Manufacturing Capacity Planning: The Operator’s Guide (No New Software Required)

If you’ve been searching “manufacturing capacity planning,” I can tell you what you found. A definition, a formula, three planning strategies with names like lead, lag, and match, and then, right on schedule, the reason that page exists: a pitch for the software company that wrote it.

I’m going to give you the parts of that worth keeping. Then I’m going to show you how capacity planning actually works on a real floor, using a real manufacturer that doubled daily throughput in a matter of weeks without buying equipment, adding people, or spending a dollar on new software.

The Formula Everyone Starts With

The textbook version goes like this: count your machines, multiply by available hours, divide by cycle time, and you have your production capacity. Multiply by your utilization rate for a dose of realism. Done.

That math is useful for sizing a building. It’s almost useless for the question you’re actually asking, which is some version of: why can’t we keep up with demand, can we take on more work, what can we promise customers, and when do we finally add people or equipment?

The formula gives you theoretical capacity, what the system could do if nothing ever went wrong. What pays the bills is demonstrated capacity, what actually comes out the back door week after week. The gap between those two numbers is where most shops live. No spreadsheet closes it, because the gap isn’t a math problem.

After nearly 30 years of doing this work, here’s the pattern I keep seeing: most capacity problems aren’t capacity problems. They’re synchronization problems. The machines and the people are already there. The work just doesn’t flow through them in any controlled way. Which means before you can plan capacity, you have to find out how much you actually have. Most companies have never seen that number.

Where I Start: The Data and the Walk

When a client tells me they can’t keep up, I’m trying to understand two things before anything else: current volume and work in process.

The first part is historical data. What has throughput actually been, week over week, for the last several months? Not the standard, not the quote assumption. What demonstrably shipped.

The second part is a physical walk. I want to see where work in process is sitting. Not the WIP number on a report, the actual piles. Which departments have work stacked in front of them? Which ones are starved and waiting?

Here’s what makes this hard at a lot of companies: they aren’t really measuring work in process, usually because of how their accounting treats it. Or they track WIP as a single dollar figure, which tells you almost nothing. That number swings with order mix and order value. It doesn’t tell you how work is flowing. When you map WIP at each touchpoint instead, each department or work center where a job changes hands, the flow problem shows itself. The piles point at it.

While I’m walking, I’m also asking what’s changed. Are we down four people from last year? Is a key piece of equipment limping along in disrepair? Capacity erodes quietly, and nobody updates the assumptions.

That’s the whole diagnostic. Throughput history, WIP by touchpoint, and an honest inventory of what’s changed. No software required. You could do it this week.

A Real Example: Doubling Throughput Without Spending a Dollar

Let me show you what this looks like in practice, because the numbers matter.

A custom manufacturer I work with was averaging four jobs per day through their final production step. Demand was running six to eight jobs per day. You don’t need a consultant to do that math: the backlog grew every single week, and lead times grew with it.

When we mapped their WIP, they had 31 jobs sitting inside manufacturing, stuck in queues between four production steps. One person was running three of those steps as a single combined process, so a job could get two-thirds of the way through and stall when the next thing came up. Worse, jobs kept getting released with material shortages. Work would make it partway through the process, hit a missing component, and get pulled back out. Every one of those starts and stops added days.

We made two changes. Neither cost anything.

First, we split the combined steps into separate operations with defined queues, so work moved through clear stages instead of living inside one person’s juggling act.

Second, we stopped releasing any job that didn’t have everything it needed. Full kit: all materials, all information, complete before release. If a job couldn’t run start to finish, it didn’t enter the system. It sat on the bench until it was ready.

Within weeks, throughput doubled to an average of 10 jobs per day. During part of that stretch they were down a couple of people, and they still outpaced everything they’d done before. No new equipment, no new hires, no overtime push. The operations leader told me the strangest part was the quiet: “I don’t even know what to do. It’s so quiet.” The flurry of expediting was gone because the chaos that made it necessary was gone.

The leadership team now believes 16 jobs per day is realistic with the same resource load. That’s four times where they started, and the only thing that changed is how work gets released and managed.

That capacity was there the entire time. No formula would have found it, and no software purchase would have either. It was buried under the way work moved through the building.

The Counterintuitive Part: Some Departments Should Be Less Efficient

Everything in that story runs on one idea, and it’s the one people struggle with most.

Every operation has a control point, the step that governs the pace of the whole system. Some people call it the bottleneck or the constraint. Your entire operation can only produce what that one step can produce. Nothing else sets the pace, no matter how hard everyone else works.

A production manager at that same client said it better than I could: one department “could churn out a hundred jobs per day, but if the final step can only do five, trying to increase the efficiency of anything else doesn’t do anything to help.”

