“It's only an hour a day.”
That's the sentence. It comes up in every conversation we have about workflow, and it's always said as a reason not to do anything. An hour a day sounds like nothing. Multiply it out and it's a number you can't unknow.
Admin isn't the problem
Every job has admin in it. Nobody's role is a hundred per cent the thing they were hired for, and a business where nobody does any admin isn't a business. That's not what this is about.
It's about what happens when admin starts biting into the top of the role. When the paraplanner is formatting documents instead of building strategy. When the adviser is writing file notes instead of sitting in front of clients. When the ops manager is fixing broken data instead of running operations.
Admin at the bottom of a role is overhead — fine, priced in, leave it alone. Admin displacing the top of a role costs you twice: you're paying senior rates for junior work, and you're not getting the senior work. That's the one worth putting a number on.
A worked example
Say you run a small advice practice. Once a month, someone sits down with commission statements from six insurers and reconciles them line by line against what landed in the bank, then chases the ones that don't match. It's a solid two days a month, which averages out to about an hour a day.
That person is on $95,000. A standard day is 7.6 hours. An hour a day is one-seven-point-sixth of what you pay them:
$95,000 ÷ 7.6 = $12,500 a year.
One person. One task. Before you count the commission that was short-paid and never chased, because nobody had time to get to it.
If you want the hourly figure: $95,000 across a working year of 219 days × 7.6 hours is about $57 an hour — higher than their hourly pay rate, because you pay them through annual leave, public holidays and sick days too. The hours you get have to carry the cost of the ones you don't.
Change the inputs and it holds anywhere:
- A bookkeeper on $75,000, 45 minutes a day on the same kind of matching — $7,400 a year
- A legal assistant on $85,000, an hour a day formatting and filing — $11,200
- An ops manager on $130,000, two hours a day pulling reports together — $34,200
Same arithmetic every time. Salary, times the share of their day.
{ The cost of time }
This task costs
$12,500 a year
on wages alone. The real number is higher.
That’s the conservative number. Wages only. It doesn’t count the 25–35% on top that comes with employing someone (super, insurance, the software they log into). It doesn’t count what they’d be doing instead, or what the mistakes cost.
This is a teaching example. The real builds we've done have removed multiples of this.
Now stack it
That's one person on one task. Three people each losing an hour a day to their own version of it is $37,500 — just under 40% of a full-time role, spent on work nobody was hired to do.
Then add the two hours a week your ops manager spends fixing CRM data, and the ninety minutes a day your senior adviser burns on file notes that could be transcribed automatically. It stops being admin and starts being a salary you're paying for nothing.
Most Australian small businesses we speak to have three to six of these running quietly in the background. Nobody's ever added them up.
Be honest about what the number is
It's capacity, not cash. Removing 219 hours of admin doesn't put $12,500 back in the account unless somebody leaves. What it puts back is 219 hours of a person you hired to do something better than data entry.
The part that is cash is the error cost. The commission keyed in wrong that surfaces six months later, and the two days someone spends unwinding it. That happens because it's the same task, done by hand, every month for three years, not because anyone's careless.
And it compounds
A fix is a one-time cost. The saving is annual.
That $12,500 doesn't come back once. It comes back next year, and the year after. Ten years of the same task is $125,000 — and that's in today's dollars, so it's an understatement, because salaries go up and the task doesn't get faster.
This is the bit most owners miss. They weigh a fix against this month and can't see the case. Weigh it against the next ten and it isn't close.
Two options
Once you know the number there are two moves.
Live with it. Keep paying the bill. Sometimes that's the right call.
Improve it. Where to start depends on the size of the number:
- Under 5 hrs/wk (around $12,500) — live with it, or fix the workflow. A cleaner handoff, a tighter template, one fewer step.
- 5 to 10 hrs/wk — look for an off-the-shelf tool, or an integration between two systems you already pay for.
- Over 10 hrs/wk ($25,000 and climbing) — a custom build starts to pay for itself in year one.
And sometimes the right move is killing the task altogether, because it shouldn't exist.
Where we sit on all this
We're not going to tell you to drop everything and go all-in on AI. Plenty of tasks don't need a system built for them, and we'll say so.
Work out your number first. Then decide whether it's worth doing anything about it.
You've got a number. Send it to us with a one-line description of the task and we'll tell you whether it's worth building something — free, no pitch.