VIZIO’s 2017 order required prominent notice and express consent where a TV setting had promised suggestions
On February 6, 2017, the FTC and New Jersey settled charges that VIZIO collected viewing data behind a setting described as program suggestions. The order required prominent disclosure and affirmative express consent, a standard later orders reused.
Looking Back revisits past weeks one at a time, up to the most recent, and asks what it meant for the banks, health systems and public agencies we serve. The series was written in 2026, with the benefit of hindsight, and each article says which week it looks back on.
Two-minute summaryThe dek and the first sentence of each section, quoted from the article
On February 6, 2017, the FTC and New Jersey settled charges that VIZIO collected viewing data behind a setting described as program suggestions. The order required prominent disclosure and affirmative express consent, a standard later orders reused.
- The week The release’s title said the charges concerned viewing histories collected on 11 million smart televisions without users’ consent. [1] The agencies alleged that the tracking was unfair and deceptive, in violation of the FTC Act and New Jersey consumer protection laws. [1] These were allegations in a complaint, resolved by a stipulated order.
- The label: a setting named for a benefit The release described the setting at the center of the case.
- The default: collection that ran without a decision For regulated organizations, the retrofit is the detail with the widest reach.
- The recipients: where the data went after collection The post also described the contracts.
- The remedy: what the order asked VIZIO to prove The release listed the order’s requirements.
- How the case reaches different organizations The VIZIO order bound one company.
- Objections, in order The objection is that VIZIO concerned entertainment, and that its lesson does not transfer.
- What digital and privacy leaders could check The four parts give a short list of checks.
- With hindsight In April 2021, the Supreme Court decided AMG Capital Management v. Federal Trade Commission.
The week
On February 6, 2017, the Federal Trade Commission and the New Jersey Attorney General announced a settlement with VIZIO, a maker of smart televisions. [1] The release described a payment of $2.2 million. [1] It said $1.5 million would go to the FTC and $1 million to the New Jersey Division of Consumer Affairs, with $300,000 of that amount suspended. [1]
The release’s title said the charges concerned viewing histories collected on 11 million smart televisions without users’ consent. [1] The agencies alleged that the tracking was unfair and deceptive, in violation of the FTC Act and New Jersey consumer protection laws. [1] These were allegations in a complaint, resolved by a stipulated order. The release did not describe a finding by a court.
According to the release, the complaint said that starting in February 2014, VIZIO made smart TVs that captured second-by-second information about the video they displayed. [1] The release listed the sources of that video: consumer cable, broadband, set-top box, DVD, over-the-air broadcasts and streaming devices. [1] The collection reached what was on the screen, whatever its source.
The release also said VIZIO appended specific demographic information to the viewing data, such as sex, age, income and household value. [1] It said VIZIO sold the data to third parties, who used it for purposes that included targeting advertising to consumers across devices. [1]
The consent question turned on a setting. The release said VIZIO touted a feature called “Smart Interactivity” that “enables program offers and suggestions.” [1] It said the company failed to inform consumers that the settings also enabled the collection of their viewing data. [1]
The FTC published a business guidance post the same day, titled “What Vizio was doing behind the TV screen.” [2] The post described the allegations in more detail. It said VIZIO had also retrofitted older models by installing its tracking software remotely. [2] It said the televisions collected a selection of pixels on the screen on a second-by-second basis and matched them to a database of television, movie and commercial content. [2]
The post described the scale the complaint alleged: as many as 100 billion data points each day from millions of televisions. [2] It said VIZIO provided consumers’ IP addresses to data aggregators, which could match those addresses to an individual or household. [2] It said VIZIO’s contracts prohibited re-identification of consumers and households by name, but allowed other details, including sex, age, income, marital status, household size, education and home ownership. [2]
The post also said the feature did not provide the “program offers and suggestions” it promised. [2] The setting’s description named a benefit. In the agencies’ account, the benefit was absent and the collection was the main effect.
