A customer replies to your text, but not with "STOP." Maybe it's "please don't contact me," maybe it's something messier and half-relevant to something else entirely. Do you have to treat that as an opt-out? Can you keep texting until they use the actual keyword your compliance footer told them to use?
For years the safe answer was: no keyword, no obligation. That assumption doesn't hold anymore, and a federal court just spent an opinion drawing the line between a reply that counts and one that doesn't. Here's what the rule actually says, what the new case tested, and what it means for how you configure opt-out handling.
The Old Playbook: "Text STOP" as the Only Way Out
Most SMS compliance footers still read some version of "Text STOP to opt out" — and for a long time, that language did double duty. It told consumers what to do, and it gave businesses a defensible position that anything else wasn't a real revocation request. If a consumer texted back "quit bugging me" instead of the magic word, plenty of programs kept sending and pointed to the footer if anyone complained.
That position was never bulletproof, but it worked often enough that a lot of opt-out logic got built around a short, fixed keyword list and nothing else.
The FCC Changed the Standard: Any Reasonable Means
In February 2024 the FCC adopted a Report and Order in CG Docket No. 02-278 (FCC 24-24) that formalized what "reasonable means" of revoking consent looks like for autodialed and prerecorded calls and texts. The rule took effect April 11, 2025, and the core piece has been governing SMS programs ever since.
The standard: a consumer can revoke consent through any reasonable method that clearly conveys they don't want to be contacted anymore. Businesses can no longer designate one exclusive opt-out mechanism and ignore everything else. Specifically:
- Replying STOP, QUIT, END, REVOKE, OPT OUT, CANCEL, or UNSUBSCRIBE to a text is automatically — "per se" — a reasonable means of revoking consent. No argument needed.
- Other words or phrases aren't excluded. If a reply doesn't use one of those seven words but still clearly asks to stop being contacted, it can count too.
- When it's disputed whether a given reply was reasonable, the standard is a totality-of-circumstances test — what did the message actually say, in context — with a rebuttable presumption in the consumer's favor if they can show the request was made.
- Once a revocation is valid, businesses have 10 business days to stop sending.
The seven magic words didn't go away. They just stopped being the only words that matter.
What "Reasonable" Doesn't Mean: Tom v. DeLancey Street Group
The test case for how far "any reasonable means" actually stretches came down on August 5, 2026, in Tom v. DeLancey, 2026 WL 2255107 (E.D.N.Y.) — the first real stress test of the post-2025 rule that most compliance teams have seen.
DeLancey Street Group, working with a lead generator called MCA Justice, texted David Tom offering to cut his business payments: "Would it help if we cut your daily/weekly payments by 80% and convert to monthly? Reply 'Yes' for more information. Text STOP to opt out." Tom replied "Yes," and more texts followed. At some point he sent back: "Don't text me email me only– about to board a plane no cell signal so WiFi works a lot better."
No magic word. So the case turned entirely on whether that reply was a "reasonable means" of revoking consent under the totality of the circumstances. The court said no, for two reasons: it never used STOP or any of its per se equivalents, and read in context, it looked like a request to switch to email while Tom was traveling — not a clear statement that he wanted all future contact to end. The court dismissed the claim.
The same opinion also rejected a separate theory that the text violated caller ID rules by not identifying the sender clearly enough. The court held there's no private right of action for that under the Truth in Caller ID Act, consistent with what TCPAWorld describes as the majority position among courts that have looked at it.
The takeaway isn't "ambiguous replies are safe to ignore." It's narrower than that: the reasonable-means standard requires a clear expression of intent to stop, and courts are willing to look past magic words to figure out whether that intent was actually there. A message that's genuinely ambiguous about a temporary channel switch is a different animal from a message that plainly says "leave me alone."
The Part of the Rule Still on Hold
One piece of the 2024 order — sometimes called the "revoke-all" requirement, at 47 C.F.R. § 64.1200(a)(10) — would require a business to treat an opt-out on one topic as applying to all future unrelated messages from that same sender. The FCC waived that specific piece after financial institutions argued they couldn't retool multi-department messaging systems in time. The waiver has been extended twice and currently runs through January 31, 2027.
