Why Added Capacity Doesn't Always Produce More Output

A supplier was producing 47 parts a week. After changing the production process, it was producing more than 470.

That is the result GE Aerospace described in its 2025 letter to shareholders. The supplier made a specialized component used around turbine blades in the LEAP engine. Customer demand was much higher than the supplier's output. Forty people from GE Aerospace and the supplier examined how the work moved, identified four areas of constraint through value-stream mapping, and changed the production process.

GE attributed the improvement to changes in the production process that unlocked capacity. Its account does not report a new factory or machine purchase as the explanation for the gain. The letter does not identify the four individual constraints or calculate the effect on engine deliveries.

The example poses a question for every CEO considering another shift, production cell, warehouse or group of hires:

Before paying for more capacity, do you know what is stopping the capacity you already have from becoming output?

A second shift can be staffed. A new line can be ready. A larger warehouse can be open. Yet customer orders still wait, work accumulates before one stage, and completed units may require correction or approval before they can be delivered.

Sometimes the demand simply did not arrive. Sometimes a supplier, production stage or required inspection limits the system. Sometimes work has been produced but has not become good, accepted output. These conditions require different decisions.

I distinguish three achievements: installed capacity, usable capacity and delivered output. The question is not how much activity a new resource creates. It is how much additional good, accepted output the whole company can deliver.

Busy is not the same as delivered.

This research examines four public-company situations to test where the conversion can stop. A documented constraint is not automatically Structural Drag. The operating arrangement becomes relevant only when evidence shows avoidable work that a different arrangement could remove without reducing what customers receive or what the business must protect.

1. The Capacity-Landing Map: installed, usable, delivered

In my Business Structural Drag framework, Structural Drag is work a company repeats because of how it is arranged—work a different arrangement could remove without changing what the customer receives or what the business must protect.

The first distinction in a capacity investment is between resources available, output those resources can reliably produce, and output the whole company can deliver.

Imagine one machine can produce 100 additional parts a day, but the next operation can inspect only 60. Installing the machine may increase work in process without increasing completed units. If inspection is a required quality safeguard, eliminating it is not a legitimate answer. Increasing inspection capability, improving upstream first-pass quality, or changing how work reaches inspection might be.

A new machine in front of a bottleneck buys a bigger queue.

This is an illustration, not a calculation about any company in this paper.

The limiting stage may also move after an improvement. A successful change at one step is not proof that the enterprise can now deliver the same increase. Follow the same unit of work through the entire path.

2. Three explanations for capacity that does not land

A disappointing capacity result needs three competing explanations before anyone proposes a redesign.

A supplier shortage can prevent a factory from producing, without the buyer's structure having caused the shortage. Quality inspections may delay release because the work genuinely needs inspection. Spare capacity may be economically sensible because demand is seasonal or because the business must protect service levels.

The structural question comes later: does the arrangement of commitments, decisions, handoffs, sequencing or correction work create avoidable repetition around the limit? If the answer is no, the right response may be to add capacity, renegotiate supply, invest in quality, wait for demand, or deliberately change nothing.

3. Four public-company situations that reveal different limits

I examined GE Aerospace, Rivian, Boeing and GlobalFoundries. The examples are deliberately different: a measured process constraint that was improved, a component shortage that later ceased to constrain production, aircraft requiring rework or delayed by unavailable inputs, and facilities operating below rated capacity because of demand conditions.

I have not assessed these companies internally or assigned them Structural Drag profiles. Reported events belong to the companies. The structural questions are my interpretations. No company-wide uplift or recoverable Structural Drag amount is inferred from these disclosures.

4. What the four situations establish—and do not establish

GE Aerospace shows that a measured process constraint can be improved. Rivian shows a documented input constraint entering and then leaving production. Boeing shows why built units may not yet qualify as delivered output. GlobalFoundries shows why spare capacity may reflect demand, not an internal bottleneck.

They are mechanism illustrations, not a sample of four companies that all added resources and failed. They also differ in scale, products, reporting units and time periods; their numbers should not be compared as if they were the same productivity metric.

The shared executive question is:

What limited the additional good output—and would that limit remain if the work were arranged differently?

The mechanisms do not depend on company size: a second shift, a new production cell or a larger warehouse can meet the same three explanations. The first clause locates the constraint; the second tests whether the arrangement creates removable work.

5. Where Structural Drag could enter

The six Structural Drag profiles suggest possible places to investigate, not diagnoses of the named companies.

No Structural Drag established remains a legitimate result. Demand, essential safety and quality controls, scarce supplier inputs, mandated qualifications, temporary ramp-up or deliberate resilience may be sufficient explanations.

The presence of a bottleneck is an observation. Structural Drag is a hypothesis that must survive a competing explanation.

