Small Business Admin Time Audit: A 5-Day Method

A small business admin time audit takes five days, a notebook and three columns. It is the cheapest diagnostic available to an owner and almost nobody runs it.

Every owner knows admin eats their week. Almost none of them know which admin, and the difference between those two states is the difference between a vague grievance and something you can act on.

The exercise below takes five days and a notebook. It is not sophisticated. Its entire value is that it replaces a feeling with a list, and the list is almost never what people expect.

The audit

For five working days, every time you switch to a task that is not the actual work of the business, write down three things: what it was, when you started, and roughly how long it took.

That is the whole method. Three columns.

Two rules make it work. Write it down when it happens, not at the end of the day, because reconstructing a day is where the distortion enters. And count the small ones — the ninety seconds spent finding an invoice number, the two minutes spent explaining to somebody where a file is. Those are the ones the memory throws away and they are frequently the bulk of the total.

At the end of the week, group them. Not by feeling — by what actually triggered them.

What people expect to find

Invoicing. Almost everyone predicts invoicing.

It is a real cost, it is genuinely irritating, and it is memorable because it happens at a specific time and it feels like a chore. That memorability is exactly why it gets overestimated.

What the audit usually turns up instead

Answering the same question again. Not a hard question — the same one. Where are you based. Do you do that kind of work. How much roughly. When could you come. Are you open Saturday. Individually thirty seconds. In aggregate, across a week, it is very often the single largest line on the page, and almost nobody predicts it because no single instance felt like admin at the time.

Chasing. Chasing a quote that went out and went quiet. Chasing a supplier. Chasing a client for the information you need to start. Chasing is expensive in a way that does not show up as time, because it also occupies the part of your head that was supposed to be doing something else.

Re-entering information you already have. Something arrives on WhatsApp. It gets typed into a quote. The quote gets typed into an invoice. The invoice gets typed into the accounting package. The same address, four times, by a person.

Finding things. Which version of the quote did we send. Where is the photo they sent in June. What did we charge the last customer in that street.

Being interrupted. The audit will not capture this directly, but you will see it in the shape of the log — a run of two-minute entries scattered through an afternoon that was meant to be one job. The two minutes is not the cost. The cost is that the afternoon produced nothing.

Why so much of it is message-shaped

There is a reason the top of that list keeps coming back to answering people.

DataReportal's Digital 2026 report puts WhatsApp reach at roughly 96% of South African internet users. That is a higher share than almost any market in the world, and it has a consequence most owners have absorbed without ever naming: your business is reachable, personally, on a device in your pocket, continuously.

An email inbox has a lid on it. You open it, you deal with it, you close it. WhatsApp does not work that way — it is the same application your family uses, it notifies, and a customer messaging at 20:15 has no sense of having done anything unusual.

That is not an argument against the channel. It is where your customers are and it is why response times here can be measured in seconds rather than days. It is an argument for noticing that the admin load it creates is real, is unbounded by office hours, and does not appear anywhere in a time audit unless you deliberately log it.

Which is exactly why the third rule of this exercise is to count the small ones.

Why the boring answers are the valuable ones

There is a natural instinct to want the answer to be something impressive. Automation marketing feeds it — the case studies are always about something sophisticated.

But the categories above are boring on purpose, and boring is what makes them tractable. A repetitive question with a stable answer is the easiest thing in the world to hand off. A follow-up on a fixed schedule is a rule. Re-entering the same address four times is an integration problem with a known shape.

The hard problems in your business — pricing a difficult job, deciding whether to take on a client, judging whether a supplier is about to let you down — should stay with you, and no honest supplier will offer to take them.

The SARS and compliance corner

One category the audit reliably under-counts, because it does not happen weekly.

VAT returns, provisional tax, the annual return, UIF, the POPIA obligations nobody has got to yet. These are lumpy — nothing for six weeks, then a day gone — which means a five-day audit will usually miss them entirely.

Add them by hand at the end. Estimate the hours across a year and divide by fifty-two. The number is usually larger than owners expect, and it belongs on the same page as the daily drains because the decision about what to hand off should be made across the whole picture rather than the part that happened to fall inside your audit week.

Deciding what to do about it

Once you have the list, sort it by frequency times duration, not by irritation. The most annoying task is rarely the most expensive one.

Then, for each line at the top, ask one question: does this require judgment, or does it require somebody to be available?

Things that require judgment stay with a person. Things that only require availability are the candidates — and "availability" covers a lot more than people expect. Answering the same question is availability. Sending a follow-up on day three is availability. Copying an address from a message into a quote is availability.

That single distinction will sort most of your list in about ten minutes, and it is a better filter than any vendor's questionnaire because it is about your work rather than their product.

What it is worth

Multiply the hours by something. Not your billing rate — you were not going to bill those hours, and pretending otherwise inflates the number in a way that will not survive scrutiny.

Use whatever the cheapest person who could do that task properly would cost, or use your own honest opinion of what an hour of your attention is worth on a Tuesday. Either produces a defensible figure. The point is not precision, it is whether the number is large enough to act on. Your figure, not an industry average.

Where it usually lands

Two things reliably come out of this exercise.

The first is that the largest single line is almost always repetitive enquiries, which is the thing this whole company is built around and which is why we keep pointing at it. That is a bias worth declaring rather than hiding.

The second is that the second largest line is usually document work — quotes, invoices, job cards, the same information moved between systems by hand. That is a different problem with a different shape, and it is what document automation and operations work exist for.

Both of those pages have prices on them, so you can compare the cost of fixing a thing against the cost of continuing to do it, which is the only comparison that matters.

Do the audit first

The reason to run the five days before speaking to anybody is that it changes who is in charge of the conversation.

An owner who walks in saying "admin is killing me" gets sold a solution. An owner who walks in with a list — eleven hours last week, four of them answering the same six questions, three of them chasing quotes — is running the meeting, and can tell immediately whether what they are being offered addresses the top two lines or something further down that happens to be what the supplier sells.

There is a version of this in the Playground that compresses a week into a couple of minutes, if you want the feeling before you commit to the notebook. But the notebook is better, because the numbers are real.

One notebook, five days, three columns. It costs nothing and you do not need us to run it.