Across decades of work with large global mining companies, one pattern stands out: They oscillate—repeatedly and expensively—between centralized and decentralized operating models.

These shifts are rarely subtle. Changes in senior leadership, periods of cost pressure, or major transformation programs often trigger decisive moves toward one end of the spectrum. Functions are pulled into corporate or commodity teams, centers of expertise are created, and assets are asked to “focus on execution.” At the other extreme, aggressive decentralization can fragment the organization, duplicating effort, stretching scarce expertise, and eroding the benefits of scale and consistency.

Although there is no single operating model that works in every context, one principle consistently holds true in mining: Value is created at the asset. Operating models perform best when capability, decision making, and accountability are aligned as closely as possible to operations, meaning where decision rights and accountability reside, not simply where the work is done.

This view is somewhat at odds with prevailing perspectives in the industry. We find that surprising. In our experience, mining companies deliver stronger outcomes when they start from a default of localization—treating assets as businesses with clear accountability for performance—and centralize only where there is a clear and compelling reason to do so.

Recent benchmarking helps explain why this matters (see Figure 1). Relative to other industries, mining company operating models appear to perform strongly in providing strategic clarity, as well as clear organizational structures and processes. However, they underperform in outcomes such as inspiration and adaptability, as shown in Figure 3 below. That pattern is telling. The challenge is not setting direction or creating structure but rather ensuring that such elements remain relevant enough to the asset to preserve ownership, responsiveness, and the ability to adapt. The further that day-to-day accountability is removed from the asset, the harder it is to sustain performance.

Start at the asset

The starting point is simple: If a supporting capability has a direct and material impact on day-to-day operational performance, it should be aligned as closely as possible to the operation, with clear accountability to the asset.

The more closely a capability is aligned to the asset, the better it addresses local constraints, trade-offs, and opportunities. Decisions are faster, accountability is clearer, and performance is easier to own. This mirrors what we see in effective private equity, where asset-intensive businesses are run as discrete units with clear decision rights, appropriate resourcing, and full accountability for returns.

This is not an argument against centers of excellence, shared services, or enterprise-level capability. It is an argument about sequencing and burden of proof. In mining, localization should be the default. Centralization should be the exception.

Centralize by exception

In our experience, there are only three defensible reasons to centralize a capability in a mining operating model (see Figure 2).

1. Scale benefits clearly outweigh the need for localization

Centralization can create value where scale effects are real and material, such as when the work is highly repeatable, transactional, or benefits disproportionately from aggregation.

However, scale alone is not sufficient justification. In mining, local context matters deeply. The more a capability depends on knowledge of the ore body, equipment, workforce, or operating rhythm, the faster scale benefits diminish. Centralization only makes sense when the value of scale clearly exceeds the cost of distance from operations.

2. Standardization or coordination that creates real value

Some activities benefit from standardization because variation erodes value. Common standards can reduce risk, improve comparability, and simplify interfaces across the organization.

The challenge is that standardization is often pursued too broadly. In many cases, variation exists for good reason, driven by differences in asset type, mining method, or system bottlenecks. Standardization should be applied selectively, where consistency adds more value than local optimization, and revisited over time rather than assumed to be permanent.

In some contexts, the benefit is not consistency but coordination across interconnected assets. We see this most often in multisite operations that share processing or logistics infrastructure. Where assets feed a common plant, rail network, port, or blending system, optimizing each site independently can suboptimize overall throughput, recovery, or realized price.

It’s often better to optimize the whole system rather than each asset on its own. The economic objective shifts from maximizing individual asset performance to maximizing system value. Capabilities such as integrated planning or system-level scheduling may need to be centralized above the asset to optimize the whole rather than the parts. Even then, the scope should be tightly defined, with operational accountability remaining firmly at the asset.

3. Expertise is scarce or constrained

In certain areas, the limiting factor may be talent. When expertise is genuinely scarce, pooling it can create more value than duplicating it across assets.

Talent constraints should inform how an operating model is implemented, not define the target state. The operating model should first be designed around what best serves the business and the asset, and only then be adjusted for the availability of capabilities, whether within the organization or in the market. Scarcity should be tested regularly; what begins as a sensible response to limited expertise can quietly become a permanent center long after the original constraint has eased.

Centralization is not binary

Deciding to centralize a capability is only part of the question. The next decision is how broadly that capability should be shared.

Capabilities can be shared within a single asset, across multiple assets, at a regional level, at a commodity level, or across an entire group. Each step up the organization increases scale and distance.

Similarity matters. The more similar the operations, the greater the potential value of shared capability. The more dissimilar they are, the more likely it is that centralization will dilute impact rather than enhance it. This is particularly relevant where organizational boundaries are drawn along commodity lines, even though operating models, bottlenecks, and technical requirements may differ materially from asset to asset. 

Implications for asset performance and leadership

Centralized operating models in mining make a clear trade-off: Scale and consistency are prioritized over local ownership. The result is often fragmented accountability, siloed optimization, and performance improvements that look positive in individual metrics but fail to translate into economic value.

This trade-off shows up in the benchmark data. As Figure 1 shows, mining companies appear relatively strong on some operating model components such as strategic clarity, organizational structure, and process discipline. But on certain operating model outcomes, including inspiration and adaptability, mining companies are weaker (see Figure 3). That pattern suggests the industry is not struggling to create order. The challenge is ensuring that operating models do not become so optimized for control and consistency that they weaken local ownership, responsiveness, and the conditions required for sustained performance at the asset.

Asset leadership roles are reshaped in the process. General managers remain accountable for outcomes but are increasingly asked to execute decisions made elsewhere, shifting the role from owning performance to managing compliance and weakening both asset performance and the leadership pipeline.

High-performing models avoid this trade-off by keeping authority close to operations, while centers remain lean and enabling, allocating capital, setting a small number of nonnegotiable standards, and building capability. Work may be delivered remotely, but accountability stays with the asset, and centers earn their place by supporting performance rather than running the business.

Implications

As mining companies face growing complexity, tighter margins, and rising execution risk, operating models matter more than ever. Although there are many viable ways to organize, the most robust starting point is clear: Value is created at the asset.

Designing operating models from the ground up—and centralizing only where there is a compelling reason to do so—strengthens accountability, improves decision making, and drives better performance. Centralization should not be the aspiration. It should be the exception, applied thoughtfully and challenged often.