Yield, Throughput and Bottleneck Diagnostic

Define the output the factory must deliver

A bottleneck exists in relation to a required rate of output over a defined period. A production line may have enough weekly capacity but still miss a daily dispatch window. A factory may produce the required number of units but fall short after scrap, rework and customer specifications reduce the quantity that can be sold.

Marketways therefore begins with the demand that the production system must satisfy. We define the products, volumes, due dates, quality requirements and operating horizon included in the decision. This definition prevents management from improving gross production while customer orders remain late or finished-goods availability remains unchanged.

The product mix is essential. Two products may use the same equipment for different lengths of time, follow different routes or require different changeovers. Consequently, the resource that limits one production plan may have spare capacity under another. Management needs a diagnosis that reflects the mix it intends to produce rather than an average unit that the factory never actually makes.

A queue identifies congestion, not necessarily its cause

Material accumulates before a process when arrivals exceed departures. The process may lack capacity. However, a long queue can also arise because planners release work too early, batches are too large, a downstream station blocks the flow or defective units return for rework. Increasing the apparent bottleneck's speed will not resolve these other causes.

Marketways examines the states of adjacent activities as well as the queue itself. A machine may wait because the previous process did not supply material. It may stop because the next buffer is full. It may remain busy for long periods because repeated rework consumes capacity. These states have different causes and require different responses.

We also distinguish persistent constraints from temporary disruption. One breakdown can create a queue that takes several shifts to clear. By contrast, repeated congestion under comparable demand may indicate a structural capacity or scheduling problem. Management should not approve permanent capital expenditure to correct an isolated event, and it should not treat a recurring constraint as random variation.

Reconstruct the route from demand to saleable output

Production records often sit in separate planning, manufacturing, maintenance, quality and inventory systems. Their timestamps may use different definitions. For example, one system may record when an order enters a workstation, while another records only when the completed batch is reported. Marketways reconciles these definitions before calculating waiting, processing or loss.

The analysis follows each product through its permitted route. We connect order release, queue entry, set-up, processing, inspection, rework, storage and dispatch. This reconstruction shows where time and material were consumed and whether each loss originated at that activity or arrived from an earlier decision.

Event data alone can also mislead. A recorded processing period may include a planned break, a micro-stop or time spent waiting for an operator to close the transaction. Marketways tests the digital record against production rules, operator knowledge and selected physical observations. The resulting model must represent the factory's actual operation before management relies on it.

Separate capacity, availability, performance and quality losses

A resource can limit output for several reasons. Its designed cycle may be too slow for the required volume. Breakdowns may reduce available production time. Short stops or lower operating speed may reduce performance. Scrap and rework may consume time without creating saleable output.

These losses are not interchangeable. Additional equipment may address a genuine capacity shortfall, but it will not correct repeated breakdowns or an unstable process. More preventive maintenance may improve availability, but it will not resolve excessive changeover time caused by the production sequence. Marketways quantifies each loss separately so that the proposed action addresses the mechanism that constrained output.

The same distinction applies to labour and supporting activities. A machine may appear constrained because no qualified operator, material handler, tool, laboratory result or maintenance technician was available when required. The diagnostic therefore includes the resources and approvals that allow production equipment to operate, rather than treating machine nameplate capacity as usable capacity.

Product mix and changeovers can alter the limiting resource

Factories that make several products divide capacity across different routes, processing times and set-ups. A plan with more of a labour-intensive product may constrain an assembly stage. A different mix may place the greatest load on machining, testing or packaging. The production stage limiting saleable output can therefore change when the mix changes.

Changeover policy adds another dependency. Longer campaigns reduce the number of set-ups, but they can also increase inventory and delay products scheduled later. Shorter campaigns improve responsiveness, but frequent set-ups consume capacity. Marketways evaluates this trade-off against demand, due dates, batch rules and available buffers.

Management also needs to distinguish a primary constraint from a closely loaded secondary resource. An intervention at the primary constraint can transfer the limitation to the secondary resource. If the analysis stops at the first improvement, the expected increase in final output may not occur. We therefore test the revised flow after each proposed change.

Local efficiency can reduce overall performance

Managers commonly measure each machine or department by utilisation, units per hour or local cost. Those measures can encourage every resource to remain busy even when the next process cannot accept more material. The factory then produces additional work-in-progress, extends lead time and conceals quality problems without increasing customer shipments.

For this reason, Marketways evaluates a local improvement through its effect on the complete production system. Faster processing at a non-constrained machine has little value if material still waits at the true constraint. The same change can be harmful if it overloads inspection, increases rework or consumes material needed for a higher-priority order.

The relevant outcome is saleable throughput achieved with acceptable lead time, quality, work-in-progress and operating cost. Local utilisation remains useful for diagnosis, but management should not treat it as the objective of the factory.

Test the intervention before committing capital or changing the schedule

Historical data can identify recurring associations between production conditions and lost output. However, an association does not prove that changing the suspected cause will release throughput. The factory may have changed several conditions at the same time, or the recorded loss may reflect a downstream constraint.

Marketways develops a test suited to the operational decision. The test may compare controlled production periods, vary a batch or sequencing rule, examine a natural change in staffing or maintenance, or use a validated discrete-event simulation. Simulation is especially useful when a live trial would disrupt customer orders or when management must compare several interacting changes before selecting one.

Each scenario preserves the relevant product mix, routing, failure patterns, changeovers, buffers and production rules. We then compare alternatives such as adding a shift, reducing set-up time, changing the sequence, increasing a buffer, improving reliability, reallocating labour or adding equipment. This comparison shows which intervention can improve final output and which merely transfers congestion.

Measure whether the factory released saleable throughput

An intervention succeeds only when the factory converts the released capacity into the required output. Marketways therefore measures saleable throughput, schedule attainment, order lead time, work-in-progress, quality yield and overtime before and after the change. We also check whether the stage limiting output changed as expected.

The comparison must cover a representative operating period. A short trial may exclude difficult products, infrequent failures or high-demand conditions. Management should retain the change only when the benefit survives the variation that the factory normally faces.

How Marketways identifies the constraint that matters

Marketways begins with the output, product mix and service conditions management must achieve. We reconstruct production routes, reconcile event and loss records, distinguish persistent constraints from temporary disruption, and identify the mechanism consuming usable capacity. We then test practical interventions against the complete flow and quantify their effect on saleable throughput, lead time, work-in-progress and quality.

The engagement can combine Operational Performance Diagnostic with Process and Workflow Analysis and Redesign, Simulation and Scenario Analysis and Statistics and Econometrics. Where the constraint reflects demand and production-plan misalignment, Sales Forecasting and Demand Planning can connect the operating diagnosis with the required production mix.

Related guides

• Quality Variation and Root-Cause Analysis
• Sales and Operations Planning
• Process Mining for Agentic AI Workflows