The release listed what the stipulated order required. [1] VIZIO had to prominently disclose and obtain affirmative express consent for its data collection and sharing practices. It was prohibited from misrepresenting the privacy, security or confidentiality of consumer information it collected. It had to delete data collected before March 1, 2016. It had to implement a comprehensive data privacy program and conduct biennial assessments of it. [1]
The Commission vote to approve the complaint and the stipulated order was 3-0. [1] Acting Chairman Maureen K. Ohlhausen issued a concurring statement. [1] The documents were filed in the U.S. District Court for the District of New Jersey. [1]
The business post drew lessons for companies. It advised explaining data collection practices up front, in language people could understand. [2] It advised getting consumers’ opt-in consent before collecting and sharing highly specific information about their entertainment preferences. [2] It said companies could hardly claim to offer a choice if the tools needed to exercise it were hard to find. [2] It also said established consumer protection principles apply to new technology, and it referred to the FTC’s recent workshop on smart TVs. [2]
The legal frame behind the unfairness charge is in Section 5 of the FTC Act. Section 5(n) says the Commission may not declare a practice unfair unless it causes or is likely to cause substantial injury to consumers. [3] The injury must not be reasonably avoidable by consumers themselves, and it must not be outweighed by countervailing benefits to consumers or to competition. [3]
Section 5 also sets the edge of the FTC’s reach. Section 5(a)(2) excludes banks, savings and loan institutions described in the Act, and federal credit unions described in the Act, among others. [3] That exclusion matters for the sectors we write about. A bank was not in the FTC’s Section 5 jurisdiction in the way VIZIO was, though the vendors that build its devices, apps and websites may be.
The VIZIO case is easy to file as a television story. We read it as a case about where consent lives in a product. The agencies did not allege that VIZIO hid its data practices in an unreadable policy. They alleged that a setting described one thing and did another, that the collection ran without a fresh decision by the viewer, and that the data left for buyers the viewer never chose. [1] [2]
Our argument is that the order located consent in the user’s path, not in the paperwork. It required a prominent disclosure and affirmative express consent for collection and sharing. [1] That standard reached well past televisions. Banks, health systems and public agencies run apps, websites, kiosks and connected devices that collect behavior data, and they buy those products from companies the FTC can reach. [3]
We take the case in four parts. Each is one place where the consent failed, as the agencies described it:
- The label: what the setting was called and what its description promised.
- The default: what happened when the viewer did nothing, including on televisions already in homes.
- The recipients: who received the data and what they could do with it.
- The remedy: what the order required, and what that implies about a consent record that would hold up.
The case was settled, not tried. The facts we describe are the agencies’ allegations, as the FTC summarized them, and the terms are those of a stipulated order. [1] [2] We draw lessons from the allegations because they describe design choices any organization could make. We do not treat them as findings about VIZIO.
The first part concerns words on a screen. The second concerns behavior without words. The third concerns the data’s path after collection. The fourth concerns the evidence that the first three were fixed. Together they cover the life of a consent decision, from the prompt to the proof.
The label: a setting named for a benefit
The release described the setting at the center of the case. VIZIO touted “Smart Interactivity” as a feature that “enables program offers and suggestions.” [1] It did not tell consumers that the setting also enabled the collection of their viewing data. [1]
That allegation is narrower than a missing disclosure. A label existed. A description existed. The complaint’s point, as the FTC described it, was that both pointed at a benefit and away from the collection. [1] [2] The FTC’s post added that the feature did not provide the promised program offers and suggestions at all. [2]
Why the name carried legal weight
A deception theory asks what a reasonable consumer would take from what the company said. An unfairness theory asks about injury the consumer could not reasonably avoid. [3] A name that describes a benefit makes the collection harder to avoid, because the person deciding never learns what the decision covers.
We read the case as treating the label as part of the disclosure, not as marketing around it. The words a user sees at the moment of choice did the legal work. The privacy policy, wherever it sat, could not cure a setting whose own description pointed elsewhere.
Where labels live in regulated organizations
Banks, health systems and agencies write labels too. A mobile banking app may offer a setting to “personalize offers.” A patient portal may offer to “improve your experience.” An agency website may ask to “remember your preferences.” Each label names a benefit.
The question VIZIO raised is whether the label also names the collection and the sharing it switches on. If a toggle enables analytics that leave the organization, the toggle’s words are part of the consent. A legal review that reads only the policy misses the place where users decide.
That review is a product task as much as a legal one. Someone has to inventory the switches, read their labels as a user would, and compare each label with what the switch actually does in code. That comparison needs a named owner, because it falls between product, engineering and legal teams.
What a label review would record
A label review can be specific. For each setting, it would record the words shown, the screens where they appear, and the data flows the setting controls. It would record the destinations of each flow, inside or outside the organization.
It would also record the date of each version of the wording. VIZIO’s case turned on what consumers saw at particular times. An organization that cannot show what its users saw on a given date cannot show what they agreed to.
