Which processes should you automate? A prioritisation method and ROI calculation
Automation tools get easier every year, but not every task is worth automating. Automating work that is rare, constantly changing or heavily dependent on human judgement can create more maintenance than it saves. This article presents a simple scoring method for deciding which processes to tackle first, and a return on investment calculation.
What makes a process a good candidate?
A good automation candidate usually has most of these traits:
- High frequency: It happens many times a day or week.
- Clear rules: It can be described as "if this happens, do that".
- Digital input: The data comes from a form, email, spreadsheet, database or API.
- High cost of errors: When done by hand, forgetting or mistyping costs money or customers.
- Connects several systems: Data is read from one system and typed into another.
With AI steps, the "clear rules" requirement has become more flexible: understanding the topic of an email, reading the amount on an invoice or estimating the urgency of a request can now be part of an automation. Where the final decision is critical, keep a human approval step.
Scoring table
List your candidate processes and score each from 1 to 5 against these five criteria:
| Criterion | 1 point | 5 points |
|---|---|---|
| Frequency | A few times a month | Dozens of times a day |
| Duration | 1–2 minutes each time | 15 minutes or more each time |
| Rule clarity | Requires judgement every time | Entirely rule-based |
| Cost of errors | Negligible | Loses money or customers |
| Technical access | Data on paper or in a closed system | Data readable via API or spreadsheet |
The processes with the highest total are the best candidates for a first pilot. For processes that score high but have poor technical access, the data may need to be digitised first.
ROI calculation
A simple return on investment calculation needs:
- Time per occurrence (minutes)
- Occurrences per month
- Hourly cost of the person doing the task
- Build cost and monthly running cost of the automation (server, API, maintenance)
Hours saved per month = (time per occurrence × occurrences per month) ÷ 60
Monthly saving = hours saved per month × hourly cost − monthly running cost
Payback period (months) = build cost ÷ monthly saving
Worked example
The figures below are illustrative values chosen to demonstrate the method; use your own business data for a real calculation.
Suppose enquiries from a web form and email are copied into the CRM by hand, forwarded to a sales rep and acknowledged to the customer:
- Time per occurrence: 6 minutes
- Enquiries per month: 400
- Hours saved per month: 6 × 400 ÷ 60 = 40 hours
Apply the hourly staff cost and the automation's running cost to those 40 hours and the payback period becomes clear. Also note the gains that are not in the formula but are often the largest:
- Faster response: Replying within minutes can increase conversion.
- No lost enquiries: Requests that slipped through the manual process drop to zero.
- Data quality: Complete and consistent CRM records make reporting reliable.
Common mistakes
- Starting with the most complex process: The first project drags on and the team loses momentum. The first pilot should be small and measurable.
- Automating a broken process: Remove unnecessary approval steps and duplicate entries before automating.
- Not measuring the baseline: Without measuring the time before automation, the benefit cannot be proven.
- Ignoring failure scenarios: Decide up front what happens when a service is down or unexpected data arrives.
- Leaving it without an owner: Every automation needs an owner and documentation.
From pilot to scale
The recommended path:
- List and score 5–10 candidate processes.
- Choose one of the highest-scoring processes as a pilot and measure its current duration.
- Launch the pilot within 1–2 weeks and measure for a month.
- Add the second and third processes based on the results, reusing shared components (CRM connection, notification templates).
Conclusion
The value of automation is measured not by how many workflows are built but by how much time and error cost they remove. A simple scoring table and ROI calculation show where to start and base the investment decision on numbers.
uurstudios reviews your processes, identifies the steps best suited to automation, builds the pilot and measures the results. For tool choice and setup, read business process automation with n8n; for services, see the overview.