One Sustainability Policy Will Not Fit Every Operating Context

Enterprise consistency should govern the outcome and the decision rules, not force every site, region or business unit to use the same intervention.

Uniformity is attractive to large organisations. One target, one policy, one scorecard and one implementation standard can look like disciplined governance.

But a consistent enterprise objective does not imply that identical local action is optimal.

Different sites operate with different asset ages, energy mixes, product portfolios, customer demand, infrastructure, workforce capability, climate conditions and regulatory environments. An intervention with high leverage in one location can be marginal in another. A target that is easy for one business unit can require disproportionate capital in another.

The strategic challenge is to preserve enterprise direction without confusing standardisation with effectiveness.

The Strategic Context

Xu and colleagues' 2017 geographical analysis of manufacturing CO2 emissions in China provides a clear empirical example. The researchers used geographically weighted regression because a conventional global model would produce average relationships and obscure local variation.

Their results showed that the relationship between major drivers and manufacturing emissions varied across regions. Economic growth had a positive effect across the analysis, but the estimated strength declined from eastern to central and western regions. Urbanisation showed a stronger estimated effect in the western region, while energy efficiency had stronger effects in eastern and central regions than in the west. Industrialisation also varied spatially.

The policy implication in the paper is differentiated regional action.

For enterprise leaders, the broader lesson is more important than the historical Chinese context:

When causal drivers differ, identical interventions can be inefficient even when the strategic objective is common.

Other sources in the batch reinforce the same pattern.

Lee, Han and Wang found landfill greenhouse-gas estimates to be highly sensitive to waste composition, climate, gas collection and methane oxidation assumptions. A waste-management strategy therefore cannot be evaluated independently of local landfill conditions.

Li and colleagues' aggregated HVAC modelling also shows context sensitivity. Building thermal behaviour, thermostat dead-bands, outdoor conditions and baseline settings affect aggregate load response.

The conclusion is not that enterprise standards are undesirable. It is that standards should be applied at the right level.

What Leaders Commonly Misread

The first mistake is equating fairness with identical targets.

Suppose a group operates five factories and imposes the same 20% energy-reduction target on each. The rule is equal. The effort may not be.

One plant may have recently installed efficient equipment and already captured most low-cost opportunities. Another may have old compressed-air systems, poor controls and significant waste. A third may have expanded production while the others are stable.

An identical percentage can therefore produce weak capital allocation.

The second mistake is centralising the intervention rather than the outcome.

Head office may mandate one technology because it worked in a pilot site. Yet the real advantage may have come from local conditions that are absent elsewhere.

The third mistake is allowing local variation to become an excuse for weak accountability.

Context matters, but "our site is different" cannot become a permanent defence against performance improvement. Differentiation needs evidence.

Reframing the Issue

The better operating model separates four levels of standardisation.

Enterprise ambition

The organisation defines what outcome matters and why. This should be common.

Decision principles

The organisation defines how options will be evaluated, including lifecycle value, risk, evidence, payback, environmental effect and strategic fit. These should also be largely common.

Local pathway

Sites choose interventions suited to their causal drivers and constraints. This should vary where evidence supports variation.

Assurance

The organisation verifies that local pathways remain credible and comparable. This should again be common.

The result is controlled differentiation.

That is different from both extreme centralisation and unmanaged local autonomy.

From Average Performance to Local Causation

Portfolio data is often aggregated because leaders need a manageable view. The danger is that averages can conceal both opportunity and risk.

A corporate energy-intensity average can improve while one large site deteriorates. A regional emissions average can hide a location where urbanisation or product growth is overwhelming efficiency gains. A common waste metric can hide the fact that one geography has highly effective recovery infrastructure while another sends the same material to landfill.

Executives therefore need two views simultaneously:

the enterprise outcome, which supports accountability;
the local causal model, which supports intervention.

The first tells leadership where performance is going.

The second tells management what to do about it.

Portfolio Allocation Should Follow Abatement Opportunity

A differentiated strategy can improve capital efficiency.

Imagine three hypothetical plants with the same annual emissions.

Plant A can reduce emissions materially through low-cost control-system changes.

Plant B requires a moderate equipment retrofit.

Plant C would need premature replacement of a major asset with significant stranded value.

If the organisation allocates equal capital to all three, it may spend too much at Plant C while leaving cheap improvement at Plant A unfunded.

Portfolio leadership should instead ask where each additional dollar, engineering hour and unit of disruption creates the greatest strategic return, subject to minimum standards and risk constraints.

This does not mean concentrating all effort where it is cheapest forever. Some high-cost sites may eventually require transformation. It means sequencing investments intelligently.

