Capacity is not only a property of equipment; it is also a property of how an organisation schedules, sequences and controls the equipment it already owns.
A familiar response to operational pressure is to ask for more capacity. Another pump, another line, another treatment train, another server cluster or another building can appear to be the most concrete way to remove a bottleneck.
Sometimes that is exactly the right decision. But capital expansion is also one of the most expensive ways to discover that the real problem was not physical capacity at all.
Existing assets can underperform because of poor sequencing, simultaneous demand peaks, rigid control rules, unnecessary waiting, weak coordination between operating layers or control settings that were never designed for the current objective. If those conditions are not understood, new capital can simply enlarge an inefficient system.
The executive question should therefore come earlier than the investment request:
Have we exhausted the value available from changing the operating logic of the system we already have?
The Strategic Context
Four 2017 studies from very different technical domains point toward the same management principle.
Li and colleagues modelled large populations of heating, ventilation and air-conditioning units. Their proposed "lazy state switching" approach did not require every unit to react immediately to a central set-point change. Instead, units could delay switching according to their operating condition. In simulations, aggregate power peaks during transitions were reduced to less than half those produced by immediate switching, while the authors reported no sacrifice in end-use performance.
Wu, Li and Qu addressed crude-oil transportation inside a refinery. Rather than redesign the pipeline network, they changed short-term scheduling and flow-rate choices. Their industrial case reported a 16.7% reduction in the study's energy objective compared with the reference schedule. The important feature was not the percentage itself. It was that the improvement came from changing when and how existing pumping capacity was used.
Liu and Wang modified the operating pattern of an established wastewater-treatment process through intermittent aeration, internal recycle logic and a small polishing zone. Their pilot study reported strong nutrient removal under the tested conditions and around 10% lower aeration energy than the comparison MLE operation.
Peng and colleagues, modelling arsenic and iron removal in a granular biofilm system, found that hydraulic retention time materially improved removal over an initial range, but further increases eventually produced little additional benefit. That is a reminder that adding more residence time, volume or equipment can become economically pointless once the governing response has saturated.
The technologies are different. The strategic pattern is consistent:
System performance can change materially when timing, sequencing, interaction and control rules change, even when the underlying asset base remains largely unchanged.
What Leaders Commonly Misread
The first mistake is equating utilisation with productivity.
An asset running harder is not necessarily producing more enterprise value. A pump operated at the maximum available rate may consume disproportionately more energy. A fleet of machines all responding simultaneously may create a system peak that none would create individually. A treatment process operated continuously may use more energy while denying the system a condition required for another biological function.
The second mistake is assuming the bottleneck is physical because the symptom appears physical.
Queues, missed output, excessive energy, high overtime or delayed customer response can all look like insufficient capacity. Yet the underlying cause may be variability, batch policy, scheduling, changeovers, poor prioritisation, maintenance windows, information latency or local rules that conflict with system performance.
The third mistake is asking operating teams to "optimise" without defining the enterprise objective.
A refinery schedule can optimise throughput, inventory, energy, tank usage or changeovers. Those objectives are not identical. A building-control system can optimise comfort, cost, power quality or demand flexibility. A wastewater process can optimise discharge quality, energy, chemicals, footprint or robustness. The system cannot know which trade-off matters most unless leadership makes that choice explicit.
Reframing the Issue
Capacity planning should be treated as a sequence of decisions, not a single engineering calculation.
The first question is whether the existing system is operating near its economically useful frontier.
Only after that should leadership ask whether the frontier itself is too low.
This distinction matters because the two problems require different capital responses.
If the asset base is capable but poorly controlled, the intervention may be software, instrumentation, scheduling, operating rules, maintenance discipline or capability development.
If the system is already well controlled and still cannot meet demand, then physical expansion becomes easier to justify.
A useful executive distinction is:
- physical capacity: what the installed system can theoretically process;
- effective capacity: what it can reliably deliver under real operating constraints;
- economic capacity: what it should deliver once energy, quality, maintenance, labour and risk are considered;
- strategic capacity: what the organisation needs to support its future position.
A business can have surplus physical capacity and still lack effective capacity. It can also have adequate effective capacity today but insufficient strategic capacity for the future.
Those are different investment problems.
Control Logic Is an Enterprise Asset
Control logic is often treated as an engineering detail. At scale, it becomes part of the business model.
Consider a hypothetical multi-site manufacturer facing a demand increase. One plant requests a new compressor because pressure falls during peak production. Investigation shows that several high-demand devices start together after scheduled breaks. A revised start-up sequence could flatten the peak. If that change restores pressure while maintaining output, the organisation has effectively released capacity without buying another compressor.
The same reasoning applies in digital systems. A company experiencing infrastructure peaks might add compute capacity. But if workloads can be queued, cached, shifted or prioritised, the more valuable intervention may be orchestration rather than hardware.
In both cases the asset is not merely the machine. It is the combination of machine, information, rules and timing.