Practically: if you're running more than one message stream from the same number — say, payment reminders and separate marketing texts — you're not yet required to treat a stop on one as covering the other. That changes at the end of January 2027. Everything else in the rule, including the per se keyword list and the totality-of-circumstances standard applied in Tom, is already in force.
What an Ignored STOP Actually Costs
The flip side of Tom is what happens when the opt-out request isn't ambiguous at all. In Hiller v. The Money Source Inc., the mortgage servicer agreed to pay $1.5 million to settle a class action alleging it kept placing prerecorded calls to consumers who had directly asked it to stop. There was no interpretation problem to litigate — the requests were clear, the calls continued anyway, and that's most of what a plaintiff needs.
Enforcement against clear-cut violations isn't slowing down either. On August 11, 2026, an Ohio federal court entered an $83.7 million judgment against Aaron Michael Jones, a robocall operator the Ohio Attorney General's office pursued for billions of illegal calls, payable within 10 days of the order.
Read together, the message for a text program is straightforward: courts aren't eager to manufacture opt-outs out of ambiguous replies, but they have zero patience for programs that ignore unambiguous ones.
Building an Opt-Out Process That Holds Up
A compliance footer that only lists STOP isn't wrong, but treating it as the only thing that counts is the exposure. A process that survives scrutiny needs a few pieces:
- Auto-suppress the seven per se words instantly. STOP, QUIT, END, REVOKE, OPT OUT, CANCEL, UNSUBSCRIBE — no human step, no delay, because these are automatically reasonable under the rule.
- Route everything else that reads like a complaint or a stop request to a human, fast. "Leave me alone," "remove me," "wrong number, stop" — none of those are on the magic-word list, but a fact-finder applying totality of circumstances could easily find them clear enough. You have 10 business days to act once a request is valid; don't spend nine of them not looking at the message.
- Keep a timestamped record of what was sent, what came back, and what your system did with it. Tom turned on exactly this kind of message-by-message reading. The business that can show its records match its stated process is in a much stronger spot than one relying on "we assumed it wasn't a real opt-out."
- Don't rely on your own footer language as a shield. "Text STOP to opt out" is fine as an instruction, but it can't function as the exclusive method anymore — the rule doesn't allow a business to designate one channel and disregard reasonable requests made another way.
If you're managing this through Android Texter, auto-reply rules can be configured to catch keyword-triggered suppression automatically, and every inbound message lands in a searchable, timestamped inbox — which is the record you want on hand if a stop request is ever disputed later. Configure the suppression keywords before your first campaign, not after your first ambiguous reply.
Frequently Asked Questions
Do I have to text back "STOP" specifically to opt out of texts?
No. Since the FCC's rule (FCC 24-24) took effect April 11, 2025, consumers can revoke consent through any reasonable means, not just the word STOP. Seven words — STOP, QUIT, END, REVOKE, OPT OUT, CANCEL, UNSUBSCRIBE — are automatically valid. Other clear phrasing can count too, evaluated under a totality-of-circumstances test.
What happens if a customer's opt-out message is unclear?
Courts look at context, not just keywords. In Tom v. DeLancey, a reply asking to switch to email while boarding a flight wasn't treated as a valid opt-out because it didn't clearly express a wish to stop all contact — it read as a temporary channel request. Genuine ambiguity doesn't automatically bind the sender.
How fast do I have to stop texting after a valid opt-out?
Within 10 business days of receiving a reasonable revocation request, per the FCC's 2024 rule. A one-time confirmation text acknowledging the opt-out is permitted as long as it contains no marketing content and is the only follow-up message sent.
Does opting out of one type of text stop all future texts from the same company?
Not yet, for most senders. The FCC's "revoke-all" requirement — treating an opt-out on one topic as covering unrelated messages too — has been waived and currently isn't required until January 31, 2027. Until then, per-topic suppression is generally sufficient.
Can a text be illegal just because it's missing sender ID information?
Not on its own, at least as a private lawsuit. In Tom v. DeLancey, the court held there's no private right of action under the Truth in Caller ID Act for a text that doesn't clearly identify the sender — consistent with the majority position among courts that have addressed the issue.
None of this changes what TCPA compliance actually requires: get consent, document it, and honor stop requests you can reasonably recognize as stop requests. Getting the mechanics right is on you regardless of what platform sends the message.