This is a records problem as well as a design problem. Release notes, screenshots and copy decks are the evidence. If they are not kept, the organization’s account of its own consent rests on memory.
One setting, several surfaces
A bank, health system or agency can have more than one surface. Words differ between a website, a mobile app, a call center script and a device in a lobby. A setting can carry a careful label on one surface and a vague one on another.
The VIZIO allegations concerned the words consumers saw on the product itself. [1] The parallel risk for a multi-channel organization is inconsistency. If the app says one thing and the website says another, the organization’s own records show two accounts of the same choice.
A label review therefore needs to run across surfaces, and within each one. The output is a single list of choices, each with every wording in use and the flows it controls on each surface.
Who approves the words
Copy changes move fast in product teams. A designer can rename a setting in an afternoon. If the rename does not pass through privacy review, the label and the flow can drift apart without anyone deciding that they should.
One control is simple: treat the wording of any data setting as controlled text. Changes to it go through the same review as changes to a privacy notice. The review checks the new words against the flows the setting governs, and the approval is recorded with a date.
The limits of the label point
The label is one place consent can fail. It is not the only one. A perfect label does not help if the collection starts before the user sees it. The second part turns to that problem.
The default: collection that ran without a decision
The release said VIZIO made televisions that captured viewing data starting in February 2014. [1] The FTC’s post said VIZIO also retrofitted older models by installing its tracking software remotely. [2] The post said the company did this without clearly telling consumers or getting their consent, as the agencies alleged. [2]
For regulated organizations, the retrofit is the detail with the widest reach. A person who bought a television before the software existed made no choice about it. An update arrived, and with it a new data flow.
Why the remote update changes the consent question
Consent attaches to a moment. When a product’s behavior changes after purchase, the earlier moment no longer covers it. The order’s requirement of affirmative express consent for collection and sharing, in the release’s description, does not depend on when the device was sold. [1]
We read the remedy as tying consent to the data practice, not to the sale. If a practice is new, the decision has to be new. A software update that adds collection is, on this reading, a new practice that needs its own decision.
Defaults in banking, health and public services
Regulated organizations ship software updates as well. A bank adds a fraud-analytics library to its app. A health system adds a session-replay tool to its portal to diagnose errors. An agency adds a new analytics tag to a benefits application.
Each change can add a data flow that users never chose. The users who installed the app last year agreed to what the app did then. Whether they agreed to the new flow depends on what they were told when it arrived.
The VIZIO allegations suggest a test for each release. Does it add collection, or send existing data somewhere new? If it does, what will the user see before the flow starts? A release process that asks those questions puts the consent decision into the engineering workflow.
The off switch is part of the default
The FTC’s post said companies can hardly claim to offer a choice if the tools needed to exercise it are hard to find. [2] That point extends the default question. A setting that is on by default and buried several menus deep is, in practice, a default with no exit.
For a bank or a health system, the equivalent is a preference that exists but cannot be found from the screen where the collection happens. An organization can test this directly. It can count the steps it takes a user to turn off a given flow, starting from the screen where the flow begins.
What the default costs to change
Changing a default has costs. An opt-in flow can collect less data than an opt-out flow. Analytics teams can lose coverage. Product teams can lose measurements they used to tune features.
Those costs are real. They are also costs the VIZIO order imposed directly, by requiring affirmative express consent before collection and sharing. [1] An organization that waits for an order pays the same cost later, plus the cost of deleting what it collected before.
Updates the organization did not write
Not every new data flow comes from the organization’s own code. Software kits from analytics and advertising vendors update on their own schedules. A new version of a kit can add collection that the organization never reviewed.
VIZIO’s retrofit was the company’s own decision, as the FTC described it. [2] For a regulated organization using a third-party kit, the equivalent change can arrive through a routine dependency update. The control is to pin versions, read release notes for data changes, and test traffic after each update.
Devices in shared spaces
Some devices sit in spaces that no single person controls. A hospital may provide entertainment screens in patient rooms. A bank may run kiosks in branches. An agency may run self-service terminals in an office.
Consent is harder in those settings, because the person using the device did not set it up. The VIZIO order’s requirement of a prominent disclosure and an affirmative choice assumes someone who can see the prompt and decide. [1] For shared devices, the safer design may be to switch collection off unless the organization itself has a clear basis for it.
