The case against running your business on someone's personal WhatsApp

Running customer conversations on a personal WhatsApp works until exactly the moment it does not, and it always fails in the same five ways: one person is the bottleneck, nothing is measurable, history leaves with the phone, nobody knows what was promised, and there is no handover.

None of this is an argument that WhatsApp is the wrong channel. It is the right channel; that is the problem. It works so well at ten conversations a day that nobody notices the ceiling until they are through it.

Failure one: the bottleneck is a person

A phone can be held by one person. When that person is asleep, driving, or on leave, your customer messaging is offline. The business grows, the queue grows, and the only available fix is that one person working longer.

Failure two: nothing is measurable

How many conversations came in last week? How many are still unanswered right now? What is the median time to a first reply? On a phone, all three questions are unanswerable, so nobody asks them, so nothing improves. The first time most businesses see these numbers they are worse than anybody guessed.

Failure three: history walks out

The conversations are on a device belonging to someone who can resign. Three years of what was promised, quoted and agreed with your customers is on a phone you do not own. This is an operational risk and, in most jurisdictions, a data protection problem: you are the controller of that personal data and you cannot produce it, export it or delete it on request.

Failure four: nobody knows what was promised

A customer says they were quoted a price last month. On a shared inbox anyone can search and find it. On a personal phone the answer depends on whether one specific person is reachable and willing to scroll.

Failure five: no handover

The second person joins and there is no way to divide the work. The common improvisations - forwarding screenshots, a second number nobody publishes, WhatsApp Web open on two laptops signed in to one account - each create their own mess. The two laptops one is the worst: two people answer the same customer, neither knowing, and the customer gets contradictory answers within a minute.

What a shared inbox has to do to be worth it

Not much, but it has to do these properly:

1. Several people, one number. Without anybody logging anybody else out.
2. Attribution. Who sent this. Kept as a fact about the message, so it survives the person leaving.
3. Unanswered as a first class idea. Distinct from unread, and the default sort order.
4. Reply from the number it arrived on. Obvious, and the thing multi-number setups get wrong most often.
5. Search across everything. Including conversations that started before the person searching joined.
6. Export. Because the data is your customers', and one day one of them will ask.

The honest limits

A shared inbox does not fix a staffing problem. If nobody has time to answer, a better inbox produces a better measured backlog. What it does is make the backlog visible, divisible and survivable - and the first of those usually leads to the other two.

What Njiwa's Desk does is exactly this list, and nothing more clever than it.