A New Building Does Not Move The Constraint

The decision in front of you

You have a capital request on your desk for a new line, a new building, or a major expansion, and the case says it relieves a capacity problem. Before you sign it, answer one question: which single resource is the binding constraint today, and how many hours a week does that resource actually have? If the file cannot answer in a number, the spend is not buying capacity. It is buying space around an unchanged limit.

This fails most often in the mid-market, where the capital request and the operating problem get described by different people. Finance sees a utilization number that reads high. Operations sees a schedule that will not hold. Sales sees orders it cannot promise. Each of those is real, and none of them identifies the constraint. The request gets approved because everyone agrees that something has to change.

Capital moves the plant, not the limit

Equipment and square footage are physical. Constraints are not always physical. When the binding limit is a decision, a run sequence, or a handoff between two functions, it travels with the organization. Move the same assets into better space and you have redrawn the floor plan around a limit that never moved.

I have been on both sides of this. I have pulled an old project folder on a company years after we first measured it, read what we found then, and seen the same constraints still present and still unaddressed after a relocation. The move removed none of them. The people on that floor could already name the problem. Most people on most floors can. What they could not do was get a decision made about it, and no amount of concrete fixes that.

So the useful question is not whether you are out of capacity. It is what mechanism makes you out of capacity. If the mechanism is a thermal cycle, a changeover, a line rate, or a physical bottleneck you can stand in front of and time, capital is a legitimate candidate. If the mechanism is a quality disposition that sits waiting on a signature, a weekly schedule that gets rebuilt several times before it runs, or a specification that nobody owns, capital is not a candidate. It is an expensive way to postpone a conversation.

What this does to the savings case

Here is the part that gets missed. The capital case and the savings case are usually the same document. When a later diagnosis finds the constraint in planning rather than equipment, the finding is correct and the business case is now wrong. Those two facts arrive together, and they are hard to hold at the same time, which is exactly why they get separated and only one of them gets reported.

The consequence is asymmetric. Depreciation and carrying cost begin at occupancy, at the full approved amount. The constrained resource's available hours per week are unchanged, because nothing touched it. Cost per unit rises first. The recovery the case promised cannot be measured, because the quantity it was measured against never moved. Meanwhile the actual remedy, usually a planning method, a staffing change, and one person with authority to sequence the week, was never in the budget and now has to be funded out of a plant that just absorbed a large fixed cost.

Keep the states separate before you sign. A proposal is not an approval. An approval is not an installation. An installation is not a verified result. A verified result is not a sustained one. A case that collapses all five into a single number is not a case. It is a hope with a payback period attached.

Make the file carry the constraint

Ask for four things, in writing, before the request comes back for a signature.

Name one constrained resource. State its available hours per week today, with the method used to get that number. State how many hours this spend adds to that specific resource, separately from any revenue claim. Name the person who owns converting those added hours into shipped units.

The fourth answer is the one that fails. An engineer can assemble the first three in a week. The fourth requires an executive to assign authority over run sequence to a named person, which is a harder decision than approving money, and is frequently the entire remedy by itself. When the fourth line comes back blank, you have learned something more valuable than the capital answer.

If the answers come back as square footage, room to grow, future flexibility, or a plant-level utilization percentage, defer. Not decline, defer. Send the request back for a constraint measurement first. That measurement costs a small fraction of the capital, and it is the only thing that tells you whether the capital will do anything at all.

What a well governed capital file reads like

One constrained resource is named. Its available hours per week are stated with the measurement method beside them. The hours this spend adds to that resource are stated separately from any revenue or margin claim. One person is named as the owner who converts those hours into shipped units, and that person has stated authority over run sequence. The same measurement is scheduled to run again after start-up, on the same method, so that before and after are comparable. Five lines. If one is missing, the file is not ready, and the missing line is itself the finding.

The next useful action

Pull the last capital request you approved and check whether it named a constrained resource and a measurement method. If it did not, you do not yet know what you bought. Run that constraint measurement now, on the assets you already own, before the next request reaches you. The answer changes your next decision either way, and it is far cheaper to learn it before the money moves than after.