Most WhatsApp bills have obvious slack in them, and removing it does not require sending less. These are the seven changes we see save the most, in rough order of impact.
1. Move volume to customer-initiated conversations
Conversations a customer starts are treated far more favourably than ones you start. Every place you currently push people to a form or a phone number is a candidate for a WhatsApp entry point instead: website chat button, QR codes on packaging and receipts, click-to-WhatsApp in your advertising.
This is structural rather than cosmetic, and it usually saves more than the other six combined.
2. Batch messages inside the 24-hour window
If two events happen within a day — dispatched in the morning, out for delivery that afternoon — the second message costs nothing extra provided the conversation is still open. Audit your triggers for events that could reasonably be combined or sequenced within one window.
3. Keep utility messages genuinely utility
Adding "and here's 10% off your next order" to a dispatch notification moves the whole message to the Marketing rate. Across thousands of orders that single sentence can cost more than the promotion earns. Send it separately, to people who opted into marketing.
4. Clean the list
You pay to message people who will never respond. Contacts with no engagement across several campaigns are pure cost, and they suppress your quality rating as well. Removing them lowers the bill and improves deliverability at the same time — one of the rare changes with no trade-off.
5. Cut the messages nobody needed
Review your automated sequences and ask, per message, what a customer does differently because of it. "Your order is being prepared" usually fails that test. So does a second reminder nobody acts on. Removing two low-value messages from a sequence that runs on every order is a permanent saving.
6. Fix retry behaviour
Failed sends that retry aggressively can multiply cost for no delivery. Cap retries, back off between attempts, and treat repeated failures to the same number as a data problem to fix rather than a send to repeat. Also check that test sends are not going to live numbers.
7. Segment by country
Rates vary substantially by market. A campaign that is marginal in one country can be clearly unprofitable in another. Look at your spend by destination country — most businesses have never done this, and it frequently reveals that a small share of contacts consumes a disproportionate share of budget.
What not to cut
- Exception messages. A delay notification prevents a support contact that costs more in labour than the message did.
- Delivery updates. These reduce "where is my order" volume, which is real money.
- Anything preventing a complaint. The cheapest message you will ever send is the one that stops a dispute.
Measure per outcome, not per message
Cost per conversation is the wrong denominator. Cost per order, per resolved support case, or per retained customer is what tells you whether messaging is working. A programme with a higher message cost and a lower cost per order is the better one, and optimising the headline rate can quietly make you worse off.