Nothing generates complaints faster than a business that keeps messaging after being told to stop. It is also entirely avoidable, and the failures follow a small number of predictable patterns.
What must happen, and how fast
- Recognise the intent to stop.
- Suppress the contact immediately — within seconds, not at the next sync.
- Confirm once, briefly.
- Record who, when and by what wording.
- Ensure every downstream system respects the suppression.
Recognise more than the word STOP
People do not read instructions. They write what they mean. Match case-insensitively on STOP, UNSUBSCRIBE, CANCEL, END, QUIT, OPTOUT, OPT OUT, REMOVE, and their equivalents in the languages you send in — in India, that means at least Hindi transliterations.
Also match plain phrases: "stop messaging me", "don't message me again", "remove me", "not interested". A customer who wrote a sentence rather than a keyword has still withdrawn consent, and treating that as unrecognised input is the difference between a stop and a report.
Do not make people confirm. Replying "Are you sure? Reply YES to confirm" to someone who just asked you to stop is the most reliable way to turn an opt-out into a block-and-report. Suppress first, confirm once, and offer the reversal route in that confirmation.
The confirmation message
One message, no argument, no offers, no survey. "You have been unsubscribed and will not receive further marketing messages from Northline. You will still get updates about active orders. Reply START at any time to resume."
That does three useful things: it confirms, it clarifies what continues, and it leaves the door open without pleading.
Scope the suppression correctly
Most opt-outs are from marketing, not from transactional messages. Someone who stops promotions still wants to know their order shipped. Applying a blanket suppression creates a different complaint — "why didn't you tell me my delivery was delayed" — and losing the transactional channel is worse for both sides.
If the wording is ambiguous, suppress marketing and say plainly what will continue. If they want everything stopped, they will say so, and you honour that.
Where suppression leaks
- Campaigns built from a stale export. A CSV pulled before the opt-out re-includes the person. Always resolve suppression at send time, never at list-build time.
- A second system with its own contact list. Your CRM knows; your email tool does not.
- Duplicate contacts. The same person under two records with formatting differences in the phone number.
- Scheduled sends queued before the opt-out. Check suppression at dispatch, not at scheduling.
- Re-imports. A fresh upload that overwrites the consent flag with a default.
Normalise numbers, or duplicates will bite
Store every number in one canonical international format. A contact saved as 9910146307 and another as +91 99101 46307 are the same human being and two rows in your database — and suppressing one does nothing for the other. Normalise on write, deduplicate on the normalised value, and this class of failure disappears.
Audit it occasionally
Once a quarter, take twenty contacts who opted out more than a month ago and confirm no system has messaged them since. It takes fifteen minutes and it is the only way to find a leak before a customer does.