Choose one bounded task. Measure how often people do it and how much working time it takes, including review and corrections. Compare it with a test that counts the same work. That gives you a basis for deciding whether to continue.
The method below is a simplified workflow calculation. Every number in the example is an invented assumption, not a client result, market price or promised return.
Use the same boundaries before and after
Define a clear start and finish. If the original measurement covers an incoming enquiry through to a reviewed CRM record, keep that boundary in the test. Include cases where the system fails and a person has to redo the work.
Separate working time from waiting time. A task can finish sooner without anyone spending fewer minutes on it. That may still be useful, but describe and assess the shorter wait separately. Record differences in case difficulty and volume between the two periods; comparing different kinds of work weakens the calculation.
A worked example with explicit assumptions
Suppose a company wants to reduce the time spent recording enquiries. The example uses these monthly assumptions:
- 200 cases with the same average difficulty before and after the test.
- 12 minutes of work per case today.
- 6 minutes per case after the test, including review, corrections and exceptions.
- SEK 400 per hour as an assumed value for released capacity.
- SEK 2,000 in additional running costs.
- SEK 30,000 in total one-off effort, including a value assigned to internal implementation time.
Calculate the value under those assumptions
The current workload is 200 × 12 / 60 = 40 working hours per month. After the test it is 200 × 6 / 60 = 20 hours. The difference is 20 hours, valued at 20 × 400 = SEK 8,000 using the assumed hourly figure.
Subtracting running costs leaves SEK 6,000 per month in this simplified calculation. Dividing the one-off effort by that value gives 30,000 / 6,000 = 5 months. This assumes unchanged volume, time, costs and valuation from the first month. It does not allow for a slower introduction or further change costs.
Decide how the released time will be used
Five months in this example does not prove that cash returns to the bank account. If the same people work the same paid hours, the company has gained capacity to use elsewhere.
Describe that use. Will people handle more enquiries, reduce overtime or catch up on neglected work? Count an actual cost reduction only when you can identify the expense that disappears. Do not count released time and the cost reduction from that same time as two separate benefits.
Try a worse outcome
Suppose each case takes 9 minutes after the test. That releases 10 hours per month, valued at SEK 4,000. Running costs leave SEK 2,000. The same one-off effort now corresponds to 15 months in the calculation.
At 10.5 minutes per case, the released 5 hours are valued at SEK 2,000. That covers only the running costs; the one-off effort is not recovered under those assumptions. This helps identify the improvement the test would need to demonstrate.
Set the test criteria before judging the result
Record acceptable quality, working time, exceptions and total cost. Where evidence is thin, use a range and identify what needs better measurement. Do not treat possible future revenue as certain.
To discuss a workflow, bring its approximate volume and time requirements. Leave sensitive information and customer records out of the contact form.
