Scope the Process, Not the Machine You Can Count

The correction that came back on the scope document

I sent a draft scope for an optimization trial to the leader who runs the plant. He came back aligned on the detail, with one callout. Under the scope and trial boundary section, the document named a specific packaging machine, more than once, and an earlier conversation had named a different machine. His correction was one line: the scope is not that machine, the scope is the feed of raw roll stock, trim through cook and chill.

That sentence was worth more than the document it corrected. A machine name in a trial boundary reads as precision. It is usually the opposite. It is the place the process first becomes countable, and it reached the document because that is where the numbers are clean.

Why the countable asset is not the boundary

Output at a packaging machine can be counted per unit, per hour and per shift, with yield calculated off what goes in. The feed process upstream produces something much harder to count: a balance of cuts, a rate that moves with what is running, an operator allocation that shifts during the day. When someone writes a scope, the countable point attracts the boundary. The study ends up with a clean measurement and the wrong subject.

Three consequences follow, and all of them are visible before the trial starts.

The asset cannot be isolated from its feed. Its observed rate is a function of what the upstream process hands it and in what balance, so any rate recorded there belongs partly to a process outside the boundary. The finding still gets written against the machine, which is how a feed problem becomes an equipment request.

The asset may not carry the volume that matters. The machine named in that document ran a product with falling demand, far enough that the plant was struggling to hold full time hours against it and there was a live case for moving the volume elsewhere. The opportunity the plant actually cared about ran on a different product through the same process. A boundary at that machine would have produced a real improvement on volume the business was preparing to give up.

The line does not run the way a single-asset boundary assumes. The objection was practical: you cannot spend a week studying one product when that product runs hard for a few days and then the line switches. The two products need different trim balance and different feed, so the trial produces two streams of data, not one. A boundary drawn at the machine hides that split, because the machine looks like a single thing.

Three tests before you fix the boundary

Ask for the one-sentence version. Have the person who schedules the line state the scope in their own words. Compare it to your document. If theirs is a process with a start step and an end step and yours is an asset, yours is a measurement convenience wearing the clothes of a boundary.

Check the asset's volume share and its direction. Pull the share of line hours by product through that asset and the demand trend for each. An asset can be the best available measurement point and the worst possible subject at the same time.

Check whether the name has drifted. If the asset named in this draft is not the asset named in the last conversation, nobody has anchored the boundary yet, and the document is recording a habit rather than a decision.

None of this means stop counting at the machine. Keep it as the instrument and say so explicitly. The scope sentence names the process, the measurement note names the asset, and they are two different lines in the document.

What a well-scoped trial reads like

The scope statement names a start step and an end step. It lists the product streams that run through that span and the share of line hours each one takes. It names the asset once, as the point where output is counted, with the yield basis stated beside it. It states how many data streams the trial is expected to produce, so a split result is an expectation rather than a surprise. The person who schedules the line has read it and can repeat the boundary from memory in one sentence.

Read your own scope back

Before the next trial is approved, underline every asset name in your scope document. For each one, decide whether it is the subject or the instrument, and write that word next to it. If you cannot tell which it is, the person who runs the line can tell you in a sentence, and that sentence is your scope.