Check whether the work improved, including the awkward cases.
A successful demonstration proves that one example worked. To judge an improvement, compare the same kind of task before and after the change, including the review and correction the new process needs.
Define the beginning and end.
For an enquiry, the task might begin when the message arrives and end when the owner has recorded a valid next action. Keep that definition the same in both periods. Do not compare a complete old process with only the fastest part of the new one.
Record the task type, handling time, waiting time, errors and any correction. Keep personal details out of the measurement sheet when they are not needed.
Look at a small example honestly.
Synthetic example, not client results: five comparable tasks took 8, 7, 9, 8 and 8 minutes before a change. Five afterwards took 5, 6, 5, 12 and 5 minutes, including review and corrections.
The means are 8.0 and 6.6 minutes. The medians are 8 and 5 minutes. The 12-minute exception matters: the new process is faster on these examples on average, but one case took longer. Five tasks per period are too few to establish a stable long-term result.
Keep other explanations visible.
Were the later enquiries simpler? Did a different person do the work? Did the team learn the task, or did volume change? Record those differences before attributing the whole change to the new workflow.
Compare matched task types where possible and repeat the observation over ordinary working conditions. Do not remove difficult cases merely because they make the improvement look smaller.
Check correctness alongside speed.
A faster wrong result is not a successful improvement.
- Accuracy
- Did the correct information reach the correct record?
- Completeness
- Was the task finished, or merely queued or drafted?
- Exceptions
- Were missing information and failures visible and handled?
- Customer impact
- Were duplicate messages, missed handoffs or incorrect promises avoided?
- Ownership
- Could the team use and recover the process without relying on one person?
Report the result with its limits.
Say what you measured, the period, the number of examples and what changed. Distinguish time released from money saved or new revenue. Keep the source observations so a later review can check the claim.
Use the result to choose the next action: retain the change, improve its exception handling, simplify it or stop. A measured limitation is useful information about how the workflow should develop.
