its publisher partners to make their data available through data clean rooms and audience modeling, which, according to Starling, ensures that personally identifiable information is not exposed and publisher data cannot be reverse-engineered to identify individuals. The company declined to name its third-party data modeling partners but plans to publicly share more specifics soon.
Until then, Bassik said, Starling is staying focused on making the most of show-level data to “support [media] optimization and drive better outcomes for publishers and advertisers.”