Local Conditions Can Change the Counterfactual

The landfill study adds another important dimension.

The value of diverting a waste stream depends partly on what happens to that waste locally if it is not diverted. Collection efficiency, decay behaviour, climate and gas-management practices influence the landfill counterfactual.

The same principle applies across enterprise strategy.

Renewable electricity may be highly valuable at a site dependent on carbon-intensive power and less material at another already supplied by low-emission generation. Water-saving investment may be strategically urgent in a water-constrained region but primarily cost-driven elsewhere. On-site storage may create resilience value in one grid and little in another.

A corporate initiative therefore needs a context map, not merely a rollout plan.

Decision Framework

ERANORTH recommends a Common Objective, Differentiated Pathway model.

1. Set the enterprise outcome

Define the common objective in measurable terms. Avoid prescribing the solution at this stage.

2. Segment operating contexts

Group sites or business units by the factors that materially influence the outcome: asset age, energy mix, climate, demand growth, regulation, product type, infrastructure, capability and supply-chain configuration.

3. Identify local causal drivers

Use site-level evidence to determine what is actually producing the result. Do not assume corporate averages explain local performance.

4. Build an intervention library

Define multiple acceptable pathways with evidence requirements. These might include control changes, efficiency upgrades, fuel switching, process redesign, procurement, behavioural interventions or replacement.

5. Allocate capital by value and necessity

Fund high-leverage opportunities while preserving a minimum standard across the portfolio. Explicitly recognise opportunity cost.

6. Normalise performance where needed

Adjust comparisons for production, climate, product mix or other factors so sites are not rewarded or penalised for conditions they do not control.

7. Rebalance periodically

Context changes. A site that was low priority can become high priority after demand growth, regulation, asset degradation or energy-market changes.

Governance Without Bureaucracy

A differentiated operating model needs stronger governance than a uniform mandate, not weaker governance.

Head office should define:

mandatory outcomes;
minimum compliance thresholds;
approved measurement methods;
evidence standards;
capital-allocation principles;
escalation rules.

Local leaders should own:

causal diagnosis;
intervention choice;
implementation sequencing;
operational adoption;
benefits evidence.

This creates clear decision rights while preserving local intelligence.

From Strategy to Execution

Immediate action: identify one enterprise target currently being implemented uniformly and test whether the same causal driver exists across every site.

Medium-term capability building: improve local data quality, normalisation methods and site-level modelling. Build portfolio dashboards that show both absolute outcome and contextual drivers.

Long-term strategic positioning: move from campaign-style sustainability programs to a portfolio operating model in which capital and interventions are continuously reallocated as local conditions change.

Signals to Monitor

Watch for identical targets producing radically different marginal costs; repeated requests for exemptions from central programs; pilot solutions underperforming after rollout; sites with similar output but very different emissions or resource intensity; corporate averages improving while a minority of sites deteriorate materially; local managers unable to explain the causal drivers behind their targets; and capital being spread evenly rather than allocated according to evidence.

A useful signal is the cost and difficulty of the next unit of improvement at each site. Large differences suggest the portfolio should not be managed as though every location is on the same curve.

Questions for the Leadership Team

Which enterprise objectives genuinely require identical local action, and which only require a common outcome?
What local variables materially change the economics or effectiveness of our current sustainability initiatives?
Are we comparing sites fairly after accounting for product mix, climate, asset condition and energy context?
Where are low-cost opportunities being underfunded because capital is distributed uniformly?
Which sites face structural constraints that require longer-term transformation rather than annual efficiency targets?
How do we prevent legitimate differentiation from becoming weak accountability?
What evidence would cause us to change the priority of a site or region?

Closing Perspective

Enterprise strategy needs coherence, but coherence is not sameness.

The most mature organisations standardise purpose, evidence and governance while allowing execution to adapt to the system in front of them.

When context changes causation, leadership should expect the pathway to change as well.

The question is not whether every site is doing the same thing. It is whether every site is making a defensible contribution to the same strategic outcome.

Source basis: This article is an original ERANORTH synthesis principally informed by Xu et al. (2017), Geographical analysis of CO2 emissions in China's manufacturing industry: A geographically weighted regression model; Lee, Han and Wang (2017), Evaluation of landfill gas emissions from municipal solid waste landfills for the life-cycle analysis of waste-to-energy pathways; and Li et al. (2017), Effective power management modeling of aggregated heating, ventilation, and air conditioning loads with lazy state switching, all published in Journal of Cleaner Production, volume 166.

About EraNorth Insights
EraNorth Insights publishes practical analysis on strategy, projects, operations, transformation and decision intelligence for professional and organisational use. About EraNorth.