Before your firm can text a client back, a carrier has to say yes
Everybody asks about the TCPA. You know that subject better than we do. The thing that actually stops a firm texting clients back is smaller, duller and entirely mechanical: US mobile carriers will not carry business texts from a number that has not been registered with them first, and the registration takes weeks. It is the single item that decides when this can go live.
This is not the consent question
Consent, timing and content are your responsibility as the sender and they sit under law you already advise clients on. Nothing on this page touches that, and we are not qualified to.
What this page is about is the layer underneath: the agreement between whoever sends your messages and AT&T, T-Mobile and Verizon about which numbers are allowed to send business traffic at all. That agreement has a name, a form, a fee and a queue.
What has to be registered, and with whom
Business texting from an ordinary ten-digit US number runs under a scheme the industry calls 10DLC. Two things get registered with a body called The Campaign Registry, which sits between the messaging providers and the carriers.
First the brand — the legal entity doing the sending. That means the registered company name, the tax identification number, the address and the website, and the details have to match what the authorities already hold, because they are checked against third-party records rather than taken on trust.
Then the campaign — the specific thing you are going to send. You describe the use case, provide sample message text, and state how people opt in and how they opt out. A missed-call reply is a straightforward case to describe, but it still has to be described.
What happens if you skip it
Since February 2025 the major US carriers block unregistered business traffic on ordinary numbers outright. It is not throttled, not flagged, and not delivered late.
The failure mode is the dangerous part. Everything on your side reports success. The message is accepted, the automation records that it sent, and the text never arrives on the handset. A firm can run that for a month and conclude that clients simply do not reply to texts.
What it costs and how long it takes
Checked in August 2026, and these move: brand registration is a one-off fee of around four dollars, and a campaign carries a recurring fee that is usually a few dollars a month, higher for some use cases. Optional vetting, which raises the volume a number is trusted with, is a further one-off fee of roughly forty dollars.
The money is not the issue. The time is. End to end, submission to being able to send, is commonly one to four weeks — the brand is often cleared in a day or a few, and the campaign review is the longer half. Anything that does not match, a company name recorded slightly differently or a website missing the language a reviewer expects, sends it back round again.
Which is why we start this during onboarding rather than on go-live day. A build that is finished and waiting on a queue is a bad week for everyone, and it is entirely avoidable by filing first.
Why the reviewer reads your privacy policy
Part of the campaign review is somebody looking at your website. They are checking that the way you say you collect phone numbers matches what you say you will do with them, and one specific sentence is looked for: that mobile numbers and messaging consent are not shared or sold to third parties for marketing.
A privacy policy that does not say it plainly is a routine reason for a campaign to be rejected or suspended later. It is a five-line addition to a page you already have, and it is the cheapest thing on this list to get right in advance.
What STOP has to do
Anyone receiving a message must be able to reply STOP and have it take effect immediately and permanently. This is not optional and it is not a setting.
Every message we send carries the opt-out line, and it is added by the builder rather than typed into the message template, so nobody can drop it by editing the wording. The list of numbers that have opted out is checked before every send — the provider will suppress the message anyway, but making the request at all is the wrong behaviour, not just an ineffective one.
Opt-outs are recorded from the provider's structured error code and never from the text of an error. That sounds like a detail and is not: every message we send ends in the word STOP, so any error that echoes the message back matched a plain text search, and so did an unrelated provider fault reading "stopped responding". Both were being filed as customers opting out, which suppressed the number permanently and raised no alert.
Quiet hours, which we enforce whether or not you ask
Automated messages are bounded to daytime hours where the recipient is. Both of the services we run that send messages check the window before sending, defaulting to eight in the morning until nine at night.
Missed-call text-back drops a message outside that window rather than queuing it, because the entire value is replying within seconds and a text arriving nine hours after the call is worse than no text. A call at two in the morning gets nothing, and the client guide says so on the first page rather than the last.
One limit worth stating: the window is your business timezone, not the recipient's. A firm taking calls across several US timezones can still be an hour outside at the edges, and closing that properly needs a lookup we do not have.
What this means for a go-live date
If somebody promises you business texting live this week, either the registration is already done in somebody else's name or the messages are not going to arrive. Those are the only two options.
We file in your name, on your account, because the registration is the entity's and not ours. That is slower than borrowing somebody's shared number and it is the reason the number keeps working if you stop working with us.
Before you ask
The things people actually want to know.
Do we have to do the registration ourselves?
We prepare and file it, but it goes in your firm's name against your own details, because the brand being registered is the legal entity doing the sending. You will need to supply the registered name, the tax identification number and the address, and they have to match the records the authorities already hold.
Can we use a number you already have registered?
No, and you should be wary of anyone offering that. Sending your firm's messages through somebody else's registered brand means the number is not yours, the compliance record is not yours, and you lose both the day the arrangement ends.
What if we already text clients from a mobile?
A person typing a text on a handset is a different thing from an automated send and is not what this scheme covers. The registration applies the moment the sending is automated, which is the moment this becomes useful.
Is any of this legal advice?
No. We can tell you what the carriers require to carry your traffic, because that is operational and we deal with it constantly. Whether a particular message to a particular person is lawful is a question for you, and our terms say the same.