New Line or Retool? A Decision Framework for Manufacturers at Capacity

A brand hits capacity and the conversation starts the same way almost every time: we need another line. Sometimes that's right. More often, what the plant needs is a $60K change to one station — and the $700K line would have bought a second copy of the same bottleneck.

The distinction matters because the wrong answer is expensive twice: once in capital, and again in the 18 months you spend discovering the new line didn't fix the problem. Here's the sequence we use to tell them apart.

Step 1: Measure every stage, not just the slow-looking one

Walk the process end to end and get a real rate for each stage — units per minute at steady state, with the operators who actually run it, on a normal day. Not the equipment nameplate. Not best-shift. The number you can hold for eight hours.

Then write them in a row. The lowest number is your line rate; every other stage is running at partial utilization whether it looks busy or not. This sounds obvious, and it is — but in practice most plants have never written the row down, which is why the loudest station usually gets blamed instead of the slowest one.

Two things fall out immediately. First, the constraint is often not where people assume: an upstream stage starves the line, or a downstream one backs it up, and the machine in the middle takes the blame for both. Second, you learn the size of the prize. If your constraint runs at 44 units per minute and the next-slowest stage does 60, then 60 is your ceiling — fixing the constraint gets you a 36% lift and not one unit more. Every dollar spent above that ceiling is buying nothing until you also move the next stage.

Step 2: Ask what the constraint is actually made of

"The machine is too slow" is a symptom, not a cause. Constraints are almost always one of five things, and the right fix differs enormously by which one:

• Cycle-time limited. The mechanism itself can't go faster — a physical or thermal limit. This is the only case where more equipment is the honest answer.
• Changeover limited. The machine is fast but idle. If you run six SKUs and burn 90 minutes per change, your capacity problem is a tooling and fixturing problem.
• Reliability limited. Nameplate is fine; uptime is 68%. A new line inherits the same maintenance practice and lands at 68% too — you've just doubled the thing you weren't maintaining.
• Labor limited. Throughput tracks headcount because someone hand-feeds, hand-packs, or hand-inspects. Targeted automation at one station is usually a fraction of a line's cost.
• Quality limited. You're running fast and scrapping the difference. Yield, not rate, is the constraint — and running faster makes it worse.

Get the downtime and scrap data before this conversation, not during it. A week of honest logging settles arguments that opinions can't.

Step 3: Apply the retool test

Once you know the constraint and its cause, the decision usually resolves against four questions:

• Is there headroom in the rest of the line? If the other stages can absorb more volume, you have a station problem, not a line problem. Retool.
• Is the constraint mechanical, or is it practice? Changeover, uptime, and labor constraints are practice. New steel doesn't fix practice.
• Does the product itself change? New format, new pack, new process step — that's often a genuine new-line trigger, because you're not scaling the old thing, you're making a different one.
• How far past the ceiling do you need to go? Retooling typically buys 20–60% on a single stage. If you need 3x, stop optimizing and design the line you actually need.

What each path really costs

Rough ranges from projects we've delivered, and they vary a lot with product and plant. Treat them as orders of magnitude, not quotes:

• Station retool or tooling change: tens of thousands, weeks to a few months, no production shutdown if it's planned properly.
• Targeted automation at one station: mid five figures to low six, a few months, usually installable over a weekend or a planned down window.
• New line or major buildout: several hundred thousand and up, six to eighteen months, with facility, utilities, and training riding along behind it.

Note the pattern: the cheap options are also the fast ones and the reversible ones. That matters when demand forecasts are uncertain — and they always are.

The case for the new line

None of this is an argument against building. Sometimes the answer is genuinely a new line, and the tell is usually one of these: the constraint is a hard physical limit; you need a multiple, not a percentage; the product or format is actually different; the existing equipment is at end of life and spares have gone unobtainable; or the facility itself — power, floor space, drainage — is the constraint. When two or more of those are true at once, incremental fixes turn into a treadmill and you should stop paying for it.

Before you sign anything

A short checklist worth running whoever you hire:

• Do we have a measured rate for every stage, not just the suspect one?
• Do we know the ceiling — the rate of the next-slowest stage?
• Do we know which of the five causes we're dealing with, with data behind it?
• Have we priced the retool option honestly, or only the one we already wanted?
• If this project hits its number, what becomes the constraint next? There's always a next one. Knowing it in advance is what keeps you off the treadmill.

That last question is the one most capital requests skip, and it's the one that separates a plan from a purchase.