Measuring improvement

Running work automatically is not the same as improving the operation.

The outcome that matters is whether the operation itself gets better.

Three useful lenses are cost, mistakes and bottlenecks.

Cost

How much human and system effort does the work consume?

Automation can reduce repetitive handling, re-keying, chasing and other work that consumes time without requiring much judgement.

Mistakes

Where does work need correction, rework or recovery?

Useful measures depend on the operation, but may include first-time-right rates, correction volume, failed checks or the amount of work returned for rework.

Bottlenecks

Where does work wait, get blocked or fail to keep pace with the business?

This can include queues waiting for a particular person, missing information, supplier responses, approvals or systems that cannot keep up with demand.

Measure the operation, not just the automation

Different operations will have more specific measures, such as processing time, touch rate, exception volume or turnaround time.

The purpose of measurement is to distinguish automation that merely moves work around from automation that genuinely improves the operation.