When Does Productivity Demand a New Operating Model?

Productivity pressure usually produces a familiar response: find the loss, tighten the process, improve labor deployment, remove waste, automate a task, raise accountability, and ask the operation to deliver another few percentage points. Often, that is exactly what leadership should do. Continuous improvement remains one of the most reliable ways to protect margin, release capacity, and improve service without adding structural cost.

But every operating model has an economic limit. Eventually, the next increment of productivity becomes harder to capture, more expensive to sustain, or dependent on management intervention that the system should not require. At that point, another improvement program can become less rational than changing the model in which the work is performed.

Consider an illustrative multi-site operation facing a 15 percent increase in demand with little appetite for proportional growth in labor or overhead. Each site has already spent years improving standard work, scheduling, equipment utilization, quality, and frontline execution. Early gains were substantial. More recently, each additional point of productivity has required more effort, more exceptions, and more managerial attention.

One facility adds automation but shifts the constraint downstream. Another reduces labor hours but creates more overtime and supervision. A third improves local throughput while inventory accumulates elsewhere in the network. The individual metrics still move, but enterprise economics barely do. The executive question is no longer simply where the next productivity opportunity sits. It is whether the existing operating model has become the constraint.

Productivity Has a Limit Inside Any Given System

The Theory of Constraints provides a disciplined starting point. Eliyahu Goldratt's central argument is that system performance is limited by a constraint, and that improving nonconstraints can create local efficiency without materially improving the performance of the whole. The management implication is deceptively simple: focus improvement where it changes system output, not merely where improvement is easiest to measure.

Goldratt expressed that logic through five focusing steps: identify the constraint, exploit it, subordinate the rest of the system to it, elevate it, and then repeat the process as the constraint moves. The order matters. Elevating a constraint often requires investment; exploiting and subordinating it first forces management to prove that existing resources, rules, and capacity have actually been used intelligently before capital is committed.

At the operating level, the constraint may be a machine, dock, production line, labor pool, supplier, or processing step. At executive level, the harder cases are often nonphysical. The limiting factor may be decision rights, incentive design, fragmented technology, planning rules, organizational boundaries, or the way scarce resources are allocated. In other words, the constraint may sit inside the operating model—or the operating model itself may be creating it.

The Difference Between a Hard Operation and a Misdesigned One

Executives should be careful here. A difficult operation does not automatically require a new operating model. Poor maintenance, weak standard work, inadequate planning, inconsistent leadership, capability gaps, and uncontrolled variability can all create a plateau that feels structural when it is actually managerial. Redesigning around weak execution simply institutionalizes the weakness at a higher cost.

The more useful test is whether repeated improvement is still changing enterprise outcomes. When the constraint is inside the current model, stronger execution should continue to release meaningful capacity, reduce cost, improve quality, or increase service. When the model itself has become the constraint, gains increasingly migrate rather than accumulate: one function improves while another absorbs the cost, one site optimizes while the network deteriorates, or one metric rises while return on invested capital remains unchanged.

Three signals that the problem is becoming structural

The first signal is a rising coordination tax. Senior managers spend more time expediting, reconciling, escalating, and personally connecting processes that should already be connected by design. Meetings multiply because decision rights are unclear. Local teams wait for approvals because authority has not kept pace with complexity. Exceptional leaders can keep such systems functioning, but their time becomes an invisible operating expense.

The second signal is that productivity improvements increasingly transfer cost rather than remove it. Procurement reduces unit cost but raises inventory. A facility increases throughput but transportation cost rises. Labor efficiency improves while defects or customer lead times worsen. Automation reduces direct labor but adds technical support, maintenance, or exception-handling costs elsewhere. Local metrics can remain green while enterprise value creation stalls.

The third signal is declining economic return on improvement. If each additional point of productivity requires disproportionate capital, management attention, implementation risk, or recurring complexity, leadership should compare the marginal economics of further optimization with the economics of structural redesign. Productivity is not valuable because the rate moved; it is valuable because the enterprise produces more value with the resources it consumes.

The Executive Inflection Point Is Economic

This is where the conversation must move beyond operational performance and into enterprise economics. A COO should not recommend redesign because processes feel cumbersome. The stronger case is that the current model is preventing the company from converting demand, capital, technology, and talent into acceptable returns.

McKinsey's recent COO research makes this connection explicit: productivity is central to delivering strategy and ultimately to return on invested capital. Its 2026 operations work similarly describes leading companies as moving beyond isolated optimization toward rewiring end-to-end operations when new technology, capabilities, and ways of working are required to capture the next level of advantage. The implication is not that transformation is always superior. It is that executives need an economic threshold for deciding when optimization has run far enough.

A useful way to frame that threshold is to compare four questions. First, what additional enterprise value can realistically be extracted from the current model? Second, what does capturing that value require in capital, management attention, risk, and time? Third, what strategic opportunity is delayed while the organization continues optimizing around structural limitations? Fourth, would redesign create a new performance frontier rather than merely shift cost into a different part of the system?

