Automation2 min readBy Ry
What your manual work is really costing you
A simple way to put a dollar figure on the copying, pasting and re-typing your team does every week, and how to decide what to automate first.
Every business has a few jobs that nobody chose. Someone exports orders every morning and pastes them into a sheet. Someone retypes invoices from PDFs. Someone builds the Monday report from five tabs. None of these feel expensive, because each one only takes “a bit of time.”
The problem is that “a bit of time” is multiplied by every working day, every person doing it, and every mistake it creates downstream. Put a number on it and the conversation changes.
The basic calculation
Start with the direct cost. You need four figures, and rough ones are fine:
- Hours per week the task takes, per person
- Number of people who do it
- Loaded hourly cost: salary plus taxes and benefits, usually 1.25 to 1.4 times the hourly wage
- Weeks worked per year (48 is a sensible default)
Multiply them together. Five hours a week, two people, $35 an hour loaded, 48 weeks: that is $16,800 a year on one task. Most businesses have several of these.
The costs the calculation misses
Direct hours are the smallest part. The larger costs are harder to see, which is why they rarely make it into a budget conversation.
Error cost
Manual transfer has an error rate. It is small per field, but multiplied across thousands of fields a month it is never zero. The real cost of each error is paid later: a mis-shipped order, a wrong invoice, a customer who does not come back. Count how many errors you found last month and what fixing them took.
Delay cost
If orders are entered once a day, every order waits up to a day. If the report is built on Monday, decisions are made on last week’s numbers. Delay rarely shows up as a line item, but it shows up in slower shipping and slower reactions.
Key-person risk
Manual processes live in people’s heads. When that person is on holiday, sick or leaves, the process degrades or stops. Ask what happens if they are out for two weeks.
Ceiling cost
A process that scales with headcount caps your growth. If doubling orders means doubling the people who retype them, margin shrinks as you grow.
Deciding what to automate first
Rank each manual task on three things:
- Annual cost: the calculation above, plus a rough error cost
- Rule clarity: could you write the steps down so a new hire could follow them without asking?
- Stability: has the process been the same for six months, or does it change every few weeks?
The best first candidates are expensive, rule-based and stable. A task that needs judgement every time is usually a tool or training problem, not an automation problem. A process that is still changing should be settled before it is automated, or you will pay to automate it twice.
What a sensible payback looks like
A good automation project pays for itself within a year on direct hours alone, with the error, delay and risk reductions as a bonus. If the maths only works over three years, look for a cheaper approach: an off-the-shelf app, a no-code workflow, or a change to the process itself.
If you would like a second pair of eyes on the numbers, a business audit or a free call is a good place to start.
Written by Ry, Evenbuilt · Los Angeles
All posts