How Many Business Processes Does a Ten-Person Company Actually Have?

There is no single right answer, and that is the useful part. The number depends entirely on where you decide one process ends and the next one starts. Ask ten people inside the same ten-person company and you will get ten answers with a tenfold spread between the lowest and the highest, and none of them is lying.

Start with something ordinary. "We invoice the client." One process?

It is, if you are the person who does it and it takes four minutes. Now watch what happens the fortnight that person is away. Somebody has to work out which clients are due, find what was agreed, check the work was actually delivered, pull the numbers out of the timesheet or the project tool, raise the invoice, send it, record that it went, chase at fourteen days, chase again at thirty, and decide at what point chasing becomes a phone call from you. That is not one process. It is nine or ten, and three of them turn on a judgement nobody has ever written down.

Pick a counting rule before you count

Several rules are defensible. They give wildly different numbers.

Count by department and a ten-person company has about six. Count by system, one for each tool that holds data somebody else relies on, and you land near twelve. Count by trigger, meaning every event that makes a person start doing something, and you are suddenly at sixty or seventy.

The rule that survives contact with a real company sits between those. A process is a trigger, a sequence of steps, and an outcome that somebody other than the doer can check. If nothing leaves the desk, it is a step. If something lands in a colleague's inbox, a customer's inbox, or a system another person reads, it is a process.

By that rule most companies under twenty people have somewhere between forty and eighty. There is no study behind that range. It is what the arithmetic gives you when you walk the four spines below and count without flattering yourself.

Four spines, and the list writes itself

The money first. Something creates the possibility of revenue, and something eventually puts cash in the bank. Enquiry, qualification, quote, agreement, delivery, invoice, collection, and the separate thing that happens when collection fails. Then the mirror image for money going out.

Then the customer. First contact through to renewal or departure. Onboarding is the one everybody forgets is a process at all, because it usually survives as a habit inside whoever does it most often.

Then the employee. Hiring, first day, getting access to things, being paid, being reviewed, leaving, and having that access taken away again. The last of those is the one small companies almost never have written down anywhere.

And the calendar. Everything that happens because of a date rather than because somebody asked. Month end, VAT, payroll, renewals, backups, the weekly meeting that is supposed to produce a decision.

Nearly everything a ten-person company does hangs off one of those four. Whatever does not is worth a second look, because it is either genuinely unusual or it is one person's habit that nobody has questioned in two years.

The total is not the finding

Write your list. Then put one name beside each line: the person who would have to do it if the usual person were unreachable for a fortnight.

In most companies this size, two names will account for more than half the list, and one of them belongs to a founder. That distribution is the finding. The total was only ever the thing that made you sit down and write the list out.

Mark one more column while you are there. Which of these has been done, start to finish, by exactly one human being ever? Not undocumented. Never performed by anyone else. That is a different category of risk, and it is the one that stops the week when somebody gets flu.

The case against doing this at all

Two fair arguments, and you should hear them before you spend a Saturday on an inventory.

If you already know precisely where the business hurts, go and fix that. A company with one screaming bottleneck does not need a catalogue of sixty processes, it needs a fortnight of attention on the thing everybody already complains about in the kitchen. Counting is a very comfortable way of not fixing.

And a list you build once, then never reopen, is worse than no list. Six months on it is wrong in ways you cannot see from the outside, and somebody will make a staffing decision on it anyway. If you are not going to keep it current, keep it short. Ten lines you trust beat eighty you do not.

What the list is actually for

Sort it by what it costs you if it stops this week. Not by how irritating it is. The irritating ones are loud and usually cheap. The expensive ones are quiet, which is precisely why they are still sitting there.

Take the top three and write them down properly, meaning somebody who has never done it could follow the page and get the same result. Three processes is roughly a day's work. It is usually the highest-return day in the quarter.

If you would rather not spend the Saturday, abi. Clone builds the map from interviews with the people doing the work and hands you the same list with owners and hours already attached.