Testing what a device actually sends
The FTC’s post described collection of on-screen pixels matched against a content database. [2] Nothing about that process would be visible to a viewer. It was visible only in the device’s network traffic and in the company’s systems.
An organization can test its own products the same way. A tester installs the app or device, uses it, and records every outbound connection. The record is compared with the consent the user gave. A flow with no matching consent is a finding, whatever the policy says.
The test is repeatable, and it should run after each release. It is also the only test that catches flows added by components no one on the team wrote.
The recipients: where the data went after collection
The release said VIZIO sold the viewing data to third parties, who used it for purposes including targeting advertising to consumers across devices. [1] The FTC’s post said VIZIO provided consumers’ IP addresses to data aggregators, which could match the addresses to an individual or household. [2]
The post also described the contracts. VIZIO’s contracts prohibited re-identification of consumers and households by name, the post said, but permitted other details such as sex, age, income, marital status, household size, education and home ownership. [2]
Why a no-names clause was not enough
A contract that bars re-identification by name sounds protective. The FTC’s account shows why it did not settle the question. The recipients could still link viewing data to a household through an IP address and a set of demographic attributes. [2]
We read this as the case’s lesson about de-identification by contract. A clause that removes names but leaves identifiers and attributes does not make the data anonymous. It only changes which field carries the identity.
Recipients in regulated organizations
Banks, health systems and agencies send data to third parties for ordinary reasons. Analytics vendors measure site traffic. Advertising platforms measure campaign results. Customer communication tools send messages and track opens.
Each of those relationships has a contract. Each contract has some clause about use of the data. VIZIO’s case asks whether the clause addresses the identifiers that travel with the data, such as IP addresses, device identifiers and hashed emails.
It also asks whether the organization knows which recipients exist. The FTC’s post described data aggregators and other companies on the receiving end. [2] A recipient list built from contracts alone can miss flows created by tags and software kits that no one negotiated.
Building the recipient inventory
A useful inventory starts from the traffic, not the contracts. It lists every external destination that a website, app or device sends data to, observed in testing. It then matches each destination to a contract and a purpose.
Destinations with no contract are the first finding. Destinations with a contract but no limits on the identifiers they receive are the second. Destinations used for purposes the user was never told about are the third.
The inventory is not a one-time exercise. Each release can add a destination, as the second part showed. The inventory has to be refreshed when the product changes, or it describes a product that no longer exists.
Sensitivity of the data
VIZIO’s data was television viewing. The agencies treated it as warranting affirmative express consent before collection and sharing. [1] The data that banks, health systems and agencies hold is at least as revealing. Account activity, appointment types and benefit applications each say something specific about a person.
If viewing history needed a prominent disclosure and an affirmative choice, in the order’s terms, the case for the same treatment of financial and health behavior is stronger. That is our reading, not a holding. The VIZIO order bound VIZIO.
Matching across devices
The release said recipients used VIZIO’s data for purposes that included targeting advertising to consumers across devices. [1] That phrase describes a capability, not a single flow. Data collected on one device can be joined to a person’s activity on others.
For a regulated organization, cross-device matching widens the exposure. A visit to a health system’s website about a condition, joined to an advertising profile, can follow the person to other screens. The organization may never see the join. It only supplied one of its inputs.
Measurement and targeting are different purposes
Organizations send data out for measurement and for targeting. Measurement asks whether a campaign worked. Targeting decides who sees the next message. The two purposes carry different risks, and a consent prompt that covers one does not cover the other.
VIZIO’s data, as the release described it, went to third parties who used it for purposes including targeting. [1] A consent record that says only “analytics” does not describe that use. Listing purposes separately makes the user’s choice meaningful and the organization’s record accurate.
The second hop
The FTC’s post described data aggregators among the recipients. [2] An aggregator’s business is to combine data from several sources and pass the result on. Data that reaches one recipient can therefore reach others the original collector never contracted with.
A sender’s contract binds the first recipient. It reaches further only if it says so and the first recipient passes the terms on. A regulated organization can ask each recipient for its own list of onward recipients and the terms that bind them. If the answer is unclear, the safer assumption is that the data travels.
