← all distillations

distillation · the evidence ledger

The FTC control-integrity record: an American standard, never pointed at a feed

Federal Trade Commission enforcement on whether an offered control actually works, covers its related surfaces, and persists; ROSCA and FTC Act section 5; the vacated Negative Option Rule, 16 C.F.R. part 425

This is the raw, evidence-tiered claim ledger behind the lawsuits hub. It is published for transparency: the source of truth, with its tiers intact, current as of 2026-08-14.

how to read the tiers

Each claim is tagged by the strength of the evidence behind it, not by how certain it sounds.

  • [ESTABLISHED]: stated in a primary source (a court order or filing, the docket, an official release) or, for a hearing, said on the official record. A fact about what was filed or said, not necessarily adjudicated true.
  • [OBSERVED]: carried by secondary coverage, or by a document read into the record.
  • [ASSUMED]: the distiller's inference, flagged as such.
  • Provenance flags ([interested-party], [advocacy-witness], [single-witness], [lawmaker characterization]) mark a claim that rests on one interested voice: reported, not adjudicated.

Tier key (this record's calibration, per the parent plan): ESTABLISHED means stated in a primary source (here, the text of a federal appellate opinion, an FTC complaint, or an entered order) or in the project's verified baseline with independent corroboration; OBSERVED means carried by secondary coverage or inferred from it; ASSUMED means the drafter's inference, flagged as such. Two calibrations are specific to this record and both cut against overreading it. A COMPLAINT IS AN ALLEGATION, so "the FTC alleged" is ESTABLISHED where "the company did" is not. A STIPULATED ORDER IS AN AGREEMENT, so it resolves a case without any court adjudicating the allegations behind it.

Reader summary

The most common objection to this project's proposal is that "controls that work" is too vague to be a legal obligation. That objection is weaker than it sounds, because American law has been enforcing a version of it for years, just never against a social feed.

The Federal Trade Commission's theory in a run of cases is not that companies lied about their controls. It is that the control itself failed, and that the failure was the violation. In its 2023 complaint against Amazon over Prime, the FTC alleged in one flat sentence that the button labelled "End Membership" did not end membership. It alleged separately that people who signed up through a Fire TV or FireStick could not cancel through those same devices. In its complaint against Epic Games, it alleged that Epic had a working undo button, discovered people were using it, and then renamed it, shrank it, moved it to the bottom of the screen and made it require a press-and-hold, after which an internal email recorded a roughly 35% decline in undo rates.

Those three allegations are, in order, the three things this site asks of a brake: that it work, that it exist on the surfaces where the choice was made, and that it not be quietly degraded. The vocabulary is different. The idea is the same one.

Then the general rule died on a technicality. In 2024 the FTC amended its Negative Option Rule to require, among other things, a cancellation mechanism as simple as enrolment. Industry challenged it, and on July 8, 2025, six days before the rule's compliance date, the Eighth Circuit vacated it in full because the Commission had skipped a required preliminary regulatory analysis. The court did not decide whether the rule was substantively lawful; it said in terms that it did not need to. It also said it did not endorse the practices the rule targeted.

So the American position today is that the principle is enforceable one company at a time, through litigation and negotiated orders, and is not written down as a general rule. And none of it has ever been aimed at a recommendation feed.

The single most useful document in this record is fifteen years old. In 2011 the FTC settled with an advertising company called Chitika, which offered consumers an opt-out from tracking. The complaint alleges the opt-out worked, told the user "You are currently opted out", and then quietly expired: the cookies were "set to expire after 10 days", after which they "automatically expired and disappeared from consumers' browsers". That is a brake that holds for a week and a half and then lets go without telling anyone.

What makes it valuable is not the failure but the fix. The consent order that resolved it specifies, as enforceable terms, that the choice must "remain in effect for a minimum time period of five (5) years", must "require no more than one additional click", must show the consumer "the current status of their choice", must disclose "that their choice is specific to the browser they are using", and must be reachable by "a hyperlink that directly takes consumers to the mechanism" from inside every targeted ad. Duration, friction, status, scope, and surface coverage, written down and agreed to in 2011. That is most of what this site's scorecard asks, in the vocabulary of an American consent order, and it is the strongest available answer to the claim that nobody knows how to draft this.

Read two limits with all of this. Complaints are allegations, not findings, and the Amazon case ended in a stipulated order, which is an agreement rather than an adjudication. And every matter here is about subscriptions, purchases and cancellation flows; extending the reasoning to a feed control is an argument this site is making, not something the FTC has done.

Brake-integrity relevance

The paper's section 5 argues that a design standard is not a standard until it says what must happen, how fast, across which surfaces, and for how long. The most common reply is that no regulator has ever written such a thing for an interface control.

