Bottleneck and throughput: how to calculate line capacity

Why it matters

The bottleneck is the resource that limits system output for the selected product mix and period. The longest cycle or largest queue can be a starting point for investigation; neither is conclusive on its own. One station may have parallel resources, while another waits for a shared worker. If products follow different routes, they do not all consume the same capacity. First draw the flow path and resource arrangement. Then assess available time, quantity per cycle, stoppages and quality together. A bottleneck is not a machine's permanent identity; when conditions change, the limiting resource may change too.

Capacity and actual throughput are separate concepts. Capacity without losses shows the upper limit achievable under explicit assumptions. Actual throughput is the number of good units leaving in a given time divided by that time. Operating below the capacity limit does not necessarily mean the machine is slow. Material shortages, a full downstream area, breakdowns or low demand can reduce output. An operating rate calculated by subtracting stoppages from time must not be presented as the result for the whole shift. Changing a rate's denominator is not an improvement. Different rates for the same period are meaningful together only when their scopes are explicitly stated.

Inventory before the bottleneck is a useful observation, but its cause must be questioned. Material arriving in batches can create a short-term queue. Incorrect sequencing or missing information can also cause a buildup. A limiting resource working continuously with material available ahead of it, while downstream resources wait for material, provides stronger evidence. Records over time are still needed. An idle station is not always a local problem that increases waste; it may be excess capacity waiting in line with demand. However, a bottleneck waiting for material or output space can prevent the system's capacity from being used.

Assess an improvement option by the system's good output. Speeding up a resource that already has enough capacity may only increase work in progress. By contrast, increasing the reliability of work feeding the bottleneck or clearing its output can improve the system result. The goal is not to keep all machines running continuously. First identify the largest system loss, then choose a small, verifiable intervention. Compare all resources again in the new capacity calculation; speeding up the old bottleneck may reach the next limit. A local reduction in cycle time and an increase in system output do not have to be the same percentage. Do not make an investment decision without making this distinction visible.

Formula