Pharma brands spend billions of dollars every year on poorly targeted TV advertising.
A new startup called BranchLab, which was launched earlier this year by three ad tech vets, wants to change that.
BranchLab creates and sells TV audience segments for health care providers and pharma advertisers without using identifiable information from patients.
One-to-one targeting solutions designed for the web don’t work in a one-to-many medium like TV. But even if they did, new privacy laws – like Washington state’s recently enacted My Health, My Data Act – are restricting the use of sensitive information, including health data.
This is the dual challenge BranchLab aims to solve, said CRO and Co-Founder Michael Parkes.
Parkes was previously the president of VideoAmp – he spent seven years there before leaving in October – and his co-founders have a similar ad tech history. CEO Josh Walsh was a co-founder of AdTheorent (acquired by Cadent in April) and CTO Chris Cagle also has a strong ad tech background, including a five-year stint at AdTheorent.
Those pedigrees helped attract investors. On Thursday, the startup announced the completion of a seed funding round led by Newark Venture Partners with participation from numerous other strategic investors, including VC firm Aperiam.
The company declined to share the amount of funding it received but said it plans to use it to build its technology and its teams, including sales roles, marketing, client success and tech.
Health care has targeting problems
But BranchLab also caught the attention of investors because it seeks to capitalize on a market that’s “extremely, woefully underserved,” Eric Franchi, investor and partner at Aperiam, told AdExchanger.
You might be wondering how, exactly, TV advertising for health and pharma could possibly be considered underserved when viewers are inundated daily with poorly targeted ads rattling off side effects of drugs to treat ailments they probably don’t even have.
The problem is a lack of sufficient data for targeting.
Neither third-party cookies nor the alternative identifiers cropping up to replace them are all that useful within the TV ecosystem, Parkes said. TV advertising relies on age and gender demos and household-level targeting, which don’t match up well with identifiers that support one-to-one targeting for digital advertising, he added. The result is low match rates and limited scale.
This data dilemma is why health and pharma advertisers settle for match rates on TV that are so low they’d be considered “a total failure” by the standards of the digital ad ecosystem, Franchi said.
Still, Parkes said, health and pharma advertisers “want to be as targeted as they can” without infringing a consumer’s privacy rights or creeping them out, which calls for a targeting solution that doesn’t rely on identifying information, he said.
BranchLab avoids identifiers
BranchLab sources health claims data from insurance companies, much like other data providers that service health care and pharma marketers. It says its data graph covers roughly 300 million US residents and 2.1 million health care providers.
What’s different about BranchLab is how it uses that data, Walsh said.
First, it puts the data in a clean room environment so it doesn’t leak the identity of any patient. Then it uses machine learning technology to create audience models based on certain attributes, such as age, gender and social determinants of health, which include urban density and proximity to health care facilities (in this case, the first three numbers of a ZIP code).
Through direct integrations with TV publishers (which BranchLab declined to name), it scores a publisher’s audience segments or viewers based on how well they fit its modeled audiences, Walsh said.
For example, an audience segment can be modeled after patients with plaque psoriasis who haven’t been prescribed treatment yet. BranchLab then compares its model with a TV publisher’s viewing data to determine which households are most likely to fit the mold, and what content they’re most likely to be watching.
Agencies can then buy this inventory directly from TV publishers themselves or from most major DSPs, Walsh said. For programmatic buys, the startup is integrated with several SSPs, and primarily uses Magnite.
“The difference is that we’re not creating a list of IDs and sending them out to publishers,” he said, because attaching claims data to a persistent identifier is risky from a privacy standpoint. Instead, BranchLab uses probabilistic models and compares them to publisher data directly.
In this way, he said, health and pharma brands find more of their target audience on TV without creating data privacy-related problems.
The company also helps drug manufacturers target ads to health care providers and practitioners, which don’t have the same privacy restrictions as actual patients seeking medical help.
BranchLab updates its data from insurance companies weekly, so it can track metrics like “script lift,” which, in this case, is based on the number of new prescriptions written during a TV ad campaign’s flight. The weekly refresh allows buyers to optimize their campaigns regularly, Walsh said.
To be fair, the absence of deterministic identifiers doesn’t change the fact that any degree of targeted advertising based on sensitive information, like a medical diagnosis, can feel a little invasive.
But as long as BranchLab continues to avoid passing identifiers that could reveal a person’s identity, it’s technically kosher. Such is the nature of pharmaceutical advertising.
