What counts as a 'fully-loaded' cost when calculating staff time?

Fully-loaded cost includes gross salary plus employer National Insurance contributions (currently 13.8% above the secondary threshold) and minimum auto-enrolment pension contributions. For a more complete picture, some businesses also include a proportion of overhead (office space, equipment, HR support). The key is to use a consistent method across all roles in your calculation.

Should I include the cost of my internal team's time during the build project?

Yes, always. Scoping sessions, feedback rounds, user acceptance testing, and training time all have a real cost even if no invoice is raised. A realistic estimate is 15–25% of the external build cost for a typical SME internal tool project. Leaving this out makes your cost column artificially low and your ROI artificially high.

How do I estimate hours saved if the process is not formally tracked?

Ask the people doing the work to log their time for two weeks, or walk through the process together and time each step. Two weeks of actual data is far more credible than a manager's estimate. If logging is not practical, interview two or three people who do the task independently and average their estimates — divergence in their answers is itself useful information about how well understood the process is.

What is a reasonable maintenance cost to budget for a custom internal tool?

A common benchmark is 15–20% of the initial build cost per year for ongoing maintenance, minor iterations, and dependency updates. This varies with complexity: a simple CRUD tool needs less; a tool with multiple integrations or frequent regulatory changes needs more. Factor this into your three-year cost model from the start rather than treating the build as a one-off cost.

At what point should I bring in an external development partner rather than building this internally?

If your internal development team either does not exist or has a backlog that means the tool will not be built for six months or more, the opportunity cost of delay is itself a calculable figure — add it to your ROI case. An external partner makes sense when speed to value matters, when the internal team lacks the specific skills, or when you want to keep your team focused on product rather than internal tooling.

How to Calculate the ROI of a Custom Internal Tool

To calculate the ROI of a custom internal tool, you subtract the total build and running cost from the measurable value it generates (time saved, errors eliminated, headcount avoided), then divide by the total cost and express the result as a percentage. In practice, most UK SMEs find the harder part is not the maths — it is identifying and honestly quantifying every cost on both sides of the ledger before a line of code is written.

Why Internal Tool ROI Calculations So Often Fail

Most internal business cases for custom software fail to get sign-off not because the tool is a bad idea, but because the proposal only captures one side of the equation. Sponsors list the build cost and then gesture vaguely at 'efficiency gains'. Sceptical finance directors reject it. A credible ROI calculation requires you to be as specific about the value as you are about the cost.

Tip

The goal is not to produce an optimistic number. It is to produce a defensible number — one you can stand behind in a budget meeting and that will not embarrass you six months post-launch.

Step 1: Map Every Manual Process the Tool Would Replace

Before touching a spreadsheet, spend time with the people actually doing the work. Walk through the process end to end and record, for each step: who does it, how long it takes, how often it runs, and what happens when it goes wrong. This is your baseline. Without it, every number you put into a business case is a guess.

Name the process (e.g. 'weekly client report compilation', 'supplier onboarding', 'stock reconciliation')
Record the roles involved and their approximate fully-loaded hourly cost (salary plus employer NI and pension contributions)
Count the hours spent per week or per month — be conservative, not aspirational
Note the error rate: how often does the manual process produce a mistake, and what does fixing it cost?

Step 2: Translate Time Into Money

This is where UK ops leads often get stuck. The simplest approach is to use fully-loaded hourly cost rather than bare salary. A member of staff on a £35,000 salary costs their employer closer to £42,000–£44,000 once employer National Insurance contributions (currently 13.8% above the secondary threshold) and pension auto-enrolment are included. Divide the annual cost by 1,650 billable hours (a reasonable working-year estimate after leave and non-productive time) to get a realistic hourly rate.

Annual Salary	Est. Fully-Loaded Cost	Approx. Hourly Rate
£28,000	~£33,500	~£20/hr
£35,000	~£42,000	~£25/hr
£45,000	~£54,000	~£33/hr
£60,000	~£72,000	~£44/hr

If the process currently consumes 10 hours per week across your team at an average fully-loaded cost of £25/hr, that is £13,000 per year in staff time alone — before you count errors, delays, or the opportunity cost of those people not doing higher-value work.

