How much infrastructure will the business actually need?

How Much Infrastructure Will the Business Actually Need?

A demand forecast does not produce one correct capacity number. Infrastructure decisions combine uncertain volume, long lead times and asymmetric errors: too little capacity can lose service and growth, while too much locks capital into assets that may be difficult to repurpose. The right answer is a capacity strategy that changes as evidence arrives.

Model demand as a distribution by customer, location, peak and time�not a compound growth line. Identify what drives each range and how quickly demand can move. Translate it into effective capacity after uptime, yield, maintenance, seasonality and network constraints. Nameplate output overstates what can be promised at required reliability.

Build a capacity ladder before selecting a large asset: improve yield and scheduling, shape demand, use inventory, reserve external capacity, lease, add modular units and then commit permanent infrastructure. Price each step by lead time, marginal cost, quality, control, reversibility and the volume at which it becomes superior to the next alternative.

HM Treasury�s 2026 Green Book recommends scenario analysis, decision trees and real-options analysis where uncertainty is significant or investment is hard to reverse. The principle is practical: preserve the right to expand, contract or switch later, but pay for flexibility only when future information can change the decision before the option expires.

Create a capacity roadmap with ranges, trigger metrics, decision dates, permitting and supplier lead times. Stress correlated growth and disruption, because the same infrastructure may support both normal demand and resilience. Enough capacity is not the peak of the optimistic forecast. It is the staged combination that protects service while keeping the economic cost of being wrong within risk appetite.