Every automation project starts with the same question: will this pay for itself?
The answer matters. Automation requires investment—software costs, implementation time, change management. If the return doesn't exceed the investment, you've wasted resources that could have gone elsewhere.
Here's a practical framework for calculating automation ROI, designed for UK SMEs who need business cases that actually work.
Step 1: Identify the Process
Start by picking a process to evaluate. Good candidates for automation share common characteristics:
High volume. The process runs frequently—daily, weekly, or triggered by common events. Low-volume processes rarely justify automation investment.
Rule-based. The steps follow clear logic. If X happens, do Y. Processes requiring significant judgment or creativity are harder to automate.
Time-consuming. The process consumes meaningful hours across your team. Automating a two-minute task won't move the needle.
Error-prone. Manual processes involving data entry, copying, or handoffs tend to have error rates of 1-5%. Errors cost money to fix.
Invoice processing, employee onboarding, report generation, and approval workflows typically tick all these boxes.
Step 2: Measure the Current State
You can't calculate improvement without a baseline. For your chosen process, measure:
Frequency. How many times does this process run per month? Count invoices processed, reports generated, requests handled.
Time per instance. How long does each instance take? Be comprehensive—include waiting time, chasing time, and rework time, not just active processing.
People involved. Who touches this process? Note their roles and approximate salary costs.
Error rate. What percentage of instances have errors requiring correction? What's the cost of each error—time to fix, penalties incurred, customer impact?
Cycle time. How long does the process take end-to-end? From invoice receipt to payment posting, for example.
Be honest with these numbers. Underestimating current costs will undermine your business case. Track the process for a week if you don't have data.
Step 3: Calculate Current Cost
Now convert those measurements into money.
Labour cost formula: (Time per instance × Instances per month × Hourly rate)
For example: 30 minutes per invoice × 200 invoices × £25/hour = £2,500/month
Error cost formula: (Error rate × Instances per month × Cost per error)
For example: 3% error rate × 200 invoices × £50 per error = £300/month
Opportunity cost is harder to quantify but often matters most. What would your team do with reclaimed time? What decisions are delayed because reports aren't ready?
Total current cost = Labour cost + Error cost + Opportunity cost
For our invoice example: £2,500 + £300 + estimated opportunity cost = £2,800+ per month
Step 4: Project the Automated State
Research what automation typically achieves for your process type:
Time reduction: Most automation delivers 60-80% time savings on suitable processes. The remaining time covers exceptions, quality checks, and oversight.
Error reduction: Automation typically reduces errors by 80-95%. Machines don't fat-finger data entry.
Cycle time reduction: End-to-end process time often drops by 50-70%. Automated steps don't wait for people to be available.
Apply these factors to your current state:
Projected labour cost: £2,500 × 25% (retaining 25% of manual effort) = £625/month
Projected error cost: £300 × 10% (retaining 10% of errors) = £30/month
Projected total cost: £625 + £30 = £655/month
Step 5: Calculate the Benefit
Monthly benefit = Current cost - Projected cost £2,800 - £655 = £2,145/month
Annual benefit = Monthly benefit × 12 £2,145 × 12 = £25,740/year
This is your gross benefit—the value automation creates before considering what it costs to implement.
Step 6: Factor in Automation Costs
Automation isn't free. Include all costs:
Software licensing. Power Automate Premium is approximately £12/user/month. Specialist tools vary. Get actual quotes.
Implementation. Whether internal or working with a Power Automate consultancy or automation agency in London, building the automation takes time. A typical SME automation project runs £3,000-£15,000 depending on complexity.
Change management. Training, process documentation, dealing with resistance. Budget 10-20% of implementation cost.
Ongoing maintenance. Automations need monitoring and occasional updates. Budget 10-15% of implementation cost annually.
For our invoice example:
- Implementation: £8,000
- Software: £1,200/year
- Maintenance: £1,000/year
First year total cost: £10,200 Ongoing annual cost: £2,200
Step 7: Calculate ROI
Net first-year benefit: £25,740 - £10,200 = £15,540 First-year ROI: £15,540 / £10,200 = 152% Payback period: £10,200 / (£25,740/12) = 4.7 months
Ongoing annual ROI: (£25,740 - £2,200) / £2,200 = 1,070%
These numbers make the decision obvious. But not every automation delivers returns this strong—which is why running the calculation matters.
When to Walk Away
Not every process should be automated. Walk away when:
- Volume is too low to justify implementation cost
- The process genuinely requires human judgment throughout
- The underlying process is broken (fix it first, then consider automation)
- The ROI is marginal and implementation risk is high
A negative or marginal ROI isn't failure—it's useful information that prevents wasteful investment.
Making the Case
Once you've calculated ROI, present it simply:
"Automating invoice processing will cost £10,200 in year one and save £25,740, delivering payback in under five months. After year one, it saves £23,500 annually."
That's a business case decision-makers can act on. No jargon, no hand-waving—just numbers that work.
A good workflow automation agency or process automation consultancy can help you build these business cases for multiple processes, prioritising by ROI and implementation complexity.