How to calculate automation ROI
The formula, the inputs that matter, a worked example and the pitfalls that make most automation business cases look better than they are. Practical for UK SMEs in 2026.
The formula
The honest formula for automation ROI is simpler than most consultancies will pretend:
ROI = (Annual benefit − Annual cost) ÷ Annual cost × 100
Where:
- Annual benefit = (hours saved × fully-loaded hourly cost) + (error-cost avoided) + (revenue unlocked).
- Annual cost = software subscriptions + (build cost ÷ amortisation period in years) + maintenance hours × hourly cost.
If the result is positive and above ~100%, the project is worth doing. Below that, the maths is telling you to either reduce scope, do it cheaper, or pick a different target workflow.
The inputs that actually matter
Most ROI templates online walk you through 20 inputs. In practice, four drive 80% of the answer:
- Hours saved per week — the dominant variable. Multiply weekly hours by 50 working weeks (assuming holidays).
- Fully-loaded hourly cost — not gross salary. Total cost to business (salary + employer's NI + pension + benefits + equipment + software per head) ÷ working hours.
- Build cost — what you pay (or your team's time at fully-loaded rate) to ship the automation.
- Maintenance burden — 0.5–2 hours per workflow per month plus the platform subscription.
The other inputs (error-cost avoided, revenue unlocked, customer-satisfaction lift) are often material but harder to estimate. Use them where they're real and obvious; leave them out where they're hand-waving.
Worked example: A £15k automation engagement
Say you're a 30-person UK marketing agency. You're considering a £15k fixed-fee project to automate four workflows: lead-to-CRM, deal-to-project-kickoff, time-to-invoice and freelancer payment.
Build cost (year 1):
- £15k consulting fee.
- 25 hours of internal time during discovery and build, across 4 staff at average fully-loaded rate £45/hr = £1,125.
- Total: £16,125 in year 1.
Ongoing cost (each year):
- Make.com Pro subscription: £100/mo × 12 = £1,200.
- Maintenance: 4 workflows × 1 hour/month × £45/hr fully loaded = £180/mo = £2,160.
- Total: £3,360 per year ongoing.
Benefit:
- Lead-to-CRM saves 4 hrs/week (account manager admin) at £35/hr fully loaded.
- Deal-to-project saves 2 hrs per new project × 1 project/week = 2 hrs/week × £45/hr (senior PM time).
- Time-to-invoice saves 8 hrs at month-end for the finance person × 12 months = 96 hrs/year at £40/hr.
- Freelancer payment saves 3 hrs/week (finance) at £40/hr.
- Hours saved: (4 × 50) + (2 × 50) + 96 + (3 × 50) = 200 + 100 + 96 + 150 = 546 hours/year.
- Weighted average fully loaded rate: about £40/hr.
- Time-saved benefit: ~£21,840/year.
Also include conservative error-cost avoided:
- Late invoices: 4 invoices/year recovered 30 days faster = £200 × 4 = £800/year (cash-flow improvement at 6% discount rate).
- Freelancer payment errors avoided: 6 incidents/year × £150 admin to fix = £900/year.
- Error-cost avoided: ~£1,700/year.
Total annual benefit: £21,840 + £1,700 = £23,540.
Year 1 net: £23,540 − £16,125 (year-1 cost) − £3,360 (ongoing) = £4,055 net positive.
Year 1 ROI: (£23,540 − £19,485) ÷ £19,485 = 21% in year one. Just positive.
Year 2 net: £23,540 − £3,360 = £20,180 positive.
Year 2 ROI: (£23,540 − £3,360) ÷ £3,360 = 600%.
Most projects clear ~20–80% in year one and shoot up in year two when the build cost is gone. Three-year NPV at 10% discount rate on this example is roughly £42k on a £15k investment — a 2.8× return.
The pitfalls that make most ROI numbers misleading
In rough order of frequency:
- Ignoring fully-loaded cost. Using gross salary instead of fully-loaded undercounts savings by ~30%.
- Forgetting ongoing maintenance. "It's done" is never done. Real workflows take ongoing time.
- Counting hours as if they were freed for revenue work. If your staff member just leaves slightly earlier, the £21/hr saving is real but doesn't convert to new revenue. Be honest about whether saved time becomes new output.
- Double-counting. Lead-to-CRM saving "4 hours" and deal-to-project saving "2 hours" might overlap if the same person is doing both. Map the actual time-by-person, not just by workflow.
- Optimism on adoption. A workflow only delivers savings if the team actually uses it. Many automations are launched, then quietly worked around. Assume 80% adoption, not 100%, in the first 6 months.
- Hidden risk-cost. New automations introduce new failure modes. Build error-recovery time into ongoing maintenance.
- Vanity savings. "This dashboard saves the partner an hour a week" — does it? Or do they ignore it? Only count savings that genuinely change behaviour.
Applying these honestly typically takes a paper-napkin 600% ROI down to a realistic 250–400% — still excellent, but defensible to a CFO.
A simple framework for your own paper-napkin calculation
For each candidate workflow, fill in:
- Workflow name: ___
- Frequency: how often it runs (per week/month)
- Time per run (manual): in minutes
- Time per run (automated): in minutes (usually near zero, but allow for review)
- Staff member doing it: title and fully-loaded hourly rate
- Error rate (manual): % of times something goes wrong
- Error cost: what it costs each time something goes wrong
- Frequency × time saved/run × hours = annual hours saved
- Annual hours saved × fully-loaded rate = direct savings
- Error-cost avoided per year = (manual error rate − automated error rate) × frequency × error cost
- Total annual benefit = direct savings + error-cost avoided
Then subtract software + maintenance + amortised build cost. If the answer is positive and above ~100% in year one, build it. If not, look at a cheaper build or pick a different workflow.
When ROI calculations mislead
Three situations where the numbers will mislead you:
- Compliance workflows. Where the cost of a single error is regulatory (FCA fine, ICO sanction, SRA investigation), the dominant variable is risk reduction, not time saved. Don't price the project on hours saved; price it on tail risk avoided.
- Strategic / customer-experience workflows. Where the workflow makes a high-stakes customer touchpoint smoother (onboarding, RFP response, executive sponsor briefing), the dominant variable is conversion or retention impact, not time saved.
- First-of-a-class infrastructure. The first workflow you build creates the platform that the next 10 workflows ride on. Don't try to make the first workflow pay for the whole platform — amortise across the first 5–10 builds.
Related reading on Watermelon
- Business Process Automation in the UK — the broader discipline.
- Workflow Automation for UK Small Business — what to automate first.
- Zapier vs Make vs n8n (2026) — platform deep-dive.
- What is a Fractional CAO? — when to hire ongoing leadership.
- Automation Consulting — what we deliver and how it's priced.
- Industries: Agencies, Ecommerce & DTC, Recruitment, Professional Services, B2B SaaS, Accounting Firms.
Ready to talk?
The free 30-minute call is the right next step. Bring one workflow you'd want to make an honest business case for. We'll walk through the maths with you — even if we're not the right consultancy for the build.
Want to pressure-test your automation business case?
30 minutes. No deck. Bring one workflow you'd want to make a business case for. We'll walk through the maths honestly.