That is why operating policy deserves the same governance attention as physical configuration.
The Portfolio Cost of Premature Capacity Expansion
New capacity competes for capital with every other strategic initiative.
It also creates future obligations: maintenance, spares, training, energy, software, insurance, inspection, property, support contracts and eventual replacement.
An expansion project that could have been avoided through control-system redesign therefore imposes two costs.
The first is the direct cost of the unnecessary asset.
The second is the opportunity cost of capital and organisational capacity that could have been used elsewhere.
Portfolio governance should therefore require major capacity proposals to show what has been done to test operational alternatives. This does not mean forcing teams through artificial "efficiency programs" when capacity is genuinely exhausted. It means making the logic visible.
If the case for expansion is strong, it should survive that challenge.
Decision Framework
ERANORTH recommends a six-stage Control Before Capacity review.
1. Define the outcome
State the requirement in business terms. Do not begin with the proposed equipment.
Examples include required throughput, discharge quality, service level, energy limit, resilience requirement or demand-growth scenario.
2. Map the operating constraint
Identify what actually prevents the outcome. Use data from the system rather than averages alone.
Look for peaks, idle periods, blocking, starvation, waiting, simultaneous starts, changeovers, control dead-bands, sequencing rules, maintenance constraints and information delays.
3. Separate controllable from structural limits
Some constraints can be changed through scheduling or policy. Others are fixed by physics, safety, regulation, geometry or material capability.
This prevents teams from wasting time "optimising" an immutable limit.
4. Test operating alternatives
Model or pilot changes to sequencing, dispatch rules, set-points, flow rates, batch sizes, priority logic, recycle ratios, run windows or demand shaping.
The test should evaluate more than output. It should include energy, quality, asset wear, safety, reliability and customer effect.
5. Identify the saturation point
Determine where additional operational effort stops creating material value.
This is critical. Optimisation can itself become waste if leadership keeps pushing a variable beyond the region where the system responds.
6. Rebuild the capital case
Only then compare:
- no investment;
- control or scheduling redesign;
- minor debottlenecking;
- modular capacity addition;
- major expansion;
- replacement with a different technology.
The decision should be based on lifecycle value, not simply installed cost.
From Strategy to Execution
Immediate action: require material capacity requests to include an operating-logic assessment. Ask what evidence shows the constraint is physical rather than procedural, informational or control-related.
Medium-term capability building: strengthen data capture, process modelling, scheduling capability and cross-functional operations engineering. Many organisations have rich equipment data but weak capability to turn it into better operating policies.
Long-term strategic positioning: design future assets to be controllable, observable and adaptable. Flexible systems often create more strategic value than assets optimised only for one forecast operating point.
Related article: Technology Is a Means, Not a Strategy
Related article: The Best Decarbonisation Technology May Be the Wrong Next Investment
Signals to Monitor
Watch for capacity requests triggered by recurring peaks rather than sustained load; assets with high nameplate utilisation but significant waiting elsewhere in the process; frequent simultaneous start-stop behaviour; rising energy per unit at high throughput; widening differences between sites using similar equipment; manual schedules that depend on a few experienced planners; control parameters that have not been reviewed after process changes; and expansions proposed before the current constraint has been measured.
A particularly useful signal is the gap between peak demand and average demand. Large gaps often indicate that timing and orchestration deserve investigation before capacity is added.
Questions for the Leadership Team
- What evidence proves that our current constraint is physical capacity rather than scheduling, variability or control logic?
- Which operating objective are we actually optimising: throughput, cost, energy, service, resilience or a defined combination?
- What capacity could be released by changing sequencing, set-points, dispatch rules or coordination between assets?
- Where does the system reach a saturation point beyond which additional operating intensity produces little value?
- What lifecycle obligations would a new asset add to the portfolio?
- If we defer the expansion and improve control first, what strategic risk do we accept?
- If we approve the expansion, what evidence would later tell us that we invested too early?
Closing Perspective
Capital is often easier to authorise than operating discipline is to build. A new asset has a price, a project plan and a visible completion date. Better control logic is less tangible, but it can be equally strategic.
The objective is not to avoid capital expenditure. It is to make sure capital is used to solve a constraint that operating logic cannot economically solve.
The strongest capacity decision is therefore not "build" or "do not build". It is a sequence:
understand the system, change the controllable logic, identify the remaining constraint, then invest with evidence.
Source basis: This article is an original ERANORTH synthesis principally informed by Li et al. (2017), Effective power management modeling of aggregated heating, ventilation, and air conditioning loads with lazy state switching; Wu, Li and Qu (2017), Energy efficiency optimization in scheduling crude oil operations of refinery based on linear programming; Liu and Wang (2017), Enhanced removal of total nitrogen and total phosphorus by applying intermittent aeration to the Modified Ludzack-Ettinger (MLE) process; and Peng et al. (2017), Enhancing immobilization of arsenic in groundwater: A model-based evaluation, 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.