For the full profile illustrations and diagnostic distinctions, see Research 01 — Why Growing Companies Become More Expensive to Run.

6. The economics: do not count capacity as contribution

Three figures can be confused when executives review a capital or headcount decision.

Installed potential. This is the promised rate from a new machine, shift, building, software implementation or team. It may be useful for planning, but it is not revenue, delivered output or profit.

Real operating improvement. A yield gain, shorter queue, higher good-unit rate or less paid rework can create usable capacity. That gain still needs the rest of the operation—and a paying customer—to convert into economic value.

Verified financial effect. Additional contribution comes from incremental saleable output actually sold at an appropriate margin, net of incremental operating costs and any displacement. A recurring cost reduction is a separate outcome. Released hours are not automatically EBITDA savings unless expenses actually fall; released capacity redeployed to profitable output must not be counted twice.

My three established economic ledgers remain separate: verified financial effects, paid capacity consumed and contribution not earned.

A useful management bridge is:

Incremental installed rate → incremental good-unit capacity at the constrained path → incremental delivered and accepted units → incremental contribution → cash.

Every arrow requires evidence. Management should also compare the result against the invested capital, working-capital requirement, time to achieve the gain and continuing maintenance or quality obligations.

There can be a valid return even without higher output: better quality, lower risk, redundancy, reduced overtime or faster response. But that is a different investment thesis. It should be stated and verified as such, rather than converted retrospectively into an unsupported throughput claim.

The connection to the Renergy™ Conversion Chain remains:

Paid capacity → Usable capacity → Delivered output → Operating profit → Cash → Enterprise value

7. Give this to your COO, CFO and operating leaders

The Capacity-Landing Test

Whitepaper 02 established this five-question test for AI. The same five questions apply to any investment meant to add capacity—people, equipment, systems or facilities. Run it with the leaders who own the work before and after the investment.

1. Where will the new capacity go? Name the route: more deliverable volume, better price and mix, quality, work brought in-house, or growth without the next hire. More room to breathe is not a route.

2. Who receives the output — and how much can they receive? Name the next stage downstream and the rate at which it can take more work. If nobody can answer, the investment is ahead of the company.

3. What fires automatically at the new volume? Promises, material, approvals, schedules. Find the commitments that will keep firing against assumptions the new capacity is about to change.

4. Who decides the exceptions? More volume brings more hard cases. Name who receives them, and the volume at which that person becomes the queue.

5. Where will you measure the result — and what else could explain it? Set a baseline before the investment and a measure where the output lands. Then name at least one competing explanation and test it: demand, mix, supplier inputs, required quality controls, capacity held in reserve on purpose, or a return not yet due.

The test can return four answers: scale the investment; redesign the handoff first; add receiving capacity or choose a different route, possibly price rather than volume; or change nothing. Each is a legitimate result.

Applying the test to one investment. Define the output promised when the decision was approved. Trace installed resources to usable capability and then to good, accepted deliveries. Examine demand, supplier inputs, yield, queues, handoffs and required controls as competing explanations. Calculate delivered output and economic contribution separately from theoretical rate and released hours. This is supporting evidence for the five questions—not a second version of the test.

The purpose of the test is not to make the new asset busy. It is to find out what limits useful output.

8. What would change my conclusion?

I would weaken a Structural Drag hypothesis if demand was insufficient; if a supplier outage or temporary technical defect fully explained the gap; if mandatory inspection or qualification was the genuine binding requirement; or if comparable work reached customers without repeat corrections after stabilization.

I would strengthen the case for structural examination if completed upstream work repeatedly arrived downstream unusable, if priority rules generated predictable rescheduling and returns, if the same exceptions waited for unnecessarily concentrated authority, or if additional resources increased work-in-process without increasing good, accepted output because of recurring handoff conditions.

A growing queue alone is not proof. An expensive investment alone is not proof. A higher equipment-utilization target is not proof.

The decisive evidence is what actually constrained the good unit—and whether a different arrangement could remove the nonessential work without sacrificing what the company must protect.

9. Before you buy the next unit of capacity

Hiring another ten people may be the correct decision. So may an additional machine, a better supplier contract, a second inspection team or a redundant facility.

But if the added resources enter a system that cannot convert them into accepted output, the company may acquire more expense before it acquires more value.

And if demand is not there, a faster operation cannot manufacture a profitable customer.

Before buying more capacity, establish what actually limits the output you can deliver.

Research 01 examined growth and operating leverage. Research 02 examined technology completion versus operating and economic conversion. Research 03 examined why growing revenue and reported profit need not become cash. Here the question is what happens between paying for capability and producing more good, delivered output.

Gower Idrees

Creator of the Business Structural Drag framework and the Renergy™ management framework

Sources, Evidence Notes & Research Methodology

Evidence reviewed through November 2, 2026