What Is That Task Really Costing You?
Copying data between systems. Re-keying invoices. Chasing approvals by email. Individually each takes minutes, which is exactly why nobody costs them. Annualised and converted into full-time-equivalent staff, the same tasks routinely turn out to be someone’s entire job.
Pick one specific task, not a whole department. The full-time-equivalent figure is usually the number that changes the conversation internally.
Manual Task Cost Calculator
One specific task. Time it, don’t estimate it.
From stopping the previous work to resuming it.
Across everyone who performs it, not just one person.
Roughly 1.3x base pay for an Australian employee.
Direct labour only. Context switching and errors are extra.
Direct labour cost only. Context switching, downstream error correction and the opportunity cost of the person’s time are all excluded, so the true cost of a manual task is reliably higher than this figure.
Three Ways to Read the Same Number
The same task can be expressed as dollars, as working days, or as a share of a person. Each lands with a different audience, and knowing which to lead with is half the battle in getting anything approved.
Dollars persuade owners and finance
An annual dollar figure sits naturally alongside every other line in a budget and makes the comparison against an automation cost immediate. It is the right lead for anyone who signs off spending, and the easiest to challenge, so make sure your hourly input is defensible.
Working days persuade operations
Telling a manager a task consumes thirty-one working days a year is far more visceral than a dollar figure. It maps onto the thing they actually manage, capacity, and it makes the opportunity cost obvious without needing anyone to agree on a rate.
Full-time-equivalent persuades everybody
Expressing a task as a fraction of a person is the number that ends debates. When a routine reconciliation turns out to consume a third of a full-time role, the question stops being whether to address it and becomes how quickly.
How the Numbers Are Built
Deliberately simple, so you can defend every figure in a meeting without needing the calculator open in front of you.
Time per occurrence, honestly measured
How long one instance actually takes including the setup and the switching, not the optimistic version people quote from memory. Timing three real instances beats estimating.
Multiplied by real frequency
Occurrences per week across everyone who performs the task. Tasks done by several people in parallel are the ones most consistently undercounted.
Annualised and costed
Weekly hours across fifty-two weeks at the loaded hourly cost of the people doing it, giving an annual dollar figure.
Converted to days and FTE
Annual hours divided by a 7.6-hour working day gives days lost, and divided by 1,750 productive annual hours gives full-time-equivalent staff.
Why Manual Tasks Cost More Than the Clock Says
The raw time figure is the floor, not the ceiling. Four effects consistently push the true cost of manual work above what a stopwatch measures.
Context switching taxes the surrounding work
A five-minute task that interrupts focused work costs considerably more than five minutes, because the return to the original task is slow and error-prone. Tasks that arrive unpredictably throughout the day (approvals, lookups, data requests) carry this tax most heavily, and it never appears in any time log.
- Interruption cost lands on the work either side, not on the task itself
- Unpredictable arrival is worse than a scheduled batch of the same volume
- Batching similar tasks reduces the tax even without automating them
- This effect is entirely absent from the calculator, which understates the total
Re-keying is where errors are manufactured
Any task involving copying data from one system into another has an error rate, typically around one per cent per field even for careful people. Those errors are discovered later, usually by someone else, usually at a worse moment, and the cost of correction is several times the cost of the original entry.
- Error rates rise sharply with volume, fatigue and time pressure
- The cost of an error is discovered downstream, not where it was made
- Reconciliation work often exists only to catch upstream re-keying errors
- Model this properly with the automation ROI calculator
The work usually falls on your most capable person
Manual processes tend to accumulate around whoever is most reliable, because that is who others trust with them. The result is that expensive, experienced staff spend hours on work that carries no judgement, while the things only they can do wait. The dollar cost understates this because it uses an average rate.
- Check who actually performs the task before choosing an hourly rate
- Senior staff doing routine admin is a capacity problem, not just a cost
- Key-person concentration also creates a genuine continuity risk
- Consider what that person would otherwise be doing with the hours
Nobody can see what the manual process produced
Manual work leaves no structured record. There is no log of how many were processed, how long each took, or where the delays occurred. That absence is itself a cost, because it prevents the process from ever being improved and makes it impossible to answer basic questions about capacity.
- Manual processes cannot be measured, so they cannot be improved
- Backlogs become visible only when someone complains
- Automation produces the measurement as a by-product
- Even logging a manual process without automating it usually pays for itself
Next Steps
Automation ROI Calculator
You know what it costs. Now model what automating it returns, including payback.
Model the return →Process Automation Priority Scorecard
Several expensive tasks? Score them to find out which one to tackle first.
Prioritise properly →AI Data Entry Automation
How re-keying between systems is typically eliminated in practice.
See how it works →Frequently Asked Questions
How do I measure how long a task really takes?
Time three real instances rather than asking anyone to estimate. Self-reported estimates for routine tasks are unreliable in both directions: people underestimate frequent short tasks because each one feels trivial, and overestimate unpleasant ones because they feel longer than they are. Include everything from the moment the person stops what they were doing to the moment they resume it, since opening the system, finding the record and saving the result are all part of the true cost even though nobody counts them.
Should I include time spent by multiple people?
Yes, and this is the most common source of undercounting. A three-step approval that touches an administrator, a manager and a finance officer costs the sum of all three, not just the longest one. Similarly, if four branch staff each do the same weekly reconciliation, the frequency input should reflect all four, not one. When a process spans people, map who touches it before you enter any numbers, because the total is usually several times what the primary owner would report.
What counts as a productive hour when calculating FTE?
The calculator uses 1,750 productive hours per year, which reflects an Australian full-time role of 38 hours a week after annual leave, public holidays, sick leave and a realistic allowance for the fact that nobody is productively occupied every minute of every rostered hour. Some organisations use 1,800 or 1,900. The precise figure matters less than using it consistently, and 1,750 is deliberately conservative, meaning the FTE result understates the impact slightly rather than inflating it.
The number looks too big. Have I done something wrong?
Usually not: this is the normal reaction, and it is why the exercise is worth doing. A task taking twelve minutes, done thirty times a week, by someone costing fifty-five dollars an hour, is over sixteen thousand dollars a year, and that is a genuinely modest example. If the result still seems implausible, check two inputs first: whether the frequency is per week rather than per month, and whether you have accidentally summed across people twice. If both are right, the number is right.
Does a high cost automatically mean I should automate it?
No. Cost tells you the prize, not whether it is winnable. A costly task that changes shape every time, depends on judgement, or runs on a system with no integration path may be a poor automation candidate despite a large annual figure. Sometimes the right answer is to redesign the process so the work disappears, to batch it, or to stop doing it at all. The priority scorecard is designed to test exactly this before you commit to building anything.
Can I use this for tasks done by contractors or outsourced providers?
Yes, and it is often more accurate because you already know the real rate. Use the actual charge-out rate you pay rather than an internal cost estimate. Bear in mind that outsourced processes frequently hide additional internal cost: the time your own staff spend briefing, checking and correcting the outsourced work. Include that as separate time at your internal rate, since it commonly adds twenty to thirty per cent to the true cost of an outsourced process.
Found an Expensive One?
Send us the task and how it works today. We will tell you honestly whether it is worth automating, and roughly what that would take.