An improvement is only as credible as the baseline it is measured against. A defensible baseline breaks the cycle into its elements, is measured across more than one run, and records the conditions — part, machine, operator, material — so a later comparison is like for like.
6 min read · Last reviewed 28 September 2026 · All resources
If cutting dominates, the levers are cutting parameters, tooling and material removal strategy. If it does not, the levers are workholding, sequencing and preparation.
Knowing which before you spend money is the entire point of the exercise. It is common for the assumption and the measurement to disagree.
Timing one part on a good day produces a number that improvement will struggle to beat. Timing across shifts and operators produces a range, which is more honest and more useful.
The variation itself is informative. A wide spread usually means the process depends on something not yet controlled.
Agreeing the target and the measurement method before any change is made prevents the argument that follows a disputed result. It also forces the question of what the improvement is worth, which sometimes ends the project early and cheaply.
Cost per part, parts per shift and freeing a bottleneck are different targets that can point to different solutions.
The after-measurement should mirror the before: same definition, same conditions, same number of runs. A result produced by a different method invites doubt no matter how good it is.
Where conditions could not be matched, say so alongside the result rather than leaving it to be discovered.
FAQ
The questions engineers and buyers ask before sending a drawing. If yours is not here, ask it directly — you will get an engineer's answer, not a brochure.
Still have a question?
Talk to Anita M.N, Managing Partner, or send the drawing and let the review answer it.
As one input, yes. It typically measures machine-running time and misses waiting, loading outside the cycle and inspection, so it tends to understate the real interval between parts.
Enough to see the spread rather than a point. Several runs across at least two operators or shifts is a reasonable minimum for most jobs.
Then the baseline takes longer to establish, and that is worth saying openly rather than substituting a single measurement. Infrequent jobs also tend to have higher setup cost, which may be the better target.
Yes. Covert measurement damages trust and produces worse data, because people behave differently when they eventually find out.
Start a project
Share what you are trying to make or fix. Our engineering team reviews every enquiry and responds with the questions that matter, a feasibility view and next steps.
