Opt-in is treated as a formality until the day the quality rating drops and you cannot reach anyone. It is worth getting right before that day.
Most guides treat opt-in as a compliance checkbox somewhere below choosing a template category. That ordering is backwards.
The WhatsApp Business Platform does not primarily police you with fines. It polices you with reach. Recipients who did not expect your message block or report it, blocks and reports lower your quality rating, and a lowered rating cuts your messaging limits — eventually to the point where you cannot start new conversations at all.
The asymmetry is what makes this worth taking seriously. Messaging a few hundred retailers who never opted in might produce a handful of orders. It might also cost you the ability to reach the two thousand who did. There is no version of that trade that pays.
Meta business messaging policy asks for opt-in that is clear on three points. The wording moves around; the substance has been stable.
It was affirmative. The person did something to say yes. A pre-ticked checkbox is not affirmative. Neither is a phone number collected for a delivery that later gets used for broadcasts.
It was specifically about WhatsApp. Agreeing to be contacted is not agreeing to be contacted on WhatsApp. The opt-in has to make the channel obvious.
Your business was identified. The person knew which business would be messaging them. Consent collected by a reseller or an aggregator on their own behalf does not transfer to you.
Notably, Meta does not mandate a particular mechanism. A web form, a checkbox at onboarding, a physical form with a signature, an IVR confirmation, or a message the retailer sends you first can all work. What matters is that you can show the three points above were satisfied.
If someone asks how a given number got onto your list, you want to answer in one line rather than a search.
Store, against each number:
That last two matter more than they seem. Opt-in language changes as forms get redesigned, and a record that says only which form was used cannot tell you what that form actually said in March. Storing the wording alongside the record makes the answer specific instead of approximate.
Retailers who message you first are the easiest case. An inbound message opens a customer service window, and asking for permission inside that window is both compliant and natural — the retailer is already in the conversation. This is also the cheapest path, since it uses a service conversation rather than a marketing template; the cost mechanics are in WhatsApp template categories and what they cost.
Opt-in gets attention because it happens at onboarding, when someone is paying attention. Opt-out happens later, quietly, and is where the damage actually accumulates.
Two failure modes account for most of it.
Only recognising an exact keyword. A retailer who types "stop sending" or "band karo" or "please stop these messages" has opted out as clearly as one who typed STOP. A system that matches only the exact keyword will keep messaging all three, and all three will eventually report the number instead.
Treating opt-out as marketing-only, silently. Stopping promotional sends while continuing order confirmations is defensible and usually correct — the retailer still wants to know their order shipped. But it has to be a deliberate decision, and the retailer should be able to stop everything if they want to. What is not defensible is continuing marketing under a utility label because the template was categorised that way.
The rule worth adopting: honour any plain-language stop request immediately and confirm once that it has been actioned. A single confirmation message is not a violation; it is the thing that stops the person reporting you because they think nothing happened.
Numbers and steps to adapt, not a template to copy.
Step three is the one most worth defending under pressure. Leading with a scheme broadcast to a freshly onboarded list is the fastest way to accumulate blocks, because the recipient has no context for why you are in their WhatsApp. Leading with something they wanted establishes it.
These are separate systems that feel like one, because opt-in quality is the biggest single input into rating.
Quality rating reflects how recipients respond to your messages — blocks, reports, and to a degree whether people engage at all. Clean opt-in produces recipients who expected the message; poor opt-in produces recipients who did not, and those recipients block.
Practically, this means volume is not the risk factor people assume. Sending a lot of expected messages to people who opted in is fine and grows your limits over time. Sending a few unexpected ones to people who did not is what causes trouble. The setup mechanics behind all of this — number registration, display name, limits — are covered in the WhatsApp Business API setup guide, and the WhatsApp Message Formatter is useful for getting the message itself readable once you have earned the right to send it.
Opt-in is not a legal formality bolted onto a messaging system. On this platform it is the mechanism that decides how many people you are allowed to talk to next month.
Collect it explicitly, record what was actually shown, lead with messages people want, and treat any request to stop as final the first time it is made. None of that is difficult. It is just easy to postpone until the rating drops, at which point recovering it takes far longer than collecting the consent would have.
The questions that come up when a distributor first sets up WhatsApp messaging at scale.
Meta requires that the person gave clear, affirmative permission to receive messages from your business on WhatsApp, that it was obvious they were opting in to WhatsApp specifically rather than to communication generally, and that your business was clearly identified as the sender. A pre-ticked box, a phone number collected for delivery purposes, or consent buried in general terms does not meet that bar.
An existing commercial relationship is not the same as opt-in, and the safest reading is that you need to collect it explicitly even from long-standing customers. The practical route is to collect it during a conversation the retailer starts — if they message you first, you can reply within the customer service window and ask for permission to send order updates going forward.
Recipients block or report your number, which lowers your quality rating. A low rating reduces your messaging limits and can move your number to a restricted state where you cannot start new conversations at all. The damage is not a fine but a loss of reach, and it affects every retailer on your list rather than only the ones who complained.
Immediately, and without requiring the person to repeat themselves. Someone who asks to stop and keeps receiving messages will report the number rather than ask twice, and a report costs far more than the messages you would have sent. Your system should recognise a plain-language stop request, not only an exact keyword, and stop marketing sends to that number straight away.
The FlowKartAI team builds WhatsApp-native ordering for Indian B2B distributors and the kirana stores they serve. We write about distribution economics, GST compliance, and the practical side of putting AI in front of retailers who have never opened an app.
FlowKartAI parses natural language WhatsApp messages into ERP-ready orders in seconds.
Try Free Demo