They had lived that. Before we worked together, they kept improving whichever department looked slow. The improvements were real. The throughput never moved, because none of those departments were the constraint.

Once you see it, the playbook inverts. You want maximum efficiency at the control point and, deliberately, less efficiency everywhere else. Idle time at a non-constraint isn’t waste. It’s protective capacity, and it’s healthy. When your best people have open time because the queues are managed, you move them to the control point and keep it fed.

This is also why I’ll tell a shop that’s drowning in orders to freeze work, which sounds backwards until you watch it happen. Pull back anything that isn’t full kit. Pull back anything that isn’t due inside the lead time window you’re trying to protect. The instinct to release everything so everyone stays busy is exactly what buries the constraint and stretches every job’s lead time. Release work at the pace the control point can consume it, and the whole system speeds up.

We’ve been trained our entire careers to seek efficiency inside our own area of responsibility. This flips that. The goal is a company that’s effective, not departments that are busy.

When It’s Actually Time to Add Capacity

So when do you buy the machine? When do you add the shift?

Later than you think, and the sequence matters more than the purchase.

Here’s what happens when you fix flow: you chew through the backlog faster than you expect, and then the system starts to starve. That’s not a problem, that’s the plan working. The constraint has moved out of your building and into your book of orders. That same manufacturer hit this point. The queues emptied, and suddenly people had capacity with not enough new orders coming in.

The sequence I walk clients through looks like this. Fix the operational reliability first. Burn down the backlog. Watch throughput climb, cash flow improve, and your reliable lead time shrink. Then take that story to the front end of the business and sell it: “It used to take us four weeks. We now do that same order in four days.” Very few competitors can match that pace, and it wins work. Then, as new orders flow in, watch how much growth you can absorb with the resources you already have. Most companies are shocked by the answer.

Only after all of that do you spend money, and you spend it at the constraint, where added capacity actually converts into throughput. A dollar spent anywhere else buys you nothing but a bigger pile of WIP.

One warning on this. People who’ve read a little theory of constraints believe the bottleneck constantly jumps around, so they chase it like whack-a-mole. It can move, and you can pick the wrong spot to elevate. But it very rarely bounces around the way people imagine. Companies that believe it does, keep spot-treating one area after another, which perpetuates the exact problem they’re trying to escape: unbalanced capacity, reduced throughput, growing lead times.

The Software Question

Half the searches around capacity planning are for software and tools, so let’s deal with it directly.

You don’t need new software to do anything I’ve described. Use whatever you already have. An ERP, an MRP, a whiteboard, a job board, Excel. The client in this story runs their queue tracking in a spreadsheet the team built themselves along with a whiteboard in manufacturing. The tool genuinely does not matter. What matters is that you’re tracking queues at your touchpoints, releasing full kit work at the pace of your control point, and watching your demonstrated throughput trend.

Scale changes the tooling, not the thinking. If you run 100 jobs a month, a whiteboard works fine. If you run 10,000, you need software to keep track, and you probably already own it. At the core, every ERP on the market is the same machine: a transaction aggregator, a giant balance sheet. The features the salesperson demos are wrapped around fundamentals that haven’t changed in decades. Pick a tool appropriate to your complexity and stop worrying that you picked the wrong brand.

I have yet to walk into a company where we couldn’t find a no-cost or low-cost way to track and execute everything flow improvement requires. Not one. If a vendor tells you the starting point is their platform, you’re hearing a sales motion, not a plan.

What to Do Monday Morning

Start with the walk. Take an hour and physically follow the work. Note where WIP is piled and where people are waiting. Count what’s sitting in front of each department. That map, however rough, tells you more about your real capacity than any formula.

Then pull two trends: on-time shipment percentage and lead time, going back at least six months. If lead times are creeping and on-time performance is sliding while everyone works flat out, you don’t have a capacity shortage. You have a synchronization problem, and there’s likely far more capacity in your building than you think.

One honest caution before you start rearranging the floor. If you haven’t done this before, the pull toward localized efficiency is strong enough that you can do more harm than good. Not catastrophic harm, but you’ll keep spot-treating departments, stay unbalanced, and wonder why nothing improves. The discipline is diagnosing the real problem before applying any technique. In my experience, the vast majority of the time, the real problem is synchronization.

Final Thoughts

Capacity isn’t a number you calculate once and pin to the wall. It’s an output of how you manage flow, and it moves the moment you change how work gets released, sequenced, and protected. The manufacturer that went from four jobs a day toward sixteen never added a machine. They changed their management style, not their asset list.

Before you price out equipment, software, or a second shift, prove what your current operation can do when work flows through it properly. Simple, not easy.

That’s it for today.

See you all again next week!

Dave