Skip to content

Why Texts Get Flagged by Carriers

Every text you send from Goliath passes through the US mobile carriers (AT&T, T-Mobile, Verizon) before it reaches a phone. Carriers run their own spam filtering on top of your registration, and they block a lot of it. This page explains what they block, why, and what you can do about it.

There are two separate gates, and a message has to pass both:

  1. Registration. Is your business and your campaign registered with the carriers (A2P 10DLC)? Goliath walks you through this in Campaigns → Registration.
  2. Content and behavior. Every individual message is scanned in real time, and weighed against your sending history and complaint rate.

You can be perfectly registered and still have every message filtered by the second gate. That is the situation most people are in when they say “my texts stopped delivering.”

Some categories are not “risky” — they are not allowed, and no rewording makes them allowed. Sending them can get your campaign rejected or terminated, and repeat violations can take your whole brand registration with them.

The prohibited list includes:

  • Unsolicited real estate inquiries — offers to buy someone’s property when they never asked to hear from you. Note the word unsolicited: the same message is fine when the recipient gave you permission to text them.
  • Lead generation and affiliate marketing
  • High-risk financial services (loans, loan forgiveness, credit repair, debt collection, debt relief)
  • Cryptocurrency and investment opportunities
  • Gambling
  • SHAFT content: Sex, Hate, Alcohol, Firearms, Tobacco (and cannabis/CBD, even where it is legal)

Carrier filtering weighs behavior first, words second. Ranked by how much it matters:

Texting people who never opted in is the single strongest signal. It drives complaints and STOP replies, which destroy your sender reputation, after which carriers filter everything from your number regardless of what it says.

Consent also cannot be bought or transferred. A purchased list does not carry consent, and consent given to another company does not extend to you. It has to be given to you, for this kind of message.

Blasting full volume from a number with no sending history looks exactly like spam. Warm up gradually.

3. Sending something different from what you registered

Section titled “3. Sending something different from what you registered”

Your campaign registration tells carriers what you plan to send. Sending a different use case, or far more volume than you declared, is a named reason for suspension.

4. Spreading the same message across many numbers

Section titled “4. Spreading the same message across many numbers”

Rotating numbers to dodge per-number filtering is called snowshoeing. Carriers detect it and fine it — up to $1,000 per violation.

The weakest of the five, but the one you control most directly:

  • Public link shorteners (bit.ly, tinyurl) — AT&T blocks these outright. Use a branded domain instead.
  • No opt-out language — every first message should carry “Reply STOP to opt out.”
  • Not identifying yourself — say who you are in the message.
  • ALL-CAPS words, excessive punctuation (!!!), emoji
  • A phone number in the message body
  • Urgency and money bait: “act now,” “limited time,” “guaranteed,” “top dollar”
  • Asking for several things at once (reply and click and call)
  • In the campaign editor, each message shows a carrier-risk indicator. Open it to see exactly which signals fired and get an AI rewrite that lowers the content risk.
  • At registration, the same check runs on the two sample messages before you submit them, so a rejection is caught before it costs you a re-registration.
  • Before you launch, the launch confirmation repeats the check across every message in the sequence.

A well-formed first message does four things:

  1. Says who you are
  2. Says why you are contacting this person
  3. Makes one low-pressure ask
  4. Includes opt-out language

Hi {{firstName}}, this is Max at Goliath Homes. You asked about selling {{propertyAddress}} on our site last week. Would it help if I sent over a few recent comps? Reply STOP to opt out.

Q: My texts were delivering fine and suddenly stopped. What happened?

Section titled “Q: My texts were delivering fine and suddenly stopped. What happened?”

You most likely crossed a reputation threshold. Filtered messages and complaints compound: each one lowers your number’s reputation, which increases the filter rate, which lowers it further. Once that starts, it affects every message from that number, not just the ones that triggered it.

No — and it will make things worse. Moving to fresh numbers to escape filtering is snowshoeing, which is itself a finable violation. Fix the underlying consent and content problems instead.

Q: Does the AI rewrite make my message compliant?

Section titled “Q: Does the AI rewrite make my message compliant?”

No. It lowers the content risk — shorteners, caps, urgency wording, missing opt-out. It cannot create consent, so it cannot make an unsolicited campaign compliant.

The recipient actively agreed to receive texts from your business about this subject — a checkbox on your form, a written opt-in, a reply to a message they initiated. Keep a record of it; carriers can ask.

Contact support. You will generally need to correct the underlying problem and re-register, which takes time and may incur another registration fee — which is why the checks above run before you submit rather than after.