Should you confirm cash-on-delivery orders on WhatsApp or by phone?

The real difference is who controls the timing

WhatsApp and a phone call can both carry a confirmation. Both can ask an open question, receive a change, and produce a record. On capability they are much closer than the usual framing suggests.

Where they differ completely is when the exchange happens. A call happens now: within a minute you know whether you have an answer. A message happens when the customer next opens the app, which might be in four minutes or after the courier has already collected.

For confirmation specifically, that is the deciding property, because a confirmation is only useful before dispatch. A channel whose latency you do not control is a channel that will sometimes deliver its answer after the answer stopped being actionable.

Where WhatsApp is clearly the better choice

When the thread already exists. If a customer messaged you to place the order, replying in that thread is not a new contact — it is continuing a conversation. Response rates on an existing thread are a different thing entirely from response rates on a cold template message, and this is the single strongest reason to use it.

When the answer needs a picture. Sizes, colours and variants are far easier to settle with an image than with a description read aloud. "Is this the one?" with a photograph beats any amount of careful phrasing.

When your customers do not answer unknown numbers. This is true of some segments and it is not a preference you can argue with. Meeting people where they already are beats being right about the better channel.

When you want the record to be readable by the customer too. A transcript is for you. A chat thread is for both of you, and a customer who can scroll back to what was agreed is a customer who complains less.

What Meta's pricing means for a confirmation flow

The pricing model is worth understanding before you build on it, and it has changed recently enough that older advice is wrong. The specifics, with the source and the date, are in the table above; the shape is what matters here.

You are billed per delivered template message, not per conversation, and only template messages are billed. Service conversations — where the customer messaged first — are free. So the cost of a WhatsApp confirmation depends almost entirely on who started the exchange. Replying inside a thread the customer opened is cheap. Reaching out cold to a customer who ordered through a website means a template message, which is billed.

That maps neatly onto the recommendation above: WhatsApp is strongest exactly where it is also cheapest, which is when the customer came to you there first.

Why we publish no Bangladesh rate

We give no per-message figure for Bangladesh, and it is worth explaining why rather than just omitting it.

Bangladesh has no standalone numeric rate in the text of Meta's published pricing documentation. It currently falls inside a regional grouping, and the actual numbers live in downloadable per-currency rate cards behind an interactive selector rather than in any page we could read. So any figure we printed would be second-hand at best.

One dated fact is worth knowing if you are budgeting: Meta's own pricing-updates page states that Bangladesh moves to a standalone rate card on 1 October 2026, with lower utility and authentication rates and a new, higher authentication-international rate. That means a figure you looked up in the middle of 2026 may be wrong by the end of it. Check the current card.

This is the same discipline we apply to every third-party number on this site: a source and a date or nothing. Getting somebody else's pricing wrong is not a small error.

Running both, which is what most merchants end up doing

The arrangement that works is not a choice between the two.

Confirm on WhatsApp where a thread already exists, because it is cheaper, less intrusive and often faster. Call the rest — the website orders, the customers who have never messaged you, and anyone whose WhatsApp message has gone unread as the dispatch window closes. That last group is the one where a call earns its cost, because it converts an unanswered message into an actual outcome.

Whichever channel carries it, the thing that decides whether the customer engages is the language it happens in. That is a bigger lever than the medium, and it is covered in Bangla, Banglish or English. The SMS comparison covers the cheaper, blunter version of the same trade-off.