This matters for deletion as well. A deletion request that stops at the first hop leaves copies further down the chain. The VIZIO deletion term, as described, applied to data VIZIO had collected; it did not describe how copies already sold would be recovered. [1]
The remedy: what the order asked VIZIO to prove
The release listed the order’s requirements. They were prominent disclosure and affirmative express consent, and no misrepresentations about privacy, security or confidentiality. [1] They also included deletion of data collected before March 1, 2016, a comprehensive data privacy program, and biennial assessments of the program. [1]
Each requirement implies a record. Prominent disclosure implies a record of what was shown. Affirmative express consent implies a record of who agreed and when. Deletion implies a record of what was deleted. A privacy program and its assessments imply records that someone outside the company could review.
Consent as a record, not a moment
A consent decision happens once on a screen. Proving it later requires a stored record. The record has to connect a person, or a device, to the words shown and the choice made, with a time.
Regulated organizations may already keep similar records for other purposes. A bank may keep records of disclosures it delivered. A health system may keep signed forms. An agency may keep application histories. The VIZIO remedy suggests extending that habit to data collection choices made inside software.
Deletion as a boundary
The deletion requirement drew a line at March 1, 2016. [1] Data collected before that date had to go. The requirement assumes the company could tell which data was collected when, and where it had been copied.
That assumption is the hard part for any organization. Data sent to a third party is outside the sender’s systems. Deleting it requires the recipient’s cooperation, and a way to confirm it happened. A data map that records recipients and dates makes a deletion order possible to carry out.
Assessments and the program
The biennial assessments in the order put an outside review on a schedule. [1] That review would ask whether the program worked, as well as whether it existed. A program that existed on paper and did not catch a mislabeled setting would not satisfy that review, on our reading.
For a regulated organization, the parallel is the internal audit plan. Data collection practices in customer-facing software can be an audit area like any other. The audit would test labels, defaults and recipients against the records the organization claims to keep.
What the remedy did not include
The release described payments to the FTC and to the New Jersey Division of Consumer Affairs: $1.5 million to the FTC and $1 million to New Jersey, with $300,000 of the New Jersey amount suspended. [1] It did not describe refunds to consumers. On our reading, the remedy’s weight was in the conduct requirements, not in compensation to viewers.
That shape matters for how organizations weigh the risk. The direct cost of a settlement like this one may look small next to the cost of changing products and deleting data. The conduct requirements are where the expense lands.
Mapping each order term to a control
The order’s terms translate into controls an organization can run without any order. Each one has an owner and a piece of evidence:
- Prominent disclosure: product and privacy teams own the wording; the evidence is the dated copy and screenshots.
- Affirmative express consent: engineering owns the choice and its storage; the evidence is the consent log.
- No misrepresentation: legal and marketing own the claims; the evidence is the review record for each claim.
- Deletion: data owners and vendor managers own it; the evidence is deletion confirmations, including from recipients.
- Program and assessments: the privacy office owns the program; the evidence is the assessment report and its follow-up.
The list follows the release’s description of the order. [1] The owners and evidence are our suggestion. They are one way to make each term testable before a regulator asks.
How long consent records need to last
A consent record is useful only while someone may ask about the choice. That period can be long. Data collected under a consent may be held, shared and used for years, and questions can arrive at any point.
A practical rule is to keep the consent record at least as long as the data it covers, and as long as any copy held by a recipient. The VIZIO order’s deletion term shows why: deleting by date requires knowing what was collected under which terms. [1]
How the case reaches different organizations
The VIZIO order bound one company. Its reasoning reaches other organizations unevenly, depending on who regulates them and what they buy. We take the main types in turn.
Banks and savings associations
Section 5(a)(2) of the FTC Act excludes banks and certain savings and loan institutions from the FTC’s Section 5 authority. [3] On our reading, a bank’s own consent practices fall to other law and other regulators, not to the FTC’s Section 5 power.
The exclusion does not cover the bank’s vendors. A company that builds a bank’s app, runs its analytics or supplies its branch devices is a separate company. Our reading is that VIZIO-style allegations could reach those vendors directly, whatever the bank’s own status.
Credit unions
Section 5(a)(2) excludes federal credit unions described in the Act. [3] The same vendor point applies. A credit union’s digital banking provider, its marketing platform and its card program partners may each be within the FTC’s reach.
Broker-dealers, wealth managers and fintechs
Broker-dealers and investment advisers are not on the Section 5(a)(2) exclusion list we cite. [3] Fintech companies that are not banks are not on it either. On our reading, the VIZIO theory is a direct risk for these firms, alongside their own securities and financial privacy rules. Counsel should check which regimes apply to a given firm.