This record is the counter-example, with a caveat that matters. The FTC has repeatedly treated all three properties as independently actionable:

  • It must work. The Amazon complaint's allegation about the End Membership button is a single sentence with no deception theory attached to it. The button's failure is the wrong.
  • It must cover its surfaces. The same complaint alleges that a cancellation path absent from the devices people enrolled on is a defect in the control, which is precisely the "scoped" criterion on the scorecard.
  • It must not be degraded. The Epic allegation is the sharpest, because it describes a control that worked being made worse on purpose, with the effect measured internally.

What this does not give the project is an adjudicated holding. These are allegations that ended in agreed orders. The strongest genuinely adjudicated point in this record runs the other way: when the FTC tried to convert the principle into a general rule, a federal appellate court vacated the whole thing without reaching its merits, and the vacatur stands.

The honest use, then, is narrow and still useful. It is not "a court has held that controls must work." It is: the United States already has an enforcement practice built on exactly this idea, so a standard that names it is not novel, and the objection that it is unadministrable is answered by the agency's own docket. What is missing is not the concept. It is anyone applying it to a feed.

Claims (the ledger)

(a) The general rule, and its vacatur

  1. [ESTABLISHED] In October 2024 the FTC amended its 1973 Negative Option Rule by a 3-2 vote, adding provisions that "bar sellers from misrepresenting material facts and require disclosure of material terms, express consumer consent, and a simple cancellation mechanism." (Eighth Circuit opinion, background section.)
  2. [ESTABLISHED] The Eighth Circuit vacated the Rule in full on July 8, 2025, per curiam (Loken, Erickson and Kobes, Circuit Judges), on the ground that the Commission "failed to follow procedural requirements under" section 22 of the FTC Act, specifically by not preparing a preliminary regulatory analysis. (The opinion writes the section as a symbol; it is spelled out here outside the quotation marks so the quoted span stays verbatim.) (Opinion, introduction and part II.)
  3. [ESTABLISHED] The court expressly declined to reach the merits: "because we hold the Commission's rulemaking process was procedurally insufficient and Petitioners demonstrated prejudicial error, we need not address Petitioners' other substantive challenges to the Rule." The unreached challenges were that the FTC exceeded its statutory authority and that the Rule was arbitrary and capricious. (Opinion, end of part II.)
  4. [ESTABLISHED] The court volunteered that the outcome was not an endorsement of the conduct: "While we certainly do not endorse the use of unfair and deceptive practices in negative option marketing, the procedural deficiencies of the Commission's rulemaking process are fatal here." (Opinion, part II.C.)
  5. [ESTABLISHED] The Rule contained a severability provision (16 C.F.R. section 425.9) and the court vacated everything anyway, "because of the prejudice suffered by Petitioners as a result of the Commission's procedural error", adding that "the party-specific vacatur requested by the Commission is not feasible" given the Rule's breadth. (Opinion, part II.C.)
  6. [ESTABLISHED] The timing is exact and worth stating precisely. The Commission had originally set a compliance date of May 14, 2025, then on May 9 deferred it to July 14, 2025 as an exercise of enforcement discretion. The opinion issued July 8, 2025. That is six days before the deferred compliance date. (Opinion, background; the "six days" is arithmetic from two dates in the opinion, not a phrase the court used.)

(b) The control must actually work

  1. [ESTABLISHED as an allegation] In FTC v. Amazon.com, the Commission alleged that the button labelled End Membership "did not end membership. Rather, it took the consumer to the Iliad Flow." (Original complaint, paragraph 119. In the amended complaint of 2023-09-20 the same text is paragraph 133.)
  2. [ASSUMED] That the sentence carries no deception element, and so states the control's failure as the wrong in itself, is the drafter's reading of its structure and placement, not something the complaint says about itself. (Flagged; a reader who wants the deception counts should read the complaint's causes of action rather than this inference.)

(c) The control must cover the surfaces where the choice was made

  1. [ESTABLISHED as an allegation] The same complaint alleged that consumers who enrolled "through devices other than computers and smartphones, such as through the Prime Video application on the Amazon FireStick and Fire TV, they could not cancel via these same technologies. Instead, they had to use the Iliad Flow or call customer service." (Original complaint, paragraph 115; amended complaint paragraph 129.)

