Automation business cases tend to be built on a single number — hours saved multiplied by an hourly rate — and that number is almost always both too low and misleadingly framed.

Here is how to cost a manual process honestly, including the parts that get left out on both sides.

Start with the real hourly cost

The most common error is using salary divided by contracted hours. That understates the true cost substantially.

A fully loaded hourly cost includes employer National Insurance, pension contributions, holiday and sickness cover, equipment, software licences and premises. And it divides by productive hours rather than contracted ones, because nobody delivers 37.5 productive hours a week.

For a UK administrator on £26,000, the fully loaded figure lands somewhere around £20 to £24 an hour, against the £13 that salary alone suggests. Use the real number.

Then add the three costs nobody counts

The cost of errors

Manual data entry has an error rate. Every rekeying step introduces one, and the cost of an error is not the thirty seconds to correct it — it is the time to notice it, investigate it, correct it downstream and occasionally apologise for it.

A wrong figure on an invoice costs a phone call, a credit note, a re-issue and some goodwill. Estimate conservatively and it still usually adds ten to twenty per cent to the true cost of a manual process.

The cost of delay

A manual step happens when somebody gets to it. That means a quote goes out on Thursday rather than Tuesday, an invoice is raised at month end rather than on completion, an enquiry sits for a day.

Delay costs money directly. Two days on average payment time is a real cash-flow number. Two days on quote response time is a measurable win-rate number, since in most small business markets the first credible quote has a substantial advantage.

The cost of it depending on one person

Usually the largest hidden cost and the one nobody puts in the spreadsheet.

If the process only happens when a particular person is available, then their holiday, their illness and eventually their resignation are all business disruptions. Ask what happened last time they were off for a fortnight. The answer is the number.

The strongest argument for most small business automations is not the hours. It is that a process currently stops when one person is on annual leave.

Now the costs on the other side

Be equally honest here, because business cases that omit these lose credibility the first time reality arrives.

Build cost. £800 to £2,500 for a straightforward flow between systems with proper connectors. More where systems lack integrations or data needs meaningful transformation.

Internal time. Specifying the process, answering questions, testing the result. Usually one to three days of somebody's attention, and it is real time that comes out of something else.

Run charges. Small — a few pounds a month per flow — but recurring. Premium connectors in Power Automate can change this materially, so check before designing around one.

Maintenance. The line most often omitted. Budget fifteen to twenty per cent of the build cost annually. Systems change, APIs are deprecated, business rules move. An automation that receives no maintenance is one that will silently stop within a few years.

A worked example

Supplier invoices arriving by email, entered by hand into the accounting system.

The manual cost:

  • 40 invoices a week, 4 minutes each = 2.7 hours a week
  • 140 hours a year at £22 = £3,080
  • Errors, roughly 2 per cent needing correction = £300
  • Delay causing missed early-payment discounts = £400
  • Two weeks a year when nobody does it and it backs up = £500 in catch-up and late payment

Annual cost: about £4,280.

The automation:

  • Build = £1,800
  • Internal time, two days = £350
  • Run charges = £120 a year
  • Maintenance at 18 per cent = £325 a year

Year one: £2,595. Ongoing: £445 a year.

Payback in about seven months, and roughly £3,800 a year afterwards. That is a good candidate.

The framing that matters

Two benefits sit outside the arithmetic and they are usually the ones that actually persuade people.

Capacity rather than cost. Unless you intend to reduce headcount — and you almost certainly do not — the saving is not cash. It is that the same team handles more work without hiring. Say that plainly rather than presenting a cost cut nobody will make.

Resilience. The process no longer depends on one person being at their desk. For a business of ten people that is worth a great deal and it never appears in a spreadsheet.

The threshold

As a working rule: build it if the annual manual cost exceeds the first-year automation cost. Think harder if payback runs past eighteen months. Do not build it if payback runs past two years — the process is either too infrequent or the build too complex, and both are signals to choose a different candidate. The situations where the answer is a firm no are set out in when not to automate.

If you would like a candidate process costed properly rather than estimated, our business automation team will do the arithmetic with you before quoting for the build. Get in touch.