What actually drives your WhatsApp bill
Meta stopped charging per conversation in July 2025. It bills per delivered template now, and the category you pick moves the number by an order of magnitude.
If you read a guide to WhatsApp pricing written before mid-2025, throw it away. Meta used to bill per 24-hour conversation. Since 1 July 2025 it bills per delivered template message, and that changes which optimisations are worth anything.
The four categories, and what they cost relative to each other
Absolute rates change and differ by country — India moved to local-currency billing in January 2026 — so check Meta's current rate card rather than any vendor's blog. What is stable is the shape:
- Marketing — the expensive one
- Promotions, offers, re-engagement, "we miss you". Every serious conversation about reducing a WhatsApp bill turns out to be a conversation about this category.
- Utility — roughly an order of magnitude cheaper
- Follows something the customer did: an order, a booking, a payment, a delivery. The trigger is what makes it utility, not the tone.
- Authentication — priced like utility
- One-time passcodes, nothing else.
- Service — free
- Your replies inside the 24-hour window after a customer messages you. No charge at all.
Three things follow from that
1. Category selection is a pricing decision
A message that follows an order is utility. Filed as marketing, it costs several times more for no benefit. Filed the other way — marketing dressed as utility — it gets rejected, which is the failure mode of trying to be clever. The rejection reasons, in order.
Audit your existing templates once. Most businesses find at least one high-volume template sitting in the wrong category, and fixing it is a config change with no downside.
2. Anything that makes them message first is worth real money
Customer-initiated conversations open a free 24-hour service window. So the click-to-WhatsApp ad, the QR code on the receipt, the wa.me link in the email signature and the "message us" button on the site are not just convenience features — they move traffic from the paid column to the free one.
This is the single highest-leverage change most teams have available, and it is usually owned by whoever runs the website rather than whoever watches the bill.
3. Per-message billing punishes chattiness
Under the old conversation model, a template that opened a window and then sent three follow-ups cost the same as one. It does not any more. Sequences that were free to pad are now billed per step.
Look at any automation that fires several templates in a row and ask whether each step earns its place. Usually two of five do.
The other meter
Your platform subscription. Every provider resells the same API from Meta, so what you are choosing between is the software on top and the commercial terms — what you are actually buying.
Two questions worth asking in writing:
- Are Meta's charges passed through at cost, or marked up? Both are legitimate; only one of them is usually disclosed.
- Is there a platform fee per message or per conversation on top? This is where a cheap-looking subscription becomes expensive at volume.
Convarza does not charge per message or per conversation. The plan covers the software and the AI usage; Meta's charges are Meta's. Plans.
A ten-minute cost review
- List every template you send, with monthly volume and category.
- Flag any high-volume template in marketing that follows a customer action — those are misfiled utility.
- Find every automation that sends more than two templates in sequence and cut the weakest.
- Count how many of your inbound conversations are customer-initiated. If it is low, the fix is on your website and your ads, not in your messaging tool.
Most teams find a double-digit percentage in that list without sending a single message fewer that anybody wanted.