(d) The control must not be degraded once it works

  1. [ESTABLISHED as an allegation] In the Epic Games administrative matter the FTC alleged that after players used an "Undo" button to cancel unwanted charges, "Epic took steps to reduce its prominence", specifically that it renamed the button Cancel Purchase and "reduced the size of the button, moved it to the bottom of the screen, and required consumers to push and hold a button on their controller (even though Epic does not require consumers to do so to purchase items in the first place)." (Epic complaint, paragraph 39.)
  2. [ESTABLISHED as an allegation quoting an internal document, which is two hops from a finding] The complaint continues that after making these changes, Epic "observed a roughly 35% decline in the net undo-rate", which the complaint attributes to an email from a Senior Product Manager. The 35% figure is Epic's own words, quoted by the FTC, inside an allegation. It is not an FTC finding and not a court's. (Epic complaint, paragraph 40.)

(e) Posture, stated per matter because it varies

  1. [ESTABLISHED] FTC v. Amazon.com is a federal district court action (W.D. Wash., No. 2:23-cv-00932). It ended in a Stipulated Order for Penalty Judgment, Document 535, filed 2025-09-25. Stipulated means agreed: the order resolves the case without any court adjudicating the complaint's allegations. (Complaint caption; order caption.)
  2. [ESTABLISHED] In the Matter of Epic Games, Inc. is an FTC administrative proceeding (Matter 192-3203) before the Commission itself, not a court action. (Complaint caption: "UNITED STATES OF AMERICA BEFORE THE FEDERAL TRADE COMMISSION".)
  3. [ESTABLISHED] Custom Communications v. FTC is the only matter in this record decided by a court on the merits of anything, and what it decided was procedural. (Opinion.)

(f) Chitika: the whole standard, drafted as a US order term in 2011

This is the most useful matter in the record and it is fifteen years old. It is the only one where the remedy, not just the allegation, reads like the specification this project says is missing.

  1. [ESTABLISHED as an allegation] Chitika offered an opt-out from behavioural-advertising tracking. The complaint alleges that when a consumer clicked it, "the message adjacent to the button changes to" a line reading You are currently opted out. (Chitika complaint, paragraph 10. The interface text is quoted inside quotation marks in the complaint, so it is rendered here without them to keep the quoted span verbatim.)

  2. [ESTABLISHED as an allegation, and this is the persistence failure in its purest form] "From at least May 2008 to February 28, 2010, respondent delivered opt-out cookies that it set to expire after 10 days. Consequently, 10 days after consumers selected respondent's opt-out, the opt-out cookies automatically expired and disappeared from consumers' browsers." (Chitika complaint, paragraph 12.) The control was not broken on the day it was set. It worked, said so, and then quietly stopped, which is precisely the failure this site's self-test is designed to catch.

  3. [ESTABLISHED as an agreed order term] The remedy requires the choice mechanism to "remain in effect for a minimum time period of five (5) years, unless the consumer deletes his or her cookies or takes deliberate action to disable the mechanism." That is a minimum duration, written as an enforceable term. (Decision and Order, Part II.C.)

  4. [ESTABLISHED as an agreed order term] The same paragraph caps friction: the mechanism "shall require no more than one additional click for consumers to exercise their choice(s)". (Decision and Order, Part II.C.)

  5. [ESTABLISHED as an agreed order term] It requires a status display. Within close proximity to the mechanism, Chitika must clearly and prominently disclose "the current status of their choice", expressed in the order as opted in or opted out of collection. (Decision and Order, Part II.C(3).)

  6. [ESTABLISHED as an agreed order term] It requires a scope disclosure, which is the term the parent plan spotted and the one most directly transferable to a feed control: Chitika must disclose "that their choice is specific to the browser they are using, and they need to implement the mechanism again if they use a different browser". (Decision and Order, Part II.C(4).)

  7. [ESTABLISHED as an agreed order term] It requires surface coverage: within ninety days, "within any advertisement that respondent serves as part of online behavioral advertising, include a hyperlink that directly takes consumers to the mechanism required by Part II.C. of this order." The control must be reachable from the surface where the targeting happens. (Decision and Order, Part II.D.)

  8. [ASSUMED, and narrower than it first looks] Mapping claims 17 to 21 onto this site's seven scorecard criteria is the drafter's translation between two vocabularies, not anything the FTC said, and it is honest only if done strictly. Done strictly it gives two clean matches, one near-match, and one inversion:

    • Persistent ("does the choice survive closing and reopening the app, another device, and time?"): matched by the five-year minimum. Clean.
    • Discoverable ("can you find the control without googling it?"): matched by the hyperlink required inside every targeted ad, and supported by the one-click cap. Clean.
    • Clear ("does it say what will change?"): only adjacent. The status display tells the user what state they are currently in, which is a different question from what a choice will do.
    • Scoped ("does it cover Reels, Shorts, Explore, notifications, and suggested accounts, or just one screen?"): not matched, and arguably inverted. See claim 22a.
    • Immediate, Material and Non-circumventing: not addressed at all.

