"Was it worth it?" Someone will ask you this question: management, controlling, maybe you yourself. And in six months you should be able to answer it without guessing and without dressing up slides. That only works if you measure the right thing today. How an iterative approach turns an optimisation into provable evidence rather than a good feeling.
No proof without a baseline
The most common reason a success can't be proven is banal: nobody measured the state beforehand. Without a starting value there's no comparison, and without comparison every improvement stays an assertion.
A solid baseline needs only a few clearly defined figures, captured before anything is changed:
- Cycle time: How long does a case take from start to finish, including waiting times?
- Error and rework rate: How many cases have to be corrected or repeated?
- Cost per case: Processing time times labour cost, plus error cost.
- Manual share: How many steps run by hand, how many automated?
These four figures are not bureaucratic busywork. They are the zero point against which every later number can be measured. Skip them and you optimise blind, and you'll later argue about impressions instead of results.
One distinction is worth knowing on top of this: these figures are lagging indicators. Cycle time and error rate only show afterwards whether things got better. They work as proof, but not for steering. To steer while the work is still underway, you also need leading indicators: the degree of automation and the manual share change the moment a step takes hold, long before cycle time visibly moves. Leading indicators warn early, lagging indicators prove at the end. Measure both and you can intervene instead of merely documenting.