Step 3: Identify the Costs You Cannot Ignore

A common mistake is treating 'build cost' as the only outgoing. A complete cost picture for a custom internal tool includes all of the following.

Build cost: the development fee or internal dev time required to scope, build, and test the tool
Infrastructure: hosting, database, and any third-party API costs — even modest cloud bills add up over three years
Maintenance and iteration: tools need updates as your processes change; budget 10–20% of the build cost annually as a maintenance reserve
Internal time to spec and test: your team's time during the project is a real cost, even if it does not appear on an invoice
Training and change management: getting your team to actually use the new tool is often underestimated

Step 4: Build the ROI Formula

Once you have both sides of the ledger, the calculation is straightforward. Use a three-year window, which is long enough to show meaningful return but short enough to be credible.

Calculate annual value generated: (hours saved per year × fully-loaded hourly cost) + (errors avoided × average cost per error) + (any headcount growth avoided)
Calculate total three-year cost: build cost + (annual infrastructure + maintenance) × 3 + internal time cost
Apply the formula: ROI = ((3-year value – 3-year cost) ÷ 3-year cost) × 100
Calculate payback period: total cost ÷ annual value = years to break even — this is often the most persuasive single number for a finance sign-off

Note

A payback period under 18 months is typically compelling for UK SME decision-makers. If your numbers show break-even at 12 months or less, lead with that figure rather than the percentage ROI — it is more concrete and harder to argue with.

Step 5: Stress-Test with a Sensitivity Analysis

Run your model at three levels: optimistic (you capture 90% of the projected saving), base case (60%), and conservative (30%). If the tool still pays back within a reasonable window even at 30% of the projected benefit, your business case is robust. If it only works at 90%, you need to either reduce the build cost or be more honest about whether the process problem is real.

Step 6: Add the Qualitative Factors (But Do Not Rely on Them)

Quantitative ROI is your primary argument. But there are real benefits that are genuinely difficult to put a number on, and a well-structured business case acknowledges them without leaning on them as the main justification.

Reduction in key-person dependency (the risk of a process living in one person's head or one spreadsheet they own)
Improved data quality enabling better management decisions
Staff satisfaction: repetitive manual work is a well-documented driver of frustration and attrition
Audit and compliance readiness, particularly relevant for UK businesses subject to HMRC record-keeping requirements or sector-specific regulation

A Worked Example: Stock Reconciliation at a UK Distributor

To make this concrete: an ops team of four at a UK distribution business spends roughly three hours each per week manually reconciling stock data across two systems. At a blended fully-loaded cost of £28/hr, that is £17,472 per year in staff time. Errors in the reconciliation cause one significant stock discrepancy per month, each requiring around four hours to investigate and correct at a senior ops manager's rate — adding roughly £2,100 per year in remediation cost. Total quantifiable annual cost of the manual process: approximately £19,500.

A custom integration tool that automates the reconciliation and flags exceptions might cost £12,000 to build, £1,500/year in infrastructure and maintenance, and £2,000 in internal time during the project. Three-year total cost: approximately £18,500. Three-year value: approximately £58,500. That is a three-year ROI of around 216% and a payback period of roughly 11 months. That is the kind of number that gets a decision made.

Warning

The worked example above uses illustrative figures to demonstrate the method. Your actual numbers will depend on your team's salaries, process complexity, and the specific tool required. Always base your business case on your own measured data, not estimates borrowed from case studies.

When the ROI Case Is Weak — and What That Tells You

If you run this framework and the numbers do not stack up, that is useful information. It usually points to one of three things: the process problem is smaller than it felt, the proposed build scope is too large relative to the saving, or there is a simpler fix (a better-configured off-the-shelf tool, a process change) that should be tried first. A credible technical partner will tell you this rather than proceed with a build that cannot justify itself.