22a. [ESTABLISHED, and it is the most important limitation of this template] On scope, the Chitika order requires disclosure of the gap rather than closure of it. The term obliges Chitika to tell the user "that their choice is specific to the browser they are using, and they need to implement the mechanism again if they use a different browser". The scorecard asks whether a control covers its related surfaces. This order settles for telling the user that it does not. A regulator with the drafting pen chose transparency about a limit over coverage of it, which is precisely the compromise this project should expect to be offered, and should be ready to argue against. (Decision and Order, Part II.C(4).) 23. [ESTABLISHED] The posture is a consent order. Chitika settled; the order was entered without adjudication and without any admission. It binds Chitika and nobody else. (Decision and Order, caption and recitals.)

Quotes (attributed record)

Attribution convention for this record. Every span below is from a complaint or an opinion. A complaint quote is what the FTC asserted; it is never a finding. Where the FTC is itself quoting a company's internal document, that is marked, because it is two hops from an adjudicated fact.

Eighth Circuit, Custom Communications v. FTC (the court's own voice)

  • "we need not address Petitioners' other substantive challenges to the Rule"
  • "While we certainly do not endorse the use of unfair and deceptive practices in negative option marketing, the procedural deficiencies of the Commission's rulemaking process are fatal here."
  • "the party-specific vacatur requested by the Commission is not feasible"
  • On what the amended rule required: "require disclosure of material terms, express consumer consent, and a simple cancellation mechanism"

FTC allegations, Amazon Prime complaint (the FTC's assertions, not findings)

  • Paragraph 119, on the button labelled End Membership: "did not end membership. Rather, it took the consumer to the Iliad Flow."
  • Paragraph 115: "they could not cancel via these same technologies"

FTC allegations, Epic Games complaint (the FTC's assertions, not findings)

  • Paragraph 39, after Epic renamed the button Cancel Purchase: "reduced the size of the button, moved it to the bottom of the screen, and required consumers to push and hold a button on their controller"
  • Paragraph 40, the FTC quoting Epic's internal email: "observed a roughly 35% decline in the net undo-rate"

FTC allegations and agreed order terms, Chitika (2011)

Complaint, paragraph 12, the persistence failure:

  • "respondent delivered opt-out cookies that it set to expire after 10 days"
  • "the opt-out cookies automatically expired and disappeared from consumers' browsers"

Decision and Order, Part II, the remedy. These are AGREED terms, not adjudicated findings:

  • Duration: "remain in effect for a minimum time period of five (5) years, unless the consumer deletes his or her cookies or takes deliberate action to disable the mechanism"
  • Friction: "require no more than one additional click for consumers to exercise their choice(s)"
  • Status: "the current status of their choice"
  • Scope: "that their choice is specific to the browser they are using, and they need to implement the mechanism again if they use a different browser"
  • Surface coverage: "include a hyperlink that directly takes consumers to the mechanism required by Part II.C. of this order"

Tensions / open questions

  • This ledger is a third of its subject. The parent research plan lists roughly twelve matters. Three are distilled here. The claim that the FTC has enforced these prongs "for years", across a body of practice, is plausible on this evidence and not established by it. Do not put the broad version on the site until more of the queue is verified. The unbuilt matters are named in the Coverage note.
  • Allegations are doing a lot of work. Two of the three prongs rest on complaints that ended in agreed orders. That is genuinely weaker than the parent plan's framing implies, and the site should say "the FTC has repeatedly alleged, and companies have repeatedly agreed to stop" rather than "the FTC has enforced".
  • Nothing here is about a feed. Every matter concerns subscriptions, purchases and cancellation. Applying the reasoning to a recommendation control is this project's argument. It is a good argument and it is not precedent.
  • The vacated rule may return, or may not. The parent plan records the replacement as at ANPRM stage with a closed comment record and no NPRM as of 2026-07-25. That was not re-checked for this pass and should be before the site describes the current rulemaking posture.
  • Was the Amazon stipulated order's content examined? Only its caption and posture. Its injunctive terms are cached and undistilled, and they are the most likely place to find an American analogue of the Dutch persistence order.
  • The Chitika order is fifteen years old and binds one company. Its terms are the best drafting available to this argument, and they are a consent order against a small advertising firm from 2011, entered without adjudication or admission. Nothing about it obliges anyone else, and the technology it describes is cookie-based tracking, not a recommendation feed. Its value is as a drafting precedent and a proof that the terms are writable, not as authority.
  • The five-year term has a carve-out that matters for any feed analogue. The duration holds "unless the consumer deletes his or her cookies or takes deliberate action to disable the mechanism". In a cookie world that is unavoidable; the mechanism is a cookie. A feed control stored server-side against an account has no equivalent excuse, which arguably makes the Dutch persistence order the stronger template and Chitika the better-drafted one.

Sources