WhatsApp Business API for grocery stores: setup, rules and real costs

The app is fine until these start happening to a grocery store:

• Orders on personal phone — no system, no scale. Hundreds of WhatsApp orders on your personal phone. Orders get lost. You can't track or organise. It's unsustainable.
• Zepto/Blinkit taking your regular customers. Your morning regulars order from Blinkit now. Not because you're worse — because they don't get reminders to order from you.
• Weekly offers not communicated proactively.

The app cannot send on a trigger, share one number across staff, or read your records. The API can. And customers keep asking the same things: customers ask whether an item is in stock, delivery slots, minimum order value, and whether they can order their usual list again.

• A verified Meta Business account in the grocery store's legal name, backed by a registration, GST or utility document.
• A number not active on the WhatsApp app — your current one works once it is removed from the app.
• A display name matching the business, and a website or public page Meta can check.
• Opt-in from every customer you message first. Home-delivery orders and the billing counter, where regular customers already share their number.

1. Connect the number through Meta's embedded signup inside Reputoo, and submit business verification — you can start sending within Meta's starting limit while it is reviewed.
2. Get your first templates approved for the 2 routine messages listed below; anything outside a customer's 24-hour reply window must be a template.
3. Wire the triggers to your records. What to connect: A billing POS and sometimes a delivery app; the order history is what makes reordering work.
4. Put staff on one shared inbox, with conversations assigned so two people never answer the same customer.

There are three ways a grocery order can arrive over the API, and most stores end up using two of them.

A WhatsApp catalogue

Products with a photo, a price and an "add to cart" button, browsed inside the chat. It works well for packaged goods with a fixed price — a 1 kg pack of atta, a brand of oil, a box of detergent. It works badly for loose produce, where the price changes daily and the weight is only known once it is on the scale. Keeping a catalogue of a few thousand items accurate is also real work; start with the few hundred items your regulars buy most.

A typed list or a photo of one

This is how customers already order by WhatsApp from a neighbourhood store: "2 kg onion, 1 Amul butter, 5 kg rice — the usual one". It needs no setup and suits everything, including loose produce. The cost is that someone has to read it. On the API that someone works from a shared inbox instead of one phone, so two staff can process lists in parallel and nobody's personal phone holds the store's orders.

A link to an order page

If you already have an online store, a link in the chat hands the customer to it. It is the most structured option and the least conversational; it suits customers who order big baskets and want to see totals before they commit.

A sensible default: catalogue for your top packaged items, lists for everything else, and the link for customers who ask for it.

Before writing a single template, list what goes out today by phone, SMS or the WhatsApp app. For most grocery stores it comes down to these:

Six of the eight are utility messages. That matters, because in India a utility message costs ₹0.115 against ₹0.8631 for marketing on Meta's July 1, 2026 rate card. A store handling 40 orders a day that sends three updates per order sends about 3,600 utility messages a month — roughly ₹414 in Meta fees before GST. A weekly offer to 1,500 opted-in customers adds about 6,000 marketing messages, or ₹5,179. The offers, not the order updates, are the bill to watch.

An out-of-stock item is the moment a grocery order stalls. The picker calls the customer, the customer does not answer, the bag waits. On the API this becomes one utility template with reply buttons:

Hi Meena, the Fortune sunflower oil 1 L in your order is out of stock today. Replace it with Saffola Gold 1 L (₹15 more), or leave it out?
[Replace] [Leave it out]

Three rules make this work in practice. Send it while the order is still being picked, not after the bag is sealed. Set a default for no reply — most stores leave the item out after 15 minutes rather than guess. And record the answer: a customer who always accepts one brand as a replacement for another should not be asked twice.

Many neighbourhood stores run a monthly credit account for regular families. Reminding people about money is where a store's tone matters most, and WhatsApp's rules happen to push in the right direction: a statement is a utility message only if it stays a statement.

• Send it on the day you agreed with the customer, not whenever someone remembers.
• Show the items or bill numbers, the total, and how to pay — your UPI ID or payment link — and nothing else.
• Do not add "this week's offers" to the bottom. Meta reads the whole template, and a promotional line turns a cheap utility message into a marketing one.
• Leave disputes about the amount to a person. A customer who replies "I didn't buy this" should reach the owner, not a script.

A store with more than one branch has a choice to make before it registers numbers.

• One number for all branches. Customers keep one contact, and the order is routed to the right branch by delivery area. Simple for customers, but the shared inbox needs clear assignment so each branch sees only its own orders.
• One number per branch. Each branch has its own WhatsApp number under the same business account, its own display name ("Sharma Stores – Sector 15") and its own templates. Clean for staff; customers who shop at two branches have two chats.

Most neighbourhood chains start with one number and move to per-branch numbers only when branches have separate teams and delivery zones. Either way, customer records and consent belong to the business, not to a branch manager's phone — which is the real gain over running branches on the WhatsApp app.

Whichever you choose, write the rule down: which branch serves which pin codes, who answers when a branch is closed, and where a customer's complaint goes if the order came from one branch and the delivery from another.

Meta rates every number on how customers react. For grocery stores, the risk is usually messaging the whole contact list about something only some of them asked for. Send reminders only to people with a booking or order, promotions only to people who opted in to them, and never use a bulk-sender tool on the ordinary app — that is the route to a banned number. The full list of what Meta checks is in our WhatsApp Business API guide.

1. Week 1 — connect and get templates approved. Move the store's number to the API, submit business verification, and get the six utility templates above approved. Keep the old phone as a backup for a few days in case anything is missed.
2. Week 2 — move the regulars. Tell your existing WhatsApp customers the number is unchanged and orders work the same way. Staff take lists from the shared inbox instead of a personal phone.
3. Week 3 — switch on updates. Order received, substitutions, out for delivery, delivered. These are the messages customers notice first, because they stop them having to call.
4. Week 4 — the first offer. Only to customers who opted in, on a fixed day, with specific items and prices. Watch replies and blocks, and adjust before the second week.

• The number is removed from the WhatsApp app and registered on the API.
• Business verification is submitted with a document in the store's legal name.
• The six utility templates are approved; none contains an offer.
• Every customer on the offers list opted in, and you know where each opt-in came from.
• Two people know how to use the shared inbox, and someone owns substitutions during picking hours.
• There is a default for unanswered substitutions, written down.