# Who owns the data: governance across studios, platforms and distributors
At 9:14pm, someone presses play on a licensed movie through a smart TV's built-in app. That single click generates at least three data trails: the smart TV manufacturer logs the device ID and viewing duration, the streaming platform logs the account, pause points and browsing path that led to the click, and the studio that licensed the film gets... maybe nothing. Or maybe a monthly aggregate report. Or, in a well-negotiated deal, a data feed it can actually query. Which of these happens is decided not by technology but by a contract clause, usually the least glamorous part of the licensing negotiation.
This lesson traces that viewing event to show how governance charters and data-sharing agreements determine who owns what, and where these deals stall in practice.
The studio (e.g. Warner Bros. Discovery, Sony Pictures) wants viewership and audience data to inform future greenlighting, marketing spend and next-window pricing (SVOD, then AVOD, then licensing to other platforms). Data is a strategic asset, not a byproduct.
The streaming platform (e.g. Netflix, Max, Amazon Prime Video) holds the direct customer relationship. It controls account-level data: what else this viewer watches, churn risk, household composition signals. Platforms guard this fiercely because it is the basis of their recommendation engines and retention models.
The smart TV manufacturer (e.g. Samsung, LG, Roku, Vizio) sits underneath both. Its operating system sees every app launched, every remote input, sometimes even screen content via automatic content recognition (ACR), a technology that samples pixels or audio to identify what's on screen regardless of source. ACR data is valuable to advertisers and is a growing revenue line for TV makers themselves.
Three parties, three separate data sets, describing the same 9:14pm event. None of them automatically has the right to see the others' slice.
"Ownership" is the wrong mental model for most media data disputes. Raw viewership logs are typically owned by whoever collects them (the platform or the device maker), governed by their own privacy policy and terms of service.
What studios actually negotiate for is access rights, defined in the licensing agreement's data-sharing schedule. Typical tiers, from weakest to strongest:
Most licensing deals settle for aggregate reporting. Full APIAPIApplication Programming Interface: a standardised interface that lets applications communicate and exchange data without knowing each other's internal workings.Voir la définition complète → query access is rare, and when it exists it usually comes with restrictions on join-ability (the studio cannot combine the streamer's viewing data with its own customer databasecustomer databaseCustomer Relationship Management: software and strategy to manage and analyse customer interactions throughout their lifecycle.Voir la définition complète → to re-identify individuals).
Three regimes matter most in 2026:
GDPR (General Data Protection Regulation, EU, in force since 2018) requires a lawful basis for processing personal data and gives EU viewers rights to access, correct or delete their data. If viewing history counts as personal data (it usually does, since it's tied to an account), any transfer between studio, platform and device maker needs a documented legal basis, typically a data processing agreement specifying purpose and retention limits. The EU's European Data Protection Board publishes guidance on cross-border transfers relevant to multi-territory licensing.
CCPA/CPRA (California Consumer Privacy Act, amended by the California Privacy Rights Act) gives California residents rights to know what's collected and to opt out of "sale" or "sharing" of personal data for cross-context advertising. This directly hits ACR data monetization by smart TV makers: if Samsung shares ACR-derived viewing signals with ad networks, that can qualify as a "sale" requiring an opt-out mechanism.
VPPA (Video Privacy Protection Act, US federal law from 1988, originally passed after a newspaper published a Supreme Court nominee's video rental history) restricts disclosure of personally identifiable video viewing records. It predates streaming but courts have applied it to platforms and apps, making it a live litigation risk anytime viewing data crosses to a third party without consent.
None of these laws say "studios own viewership data." They constrain *how* any party can move personal data across the chain, which is exactly why the contract has to do the heavy lifting.
Inside large media companies, a data governance charter is the internal document that translates external law and external contracts into operational rules: who can access which data, under what approval, logged how.
A practical charter for this three-way deal would specify:
1. Data classification: is viewing-event data classified as personal, pseudonymized or aggregate, and does that classification change at each handoff?
2. Access control: which named roles (not individuals) at the studio can query the platform's shared dataset, and through what interface.
3. Purpose limitation: data shared for "content performance analysis" cannot silently be reused for "ad targeting" without a new agreement.
4. Retention and deletion: how long the studio can retain query results after the license term ends. This is where many deals stall: platforms often want data deleted at contract end, studios want it retained for historical trend analysis.
5. Audit rights: does the studio get to inspect the platform's data pipelinedata pipelineETL (Extract, Transform, Load) is a data integration process that pulls data from sources, reshapes it into a consistent format, and writes it into a target system.Voir la définition complète → for compliance, or only receive attestations?
Where deals commonly stall: the platform's unwillingness to expose raw device-level data (fear of competitive leakage to a studio that also licenses to rivals), and the studio's unwillingness to accept aggregate-only reporting for a multi-million dollar licensing renewal decision.
Vérification des acquis
1. In the single viewing event described, what ultimately determines which party receives which slice of data?
2. Why does the streaming platform typically resist sharing granular account-level viewing data with the studio that licensed the content?
3. A studio negotiating a new licensing deal wants to actually query viewership data rather than receive a static report. Which outcome reflects the strongest position on the spectrum described in the lesson?
4. Select ALL correct answers about why the same 9:14pm viewing event generates three separate, non-overlapping data trails.
Sélectionnez toutes les réponses correctes.
5. Select ALL correct answers about automatic content recognition (ACR) as described in the lesson.
Sélectionnez toutes les réponses correctes.
Non-technical managers involved in these deals should still be able to ask for and read the following:
A simple technical check a data team might run to flag joinability risk, using pseudo-code logic anyone can follow:
# Check whether a "pseudonymized" dataset is actually anonymous
overlap = shared_columns(dataset_A, dataset_B)
if len(overlap) >= 2 and includes_rare_values(overlap):
flag("Re-identification risk: high")
else:
flag("Re-identification risk: low")If a studio's customer databasecustomer databaseCustomer Relationship Management: software and strategy to manage and analyse customer interactions throughout their lifecycle.Voir la définition complète → and a platform's "anonymized" export share two or more overlapping fields (say, zip code and device model), re-identification risk rises sharply, this is the kind of check regulators expect companies to have run before calling data "anonymized."
How Streaming Services Track You