Those questions change the conversation. A redesign that costs $20 million to unlock $3 million of annual benefit may be strategically unattractive even if the organization is inefficient. Conversely, a redesign that avoids a $100 million capacity expansion, compresses decision time, and allows the network to absorb growth may be compelling even if its direct labor savings are modest. The relevant measure is not improvement activity. It is enterprise value.

Organization Design Becomes the Supporting Question

Once leadership has evidence that the operating model is creating the constraint, the next question is what must change. Jay Galbraith's Star Model is useful because it prevents executives from treating organization design as an org-chart exercise. Galbraith's model connects five design choices: strategy, structure, processes, rewards, and people. The core principle is alignment—an organization performs better when these elements reinforce the strategy rather than contradict one another.

Suppose the strategy now requires greater speed and network flexibility, but decision authority remains highly centralized. Processes were designed for stable demand, incentives reward local utilization, technology was deployed function by function, and managers were developed primarily to control their own sites. The organization does not have five independent problems. It has one design problem expressed through five mechanisms.

That distinction matters because structural misalignment cannot be repaired indefinitely with frontline discipline. If the strategy requires cross-network optimization but leaders are rewarded for local results, better execution may actually deepen the problem. If the business requires rapid decisions but governance requires multiple approvals, faster frontline processes will not create faster enterprise response. The performance ceiling is being imposed by the rules of the system.

A Practical Example: Capacity That Already Exists

Imagine a distribution network approaching its stated capacity limit. Demand is growing, several buildings are close to saturation, service risk is increasing, and leadership is evaluating a major capital expansion. On the surface, the choice appears straightforward: improve productivity enough to delay the investment or build additional capacity.

Deeper analysis reveals that usable capacity already exists elsewhere in the network. The problem is that volume-allocation rules were created for an earlier demand pattern, labor planning remains site-specific, transportation decisions are optimized separately, commercial commitments limit flexibility, and facility leaders are rewarded primarily for their own building's performance. Technology provides excellent local visibility but weak network-level orchestration.

The physical network may not be the constraint. The operating model may be preventing the company from using the network as a system. A traditional productivity initiative could continue attacking labor rates inside the busiest sites. A capital program could add square footage or equipment. Theory of Constraints forces the more consequential question: what is actually preventing greater system throughput?

If the answer is allocation logic, incentives, decision authority, information flow, and cross-site planning, the economics change. Redesigning those mechanisms might defer capital, improve asset utilization, and create future flexibility. It may also raise transportation cost, reduce local autonomy, require new technology, and make some site-level metrics look worse. Those are not reasons to avoid redesign; they are the tradeoffs executives must price before choosing it.

When Not to Redesign

There is a strong counterargument to operating-model redesign: organizations often reach for transformation too early. New structures create new interfaces. Technology introduces learning curves. Changed responsibilities create ambiguity. New measures create unintended behaviors. Redesign can destroy institutional knowledge and consume leadership attention that might have produced a higher return through disciplined execution.

Goldratt's sequence provides a useful safeguard. Exploit the current constraint first. Subordinate the rest of the system to it. Only then elevate it. In practical terms, prove that the current system cannot economically deliver the required performance before replacing it. A company should not spend millions redesigning an operating model to solve a constraint that better maintenance, clearer standards, stronger leadership, or smarter scheduling could have removed.

But the safeguard works in both directions. Once the organization has genuinely exploited the current model, repeatedly demanding more from it can become its own form of waste. Management layers expand to coordinate exceptions, technology is added to compensate for fragmented work, capital is deployed around bottlenecks created by policy, and improvement teams spend increasing effort protecting a design that no longer matches the strategy.

The Decision

The executive challenge is therefore not choosing between continuous improvement and transformation. Both are necessary, and both can be misused. The challenge is identifying what currently limits enterprise performance and then choosing the least costly way to move that constraint.

If the constraint is waste, variability, capability, maintenance, process discipline, or a physical bottleneck, improve the operation. If the constraint is increasingly created by decision rights, incentives, organizational boundaries, information flow, technology architecture, or the allocation of resources, the operating model itself deserves scrutiny.

The decisive test is economic: Has the marginal return from further optimization fallen below the expected return from redesign, after accounting for capital, risk, disruption, and strategic opportunity? When the answer becomes yes, productivity is no longer primarily an improvement problem. It has become an operating-model decision.

Selected Sources

Goldratt Research Labs. Introduction to Theory of Constraints (TOC). Source

Galbraith Management Consultants. The Star Model and Organization Design. Source

McKinsey and Company. Productivity at the Core: How COOs Deliver Strategy. July 17, 2025. Source

McKinsey and Company. A New Operating Model for a New World. June 18, 2025. Source

McKinsey and Company. Winning the Race to Rewire in 2026: Capturing Operational Advantage. April 2, 2026. Source

Editorial standard: established frameworks, executive application, practical tradeoffs, selective sourcing.