Agree what counts as finished
Choose one output, its responsible reviewer and the acceptance standard. Define the unit of work before measuring: one client file, one reporting pack or one evidence request. Record case complexity, source availability and exceptions so easier cases cannot quietly improve the apparent result.
This is Halyard's proposed practical method, not a validated industry benchmark, regulatory standard or promise of improvement.
Measure both effort and elapsed time
For each case, record preparation minutes, review minutes and correction minutes without counting the same work twice. Their sum is total active effort. Separately record elapsed time from request to accepted completion, including waiting.
Keep setup, training, maintenance and recurring supplier costs visible outside per-case effort. Record who worked on the case and whether it met the acceptance standard. Retain failed and abandoned cases, together with the reason.
Compare like with like
Use representative ordinary cases and deliberately selected exceptions. Match periods, scope, output quality and reviewer expectations. If the same person repeats a case, note the learning effect and vary test order where practical.
Report case counts, distribution and limitations alongside any average or median. A small convenient sample can inform a scoped next step; it cannot establish firm-wide savings. Do not extrapolate across different teams or entities without checking their work.
Keep failures separate from speed
Track incorrect statements, unsupported explanations, wrong versions, incomplete outputs, permission failures and corrections. Record their severity and whether the review caught them. Critical issues cannot be averaged away by time savings.
Agree stop conditions before the test. A material information-control failure, absent approval responsibility or unmanageable correction burden should pause progression until resolved and retested.
Make the economics explicit
For matched cases, per-case effort change equals baseline total minutes minus trial total minutes. Multiply a cautiously supported typical change by eligible case volume to estimate potential capacity. This is capacity, not cash savings; people may use the time for other work.
Where leadership needs a financial model, state the assumed cost rate and subtract ongoing supplier, support and monitoring costs. Include implementation and training costs in any payback calculation. A zero or negative net benefit has no positive payback period. If key inputs are unknown, report them as unknown.
Record the decision, including a decision to stop
A useful result states the sample, period, method, complete-task effort, quality, exceptions, controls, adoption and limitations. The sponsor then decides to proceed, narrow, remediate and retest, or stop.
The downloadable worksheet is a collection aid. Store completed worksheets with client evidence in an approved private location. Publish client-specific results only after verification and permission.
Start with the measurement worksheet
Download a blank CSV with one row per case and phase. Record baseline and trial separately. Keep completed client worksheets in an approved private location.