Volume and Capacity Forecasting

Start with the capacity decision

A forecast should answer a decision that management can take. A depot manager may need next week's shift plan. A transport operator may need a three-month vehicle contract. An airport or terminal may need several years to add infrastructure. One forecast horizon cannot serve all three decisions equally well.

Marketways defines the resource, location, time interval, planning horizon and decision lead time before selecting a model. We also identify the consequences of shortage and surplus. This definition determines the forecast detail and uncertainty that management needs.

Forecast demand where capacity is consumed

A network-wide forecast can hide opposing local movements. Growth on one lane does not release a vehicle, gate or handler at another node. Even a correct national total can therefore support the wrong local plan.

Marketways forecasts demand by the operational units that consume capacity. Depending on the network, these units may include origin, destination, route, depot, terminal, customer, service class and departure window. The analysis retains the wider network totals so that local plans remain connected to the organisation's overall demand.

Reconcile forecasts across the network

Different planning teams may forecast the network total, each region and individual lanes separately. These estimates can disagree because each model uses different information and error. If the figures do not reconcile, management cannot build one capacity plan from them.

Hierarchical and grouped forecasting methods use information from several levels and then reconcile the results. The network total equals the regions, and the regions equal their constituent lanes or nodes. Reconciliation preserves local information without allowing contradictory totals.

Convert demand into workload

Shipment, parcel or passenger counts do not consume capacity uniformly. A pallet may occupy more cube than another pallet. A delivery with several stops uses more driver time than a direct movement. Passengers with checked baggage, transfers or assistance needs place different demands on airport processes.

Marketways converts demand into the workload relevant to each constraint. Measures may include weight, cube, stops, handling moves, dwell time, service minutes, baggage items or connection demand. This translation shows whether the limiting resource is vehicle space, labour, equipment, processing rate or physical storage.

Represent the timing of demand

A daily or weekly total can remain stable while the arrival pattern becomes more concentrated. Average capacity may appear adequate even though a short peak creates queues, missed departures or overtime. The model must therefore reflect the interval in which the operation can actually respond.

Marketways analyses intraday, daily and seasonal patterns at the frequency relevant to the constraint. We distinguish an increase in total demand from a change in arrival timing. Where queuing or processing interactions matter, simulation can test whether resources clear the peak before the next wave arrives.

Separate recurring patterns from structural changes

Historical patterns help when the operating environment remains comparable. However, a new customer contract, route launch, facility opening, public event, trade change or service withdrawal can shift demand beyond the established pattern. A model that treats the change as ordinary seasonality will adjust too late.

Marketways records known interventions and tests whether they created a temporary spike, a permanent level change or a different trend. Confirmed future changes enter the forecast directly. Uncertain changes become explicit scenarios with stated assumptions rather than hidden judgemental adjustments.

Use forward information without confusing intent with demand

Bookings, manifests, tenders, reservations and customer schedules can provide early evidence of future demand. However, the relationship between an early signal and completed demand may change by customer, lane, booking window or cancellation policy. The latest signal is not automatically the best forecast.

Marketways measures how each forward indicator matures before the operating date. The model uses only information available at the forecast origin and tracks revisions as new evidence arrives. This approach prevents later manifests or completed movements from leaking into an earlier forecast test.

Treat sparse lanes differently

Some routes carry demand in most periods. Others receive irregular orders separated by many zero-volume periods. Applying the same model to both patterns can create unstable local forecasts or allocate capacity to movements that rarely occur.

Marketways identifies intermittent series and tests whether they should be forecast separately, pooled with comparable lanes or represented through the probability and size of an arrival. The capacity decision may also justify a shared or flexible resource instead of dedicated capacity for each sparse lane.

Quantify the uncertainty that changes the decision

A point forecast gives one expected value. It does not show the plausible range around that value. Capacity planned exactly to the point forecast can fail whenever actual demand lands above it, while planning to an extreme case can leave expensive resources idle.

Marketways produces forecast distributions or prediction intervals and tests whether their historical coverage is credible. The selected planning threshold reflects the unequal cost of shortage and surplus. A safety-critical or time-critical movement may justify more protection than a demand stream that can be deferred or reassigned.

Match each response to its lead time

Management can respond to a short-term peak through overtime, rescheduling, subcontracting or temporary equipment. Recruitment, vehicle leases and network redesign require more time. Infrastructure investment may require several planning cycles.

Marketways aligns each forecast horizon with the actions still available at that point. The model shows when the organisation must commit, what later flexibility remains and which evidence should trigger the decision. Forecasting then supports an executable response rather than reporting a risk after the response window has closed.

Model capacity as a constrained system

Adding one resource does not always increase throughput. Extra vehicles may wait if a gate, sortation line, loader or receiving dock remains constrained. Capacity can also rise in steps because a full shift, vehicle or facility module must be added at once.

Marketways represents shared resources, dependencies, minimum operating units and step costs. Optimisation or simulation can show where extra capacity changes service and where it merely moves the bottleneck. This prevents the organisation from purchasing capacity that another constraint leaves unused.

Compare flexible and committed capacity

Permanent capacity may offer a lower unit cost when demand is stable and sufficiently high. Flexible capacity can protect the organisation when demand remains uncertain, although subcontracting, leasing or overtime may cost more per movement.

Marketways compares fixed, flexible and staged capacity under the same demand scenarios. The analysis includes commitment dates, cancellation terms, supplier availability and the operational consequences of each option. Management can then judge when flexibility is worth its premium and when repeated temporary measures justify a longer-term commitment.

Test scenarios and operating triggers

A scenario becomes useful when it changes a decision. Generic labels such as optimistic and pessimistic do not identify what happened or what management should do differently.

Marketways defines scenarios through specific assumptions such as a customer win, lane closure, delayed fleet delivery, seasonal surge or altered service promise. Each scenario produces a capacity requirement, service consequence and response. Trigger thresholds then connect new evidence with the approved action.

Use AI to update evidence within defined authority

An AI-assisted workflow can collect bookings, manifests, staffing, vehicle availability and external events, refresh the forecast and prepare an exception for review. It can also explain which assumptions or signals changed since the previous planning cycle.

The workflow should not purchase capacity or alter a service commitment merely because one forecast crossed a threshold. Marketways tests missing feeds, duplicate orders, delayed updates, contradictory sources and failed tool calls. The organisation defines who may approve overtime, external capacity, customer changes or capital expenditure.

Evaluate the forecast through the decisions it supports

A lower average forecast error does not guarantee a better capacity plan. A model can perform well across ordinary periods and still miss the peaks that cause service failure. Another model may produce a slightly larger average error but improve staffing, vehicle or gate decisions.

Marketways tests forecasts at rolling historical origins and at the horizons used by management. We assess point accuracy, interval calibration, peak performance and bias by lane or node. We also measure capacity shortages, idle capacity, overtime, service reliability and avoidable external hire. These results show whether the forecasting process improved the operating decision.

How Marketways connects demand with capacity

Marketways combines Sales Forecasting and Demand Planning with Workforce Planning and Capacity, hierarchical forecasting, scenario analysis, optimisation and simulation. We define the decision, prepare the demand and capacity evidence, quantify uncertainty and connect each forecast horizon with an approved response.

The result is a capacity plan that identifies which resource may become constrained, where and when the constraint may arise, how much uncertainty remains and what management can do before service is affected.

Related guides and evidence

• Network, Depot and Warehouse Design
• Fleet and Route Optimisation
• Sales and Operations Planning
• Disruption and Corridor Resilience
• Fleet Reliability and Replacement
• Fleet Reliability, Capacity and Disruption Planning: case study