Your Business Number Got Flagged. Now What?

The first sign is rarely an error message. It is usually a person: someone mentions they stopped getting your texts, and when you check, everything looks fine on your end.

That gap — between your system reporting success and messages not arriving — is what makes this failure hard to catch and easy to make worse.

First: confirm it is actually a flag

Several things look identical from the outside. Rule them out before acting.

Check whether sends are erroring or succeeding. A hard block usually produces errors. Silent non-delivery — sends succeeding, nothing arriving — points at filtering or a shadow restriction, which is more common and more insidious.

Check delivery receipts, not delivery. If your platform reports messages as delivered but read and delivery confirmations have stopped, the messages are likely arriving somewhere the recipient is not looking.

Check inbound. A line that normally receives replies and has gone completely silent for a day or more, while still sending, is a strong signal. People replying is the hardest thing to fake and the first thing to disappear.

Test with a colleague on each network. Send to a phone you control on a different carrier, and to an iPhone and an Android. A failure on one network and not another tells you it is carrier-side; failure everywhere points at the account or the number itself.

The first 24 hours

Stop sending. This is the whole job. Continuing to send into a flagged number converts a temporary restriction into a permanent one, and it is the single most common way people make this unrecoverable.

Do not immediately switch to a new number and resume at the same volume. The behaviour is what got flagged. Moving it to fresh infrastructure reproduces the result on a delay, and now you have burned two numbers.

Work out which layer failed. Carrier-level filtering, Apple account suspension, and a recipient-side spam classifier are three different problems with three different fixes. The tests above tell you which.

What actually gets a number back

Honestly: sometimes nothing does. Worth knowing before you spend a week on it.

For carrier filtering, the path is your provider. Carriers do not take appeals from senders directly; they take them from registered messaging providers. If you are on unregistered infrastructure there is often no route at all — which is one practical argument for using a platform that holds proper 10DLC registration.

For an Apple account restriction, there is a support path, and it is worth attempting once. Be straightforward about what the account does. Accounts have been restored as a one-time exception; accounts have also been permanently declined. Do not build a plan around it succeeding.

For on-device filtering, there is no appeal because there is no authority — each recipient's phone decided independently. The only remedy is changing what you send and to whom, and being saved as a contact.

Coming back

If sending resumes, resume slowly.

Wait a full 24 hours. Then run at half your previous ceiling for at least a week before climbing. A flagged number that returns to prior volume immediately tends to be flagged again, faster, and the second time is usually final.

If you are starting on a new number instead, treat it as genuinely new: a warm-up of roughly 25 a day in week one, 50 in week two, 75 in week three, 100 from week four. Halve those for a brand-new account and stretch it by two weeks.

One thing people miss when starting over: hardware carries history too. Apple associates accounts by device fingerprint, so a clean account on previously flagged hardware is frequently linked to the old one — sometimes within days. If you are rebuilding, the account is not the only component that needs replacing.

The part that matters more than recovery

Recovery is a tactic. It does not address why it happened.

In almost every case, a flagged business number was sending to people who had not asked to hear from it. Not necessarily maliciously — a purchased list, an old list, an import from a CRM where consent was assumed rather than recorded, an onboarding flow that texts before anyone opts in.

The reliable fix is structural: stop being the one who starts the conversation. A message in a thread the customer opened has no complaint button sitting next to it, no unknown-sender treatment, and no pattern that looks like outreach. It is not a mitigation; it removes the condition.

That is the rule senderZ enforces at the routing layer — an iMessage line never sends to a number that has not messaged it first. It came out of watching this failure three times and eventually accepting that hardening the setup was treating a symptom.

Monitoring so it does not surprise you again

Three things to watch, in order of how early they warn you:

Inbound silence. A line that normally gets replies and goes quiet while still sending. Cheapest and earliest signal available.

Delivery-receipt rate. Delivered divided by sent. A collapse while sends keep succeeding means something is absorbing messages.

Consecutive send errors. The loudest signal, and the latest — by the time this fires the damage is usually done.

We alert on all three per line, and quarantine a line automatically when the error signal crosses a threshold. The important part is not the specific thresholds; it is that somebody finds out from a system rather than from a customer mentioning it in passing.