Health systems and ambulatory networks
Health systems that are HIPAA covered entities answer to HIPAA for protected health information. The FTC’s Health Breach Notification Rule, amended in 2024 and discussed in the hindsight section, applies to entities not covered by HIPAA. [6] The consent lesson still applies to a health system’s own apps and websites. The data they collect about behavior can reveal health conditions, and the vendors that collect it are separate companies.
Life sciences and digital health companies
Consumer health apps and device makers outside HIPAA are the closest analogs to VIZIO. They sell connected products, push software updates and work with analytics and advertising partners. The FTC’s later actions, discussed below, show where this line of cases went for them. [5] [6]
Federal, state and local agencies
Public agencies are not commercial sellers of televisions. They do buy kiosks, apps, websites and connected equipment from commercial vendors. An agency’s procurement can require vendors to disclose every data flow and to obtain consent where a flow leaves the agency’s control.
Agencies also face a trust question that commercial firms do not. A resident applying for benefits may have no alternative provider. A setting that collects data without a clear choice is harder to justify when the person cannot walk away.
Technology vendors serving regulated clients
The companies that build software for banks, health systems and agencies sit closest to the VIZIO facts. They ship devices and apps, push updates and work with partners. Their clients carry the trust relationship, but the vendor controls the code.
Contracts can align the two. A client can require notice before any update adds collection or a new recipient. It can require the vendor to keep the consent records the order described, and to support deletion across its own sub-processors.
Health plans and insurers
Health plans are HIPAA covered entities for the protected health information they hold. Their member portals and apps also collect behavior data, and their marketing teams use advertising platforms. The label, default and recipient questions apply to those products as they do to a health system’s.
Insurers outside health coverage answer to their own regulators. For all of them, the vendor point holds. The companies that build their apps and run their analytics are separate companies with their own exposure.
Objections, in order
1. Television viewing is unlike financial or health data
The objection is that VIZIO concerned entertainment, and that its lesson does not transfer. We think it transfers more strongly, not less. The order required affirmative express consent for viewing data. [1] Financial and health behavior is at least as specific about a person, so the consent design questions are at least as sharp.
2. A privacy policy disclosed the practice somewhere
The release and the post do not say whether a policy mentioned the collection. They do say the setting’s own description named a different benefit. [1] [2] Our reading is that a policy cannot cure a setting whose own words point elsewhere. The order’s requirement of a prominent disclosure supports that reading. [1]
3. Banks are outside the FTC’s Section 5 power
The point holds for banks, savings and loan institutions and federal credit unions within the Act’s exclusions. [3] It is not correct for their vendors. It also does not answer the customer’s question, which is whether the bank asked before the data left.
4. Opt-in consent destroys analytics
Opt-in consent can reduce coverage, and the cost is real. The order imposed it on VIZIO anyway, and paired it with deletion of earlier data. [1] An organization that designs for consent from the start avoids the deletion cost and keeps the data it does collect usable.
5. Contracts barring re-identification protect the data
The FTC’s post described contracts that barred re-identification by name while allowing IP addresses and demographic details to travel. [2] A clause that removes names but leaves linkable identifiers does not make the data anonymous. Contracts help only when they address the identifiers that actually travel.
6. A single settlement sets no rule
That is true in law. A stipulated order binds its parties. What a settlement does is show how the agency reads its statute. The later actions discussed below show the same consent standard appearing in other orders. [5] [7]
7. Users click through prompts anyway
A user may accept any prompt without reading it. That does not make the prompt pointless. The order required a prominent disclosure and an affirmative choice, not a guarantee that every user reads it. [1] The record of a clear prompt is what lets an organization show the choice was real.
8. Prompts damage the customer experience
A prompt adds a step. Badly designed prompts frustrate people. The alternative the VIZIO allegations describe is worse for the relationship: data flowing under a label that named something else. [1] A short, specific prompt at the moment a feature starts is part of the experience, not an interruption of it.
What digital and privacy leaders could check
The four parts give a short list of checks. None requires a new system. Each requires someone to look at the product as a user would and compare it with what the code does.
- List every setting in customer-facing software that enables collection or sharing. Record its label, its description and the flows it controls.
- Compare each label with the flows behind it. Flag any label that names a benefit and omits the collection.
- For each software release, ask whether it adds a data flow or a new destination. Decide what users see before the flow starts.
- Build the recipient inventory from observed traffic. Match each destination to a contract and a stated purpose.
