Streaming TV buyers want a clearer view of where their ads run and whom they reach. Starling is betting that show-level signals can help solve both problems. Its commercial test is whether that added detail can improve acquisition results enough to justify the spend.
Michael Bassik says the market has little need for another supply-side platform, which helps publishers sell advertising inventory. His case for launching Starling rests on turning viewing signals into audiences that advertisers can buy across platforms.
For growth teams, the opportunity deserves scrutiny rather than an automatic budget shift. Content detail could sharpen targeting, but buyers still need evidence that it improves funnel performance.
Show-Level Signals Need a Buying Purpose
Starling works directly with connected TV publishers, including AMC, TelevisaUnivision, Paramount and Telly, to obtain content signals. Those inputs cover show titles, genres, apps and the content around an ad placement.
The company also draws on audience and identity data from publishers and unnamed outside providers. It uses those inputs to build custom audience segments based on viewing habits.
Bassik describes show titles paired with audience identifiers as a particularly valuable combination for buyers. That pitch connects two distinct questions: what someone watches and whether that person fits an advertiser’s target audience.
Buyers should set the commercial objective before choosing a segment. A show’s subject may suggest audience relevance, but it does not establish purchase intent. For a demand generation team, the useful question is whether that context improves access to likely customers.
Private Deals Turn Viewing Context Into Inventory
Bassik offers a buying example: an advertiser could seek viewers of particular morning news programs. Starling could aggregate the relevant inventory and make it available to buy in private marketplace deals, known as PMPs.
He sees these deals as a way to build scale. Starling also uses audience modeling to extend a brand’s target group to viewers judged similar.
Those approaches deserve separate scrutiny because they answer different planning needs:
- Content-based reach: Ask which shows and publishers contribute inventory to the proposed deal.
- Modeled reach: Ask how added viewers resemble the original audience and how that resemblance is assessed.
- Combined delivery: Ask whether reporting can distinguish the initial target from its modeled expansion.
These questions help buyers weigh reach against relevance before committing spend. A broader segment may offer more delivery, while making the audience harder to evaluate.
Publisher Economics May Change the ROI Calculation
Bassik says deals supported by content data tend to attract stronger demand and higher ad prices. That gives publishers a commercial reason to make useful signals available.
Advertisers should assess the same shift from the cost side. If richer context commands a premium, the campaign needs enough added value to cover that expense. Better placement detail alone cannot prove a lower customer acquisition cost.
An evaluation plan should connect the media purchase to the advertiser’s actual funnel. For a lead generation campaign, that could mean comparing qualified leads rather than counting every response equally. For a sales campaign, buyers should judge the test against their chosen conversion and return measures.
These are proposed tests, not results Starling has demonstrated in the source account. Its targeting approach offers a hypothesis that buyers can examine with their own campaign data.
Privacy Limits What Publishers Can Expose
Access to viewing signals remains uneven because publishers differ in what they will share. Some treat those records as a competitive asset. Others are concerned about the Video Privacy Protection Act (VPPA).
Passed in 1988, the law restricts video providers from sharing an individual’s viewing details without consent. Bassik says some publishers interpret it as a reason to withhold content signals from programmatic transactions.
Starling says it uses outside modeling providers, clean rooms and audience modeling to handle publisher data. It claims these methods prevent exposure of personally identifiable information and stop users from being identified through reverse engineering.
The company has not named its modeling partners, although it plans to release further details. Buyers should treat the privacy safeguards as company claims pending more disclosure, rather than as an independently verified compliance finding.
Build the Test Around Acquisition Quality
Starling’s approach gives marketers a reason to investigate content-led CTV buying, but the evaluation should remain narrow. Start with an audience question that existing buying methods do not answer well.
Before allocating a test budget, buyers should establish:
- Which content signals are available and where coverage is incomplete.
- How modeled viewers enter the segment and appear in reporting.
- What privacy documentation supports the proposed use of viewing data.
- Which acquisition measure will determine whether the test earns further spend.
Content transparency can make a media plan easier to inspect. Its value to a growth team depends on what happens after exposure. Set the success threshold before launch, then expand only if the campaign meets the agreed acquisition target.
