How many people a messaging queue actually needs
Messaging staffing is a concurrency problem, not a volume one. The question is not how many messages arrive, it is how many conversations are open at once and how long each stays open - and messaging conversations stay open far longer than calls.
Why it is unlike phone and email
| Phone | Messaging | ||
|---|---|---|---|
| Concurrency | One | Effectively one at a time | Several |
| Customer's pace | Now | Days | Minutes, unpredictably |
| Conversation length | Minutes | One or two exchanges | Hours, with gaps |
| Cost of a gap | Call ends | None | Context lost, window closing |
The third row is the one that breaks staffing models borrowed from elsewhere. A messaging conversation is open for an hour and needs four minutes of attention spread unpredictably across it.
Working out the number
- Count new conversations per hour, by hour, across a normal week. Not the daily total - the shape.
- Measure your handling time: total minutes a person actually spends on a conversation, not its duration.
- Pick a concurrency you believe. Four to six is realistic for anything that requires looking something up. Eight is possible for simple, repetitive traffic.
- Divide, then add the tail. Staffing for the mean gives you a queue every busy hour, and busy hours are when it matters.
Do the arithmetic per hour, not per day. A business receiving 200 conversations a day with half of them between 17:00 and 20:00 does not need the daily average of staff at any point - it needs cover in the evening and almost none at 10:00.
The cheaper levers, first
Before hiring, the three things that reduce required headcount without reducing service:
- Answer the machine-answerable. Hours, location, delivery cost. Often a large share of first messages - see auto-replies that help.
- Narrow the hours and mean them. Stated hours with a real reply beat all-hours with an unpredictable one.
- Fix the top three repeated questions at the source. If forty people a week ask where their order is, the answer is an order notification, not another agent.
The floor nobody plans for
One person cannot cover a queue, because they take leave and get ill. Any channel you promise a response time on needs two people who can answer it, and that is a floor rather than a target.
Questions people actually ask
How many WhatsApp conversations can one person handle at once?
Somewhere between four and eight concurrently for real support work, far fewer than chat widgets are often staffed at, because customers reply on their own schedule and each return costs a context switch.
Is messaging cheaper to staff than phone?
Per conversation, usually yes, because one person handles several at once. Per hour of coverage it is often worse, because the expectation of a quick reply extends across more of the day.
Njiwa
One inbox for WhatsApp, Facebook and Instagram, and an API that sends over either WhatsApp rail. 7 days free, no card.