The first Black Friday where every reply costs money

Black Friday falls on 27 November 2026 and Cyber Monday on 30 November. WhatsApp service messages become billable on 1 October. That is eight weeks between the change and the busiest support week of the year, and it is the first peak season where answering a customer on WhatsApp has a per-message price.

What peak actually does to a support queue

The volume is the obvious part. A successful promotion can multiply order volume five to ten times in a few days, and support follows it with a lag of about a week — which is the part people plan for badly, because the queue peaks after the sale is over and the temporary staff have gone.

The mix changes too, and it changes in the direction that costs the most. Where-is-my-order runs at 25–40% of ecommerce support normally and climbs past 50% in peak, because carriers slip and every slipped parcel produces a thread. Those threads are multi-turn, they restart when the parcel still has not arrived, and from October each turn is billable.

The arithmetic that is new this year

Take a store that handles 2,000 support conversations in a normal month at three replies each. That is 6,000 billable messages where last year there were none.

Now apply peak: volume up five times, WISMO share up from 35% to 50%, and threads that run longer because the answer is "it is late" rather than "it arrived". The message count does not scale linearly with orders — it scales faster, because the hard conversations are the ones that take the most turns.

None of which is a reason to panic. It is a reason to know your own multiplier before the rates are published, which Meta is expected to do around 1 September. Two numbers, from twenty recent threads: conversations per month, and replies per conversation. Everything else is arithmetic.

Four things worth doing before October

1. Count replies per resolution now, not in December

It has become a cost line. If your average is five and it could be three, that is 40% off the new bill and usually a better experience, because fewer round trips is what the customer wanted. Measure it in August when the queue is calm enough to look at.

2. Move what you can from reactive to proactive

A dispatch notification is a utility template: billable, but volume-tiered, so the rate falls as you send more. A WISMO thread is service messages: flat-rate, however many you send. One proactive update can prevent an entire thread.

The trade depends on rates nobody has yet, so do not commit to a direction in August. Do get the templates written and approved, because approval takes time you will not have in November.

3. Fix the answers, not the staffing

Teams that hold service quality through December audited in Q1 and had things tested by September; starting in October leaves time for band-aids only. The highest-leverage work now is unglamorous: make sure the agent can find an order, knows your actual cut-off dates, and knows what your returns window becomes over Christmas.

4. Decide where automation stops before you are busy

Peak is exactly when the temptation to automate the hard conversations arrives, because the queue is frightening. It is also when getting it wrong is most expensive, because everyone is watching and nobody has patience. Decide the boundary in September, when you can think.

What does not change

The rules are the same in November as in August, and peak volume is not a defence. Opt-in is still required before you message anyone. Opt-out still has to be honoured immediately, including for the promotional sends that peak tempts everyone into. Automation inside the 24-hour window is still conditional on a clear, prompt escalation path to a person.

That last one deserves attention in peak specifically, because it is when escalation paths quietly stop working — the queue is long, the human is not answering, and the handover becomes a message into a void. An escalation path that nobody staffs is not an escalation path.

The uncomfortable one

If your plan for peak is to answer less, the new pricing rewards you for it, and that is the trap. Meta charging per reply creates a real incentive to be terse, close threads early, and not follow up. Every one of those saves money and costs a customer.

The version worth aiming at is fewer messages because each one answered more, not fewer messages because you stopped talking.

Common questions

Published 11 August 2026. The October pricing change is confirmed; the per-market rates are not published yet. This post will be updated when they are.