Why WhatsApp won, and why that is fine
Nobody chose WhatsApp as their operations platform. It arrived because it was already on everybody's phone, and because the alternative - asking a technician with dirty hands to open an app and update a card - never survived contact with a busy Tuesday.
That is not a failure of discipline. It is a completely rational outcome: the tool with zero adoption cost beats the better tool with any adoption cost, every time, in a business where nobody is paid to administer software. Any advice that starts with "stop using WhatsApp" is advice that will be ignored, and rightly.
So the useful question is not how to move your team off it. It is how to use it deliberately, and to know precisely which jobs it cannot do so you stop expecting it to.
- It is already installed, so adoption is free
- The least technical person on your team is fully competent in it
- It works on bad signal, old phones and in any country
- It carries tone, which a task board cannot, and that does real work in a small team
The three things a group chat cannot do
A group chat is a conversation, and a conversation is not a system of record. Three specific capabilities are missing, and each one costs money in a different way.
It cannot assign responsibility
"Can someone ring Mrs Diaz back about the warranty?" is read by four people, each of whom reasonably assumes somebody closer to it will pick it up. Nobody does. On Friday, Mrs Diaz rings the owner.
This gets worse as the group grows, not better. A request to everybody is a request to nobody, and reading a message is not the same as accepting it - with no way to tell the two apart.
It cannot tell you what is outstanding
There is no question you can ask a group chat that returns the things that are not done. You can scroll, which finds only what you already remember to look for. So "what is outstanding?" becomes walking round asking four people: forty minutes, four partial answers from memory, and it has to be repeated tomorrow.
It has no memory you can query
"What is my approval limit?" is asked in the group and answered from memory by whoever is free. Next month it is asked again and answered slightly differently. Two members of staff are now confidently operating to different rules, and nobody will find out until one of them is wrong in front of a customer.
Fix one: name a person in every request
This is free, takes no software, and removes the largest single category of lost work. Never post a request to the group without a name attached, and treat a request without a name as not yet made.
It feels blunt for about a week, and then it feels normal. The awkwardness is the point: it is the moment somebody either accepts the job or says they cannot, and that moment is exactly what a group request lets everyone skip.
- "Can someone ring Mrs Diaz?" becomes "Ravi, can you ring Mrs Diaz today?"
- If you do not know who should do it, that is the actual problem - decide first
- Ask for a one-word acknowledgement, so reading and accepting are visibly different
- If nobody can take it, you have learned something more useful than silence
If you only do one thing from this guide, do this one. It costs nothing and it is the difference between a request and a wish.
Fix two: move work into one-to-one chats
Keep the group for what it is good at - coverage, swaps, notices, the social glue. Move anything that needs to be traced to a person into a direct chat with that person.
The reason is not tidiness. In a one-to-one chat, responsibility is unambiguous by construction: there is one other person, so there is nobody for either of you to assume about. It also means you can see, at a glance, what you have actually asked each person for this week - which you cannot do in a group.
- Direct chats for anything with a deliverable or a deadline
- The group for notices, coverage and everything social
- It also makes a reminder private, which matters more than owners expect
- Nobody has to be told about somebody else's late work
Fix three: get one message at the end of the day
A daily end-of-day message from each person is the highest-value habit in a small team, and the hardest to establish. Two or three lines: what got done, what is stuck, what is carrying over.
It usually fails for a reason that has nothing to do with willingness. People stop sending updates when the update disappears into nothing - when nobody visibly reads it, nothing changes as a result, and it therefore looks like paperwork for its own sake. If you want the habit, you have to make it visibly matter.
- Ask for two or three lines, not a form or a template
- Reply to the first few, so it is obvious somebody reads them
- Act on a blocker within a day, once - that is what makes the habit stick
- Never use an update against somebody in the same week you asked them to start sending it
There is a longer treatment of exactly this in getting staff to report, including what to do when updates stop arriving.
Fix four: write down the five things people keep asking
Spend twenty minutes listing the questions your staff have asked you more than twice this month. Approval limits, what to do when a part is late, who signs off overtime, what to tell a customer who wants a discount, how long a process should take.
Write the answers down - one page, plain sentences, actual numbers. Then put it somewhere they can reach from their phone. This single exercise removes more interruption than any tool, and it is the prerequisite for any software ever answering a question for you.
- List the questions asked more than twice this month
- Write one plain answer each, with the real figures in it
- Put it where a phone can reach it - a shared note is enough to start
- Tell people it exists, twice, because once is never enough
- Add to it whenever a new question is asked twice
How to tell whether the free fixes are enough
For a lot of teams, the four fixes above are enough, and anyone selling you software should be willing to say so. The way to find out is to measure rather than to feel, for one week.
| Measure this for a week | Probably fine | It is costing you |
|---|---|---|
| Hours you spent chasing | Under 2 | Over 5 |
| Times a customer told you first | 0 | 2 or more |
| Things outstanding you can list unaided | Most of them | Very few |
| Same policy question asked twice | Rarely | Weekly |
| People who sent no update at all | 0 - 1 | 3 or more |
If your week lands in the left column, keep your money and keep doing the four fixes. If it lands in the right, the problem is no longer discipline - it is that nobody can hold the whole picture in their head, and that is a tooling problem.
What software should add, and what it must not remove
If you do reach for software, the test is narrow. It must not require your staff to open anything new, because that is the step that killed every previous attempt. And it must answer "what is outstanding?" without anybody being asked.
This is the point where we will say plainly that it is what we built. Bosun works inside WhatsApp, in one-to-one chats, and turns those messages into a list the owner can query. Your staff carry on exactly as this guide describes; the difference is that somebody can now ask what is late and get a true answer.
- It must not need staff to install, log into or open anything - see the assistant
- It must answer what is outstanding, without asking people - see the owner briefing
- It must chase without nagging, and stop - see automatic follow-up
- It should answer the repeated questions from your own written page - see answers from your procedures
If a tool fails the first test, the rest of its feature list does not matter. That is the single most reliable predictor of whether it will still be in use in three months.