- Check whether contracts address identifiers such as IP addresses and device IDs, as well as names.
- Keep dated records of consent wording, consent choices and deletions, so the organization can show what users saw and chose.
- Test each release by recording what the app or device sends, and compare the traffic with the consent given.
- Require vendors to give notice before an update adds collection or a new recipient.
The checks also give leaders a way to measure progress. The share of settings with reviewed labels, the share of releases with traffic tests, and the share of recipients matched to contracts are each countable. Those counts belong in the privacy program’s regular reporting, beside incidents and requests.
These checks fit an existing privacy program. They move part of the work from policy review to product review, which is where the VIZIO allegations located the failure. [1] [2]
With hindsight
In April 2021, the Supreme Court decided AMG Capital Management v. Federal Trade Commission. It held that Section 13(b) of the FTC Act does not authorize the Commission to seek, or a court to award, equitable monetary relief such as restitution or disgorgement. [4] The decision is dated April 22, 2021. [4]
The 2017 VIZIO release described a payment to the FTC but did not state the statutory basis for it. [1] We do not say the VIZIO payment rested on Section 13(b). We do read AMG as narrowing one route the FTC had used to obtain money in first-time Section 5 cases. [4] Civil penalties remained available for knowing violations of FTC rules and for violations of final orders. [3]
On our reading, that shift made rules matter more. On February 1, 2023, the FTC announced an action against GoodRx, a telehealth and prescription discount provider. The FTC said it was the first enforcement action under its Health Breach Notification Rule. [5]
The FTC alleged that GoodRx shared users’ prescription medications and health conditions with Facebook, Google, Criteo and others, despite promises not to share health information with advertisers. [5] GoodRx agreed to pay a $1.5 million civil penalty for violating the rule. [5] The proposed order barred disclosure of health information for advertising. It required affirmative express consent before sharing health information for other purposes, and it barred dark patterns to obtain that consent. [5]
The order also required GoodRx to direct third parties to delete shared data, and to post a data retention schedule. [5] The Department of Justice filed the complaint and order on the FTC’s behalf in the Northern District of California. [5] The consent standard from VIZIO, affirmative express consent, reappeared in a health case, now attached to a rule that carried penalties.
On May 30, 2024, the FTC published amendments to the Health Breach Notification Rule, effective July 29, 2024. [6] The rule applies to vendors of personal health records and related entities that are not covered by HIPAA. [6] The amendments clarified the rule’s coverage of developers of health apps. They also revised the definition of breach of security to clarify that it includes unauthorized disclosures as well as data security breaches. [6]
On our reading, that definition change connects directly to VIZIO. The 2017 case, as the agencies alleged it, concerned data that left with the company’s intent but without the user’s consent. Under the amended rule, for the entities it covers, a disclosure of that kind can be a breach that triggers notice. [6] As checked on November 15, 2026, the eCFR showed the rule as amended in 2024, with no later amendment. [8]
On June 27, 2024, the FTC announced a final order with Avast, a UK-based software company. [7] The FTC alleged that Avast collected browsing information through its browser extensions and antivirus software and sold it through a subsidiary, Jumpshot, to more than 100 third parties. [7] The order banned Avast from selling or licensing web browsing data for advertising. [7]
The Avast order required affirmative express consent before Avast sold or licensed browsing data from non-Avast products for advertising. [7] It required deletion of the browsing information transferred to Jumpshot and of products or algorithms derived from it. It required notice to affected consumers and a $16.5 million payment, which the FTC said it expected to use for redress. [7]
The deletion terms show how far the remedy moved. VIZIO had to delete data collected before a date. [1] Avast had to delete data and the products and algorithms built from it. [7] For organizations that train models or build segments from collected data, that change matters. A consent failure can now reach what was built with the data.
What held up from 2017 was the consent standard. Affirmative express consent appears in the VIZIO, GoodRx and Avast orders as the FTC described them. [1] [5] [7] What changed was the force behind it: rules with penalties, a broader definition of breach for health data outside HIPAA, and deletion that reaches derived products. [6] [7]
For banks, the Section 5 exclusion in the statute stayed as it was in the text we cite. [3] For health systems, the HIPAA boundary of the health breach rule stayed in place. [6] For all of them, the companies that build and run their software stayed within the FTC’s reach, and the consent questions VIZIO raised still apply to those vendors’ products.
Spotted an error? Tell us and we will check it and correct the article where needed.