What Is a Bottleneck Operation? Definition and Examples

A bottleneck operation is the slowest or most overloaded step in a process, the one that limits how much the whole process can produce. Like the narrow neck of a bottle, it decides how fast everything flows out, however wide the rest is. The idea applies far beyond factories, to hospitals, restaurants, offices and software teams.

How a Bottleneck Limits the Whole Process

Any complex process is made of linked activities, and one of them acts as the limit on the entire system.1leanproduction.com. What Is a Bottleneck Operation? Definition and England In manufacturing, that limiting activity is usually called the bottleneck.1leanproduction.com. What Is a Bottleneck Operation? Definition and England It caps throughput, meaning how much gets finished in a given time.2Journal of Manufacturing Technology Management. Application of the theory of constraints to a bottleneck operation in a manufacturing plant

Take a sandwich shop where taking orders handles 60 an hour, making sandwiches handles 30, and the register handles 50. The shop can only serve 30 an hour, set by the sandwich step. Adding a second sandwich maker does raise output, but only until the next-slowest step, the register at 50, takes over as the limit.

The bottleneck is not always the slowest worker or machine. It is a matter of capacity versus demand, not effort or skill. A fast machine can be the bottleneck if far more work is routed to it than to others, and a step can be a bottleneck only at busy times of day or in busy seasons.

You can usually see one at work. Unfinished work piles up in front of a single step, while the steps after it keep running out of things to do. In an office, that looks like one inbox that keeps growing while the people downstream wait for something to arrive.

The pile forms before the bottleneck rather than after it. Steps upstream can produce faster than the bottleneck can accept, so the extra work has nowhere to go and waits. Steps downstream only receive what the bottleneck lets through, so they are never overloaded and often sit idle.

Balance matters too, not just which station is slowest. On an assembly line, the busiest workstation sets the time each unit takes to come off the line.3International Journal of Production Research. LB-ALBP: the lexicographic bottleneck assembly line balancing problem Yet two lines with the same busiest station can behave very differently.3International Journal of Production Research. LB-ALBP: the lexicographic bottleneck assembly line balancing problem Workloads of 10, 10, 10, 4, 3, 3 put three stations at the limit, while 10, 6, 6, 6, 6, 6 leave only one there.3International Journal of Production Research. LB-ALBP: the lexicographic bottleneck assembly line balancing problem The second layout is more reliable, since in the first a slip at any one of three maxed-out stations stalls the line.

Everyday Examples of Bottleneck Operations

In hospital emergency departments, the jam is often getting patients out, not treating them. In Australian hospitals, waiting to leave the emergency department for a ward made up about 60% of time spent in emergency at large hospitals and 20% at major city hospitals.4PubMed. Analysing the emergency department patient journey: Discovery of bottlenecks to emergency department patient flow So the critical bottleneck was moving patients from emergency into inpatient care.4PubMed. Analysing the emergency department patient journey: Discovery of bottlenecks to emergency department patient flow In simulations, halving that departure wait raised the share of admitted patients meeting time targets by up to 22% at large hospitals.4PubMed. Analysing the emergency department patient journey: Discovery of bottlenecks to emergency department patient flow Other common emergency bottlenecks include waiting for a specialist, tests outside the department, discharged patients waiting for transport, finding a bed, and handover to the ward.5PubMed Central. Modelling Granular Process Flow Information to Reduce Bottlenecks in the Emergency Department

Restaurants and other services have bottlenecks too. Multi-stage in-store service, like a fast-food counter, can be analysed for bottlenecks the same way a factory is.6Asia Pacific Journal of Marketing and Logistics. Bottleneck analysis in multi-stage service delivery using the theory of constraints The step where customers queue longest is the one to look at first.

In software teams, review and testing often become the limit. A person who must approve every code review becomes the bottleneck for the whole team.7PostHog Newsletter. Stop being the code review bottleneck When AI tools generated hundreds of code changes quickly, automated testing and reviewers’ attention were swamped.8arXiv. Orchestrating AI-Assisted Code Remediation: Socio-Technical Bottlenecks in a Large Industrial Repository Grouping changes by folder and capping files per change restored the flow of work.8arXiv. Orchestrating AI-Assisted Code Remediation: Socio-Technical Bottlenecks in a Large Industrial Repository It still took asking for reviews, agreeing on change size and repeated follow-up.8arXiv. Orchestrating AI-Assisted Code Remediation: Socio-Technical Bottlenecks in a Large Industrial Repository

Shipping has its own version. Ports are a classic chokepoint, and climate-related disruptions at ports can ripple through global shipping and trade.9Nature Climate Change. Systemic risks from climate-related disruptions at ports

Office work offers plenty of examples you may recognise: a manager who must sign off every expense or purchase, one person who alone knows how to run payroll or a key system, or a shared printer, meeting room or approval queue that everyone waits on. Take the sign-off. The manager may do the check in minutes, but requests wait for days because the manager is in meetings. Everyone who needs the purchase is held up, even though their own steps are quick. The real delay is waiting time in the queue, not the work of approving, so the fix is often changing who can sign small amounts or when approvals are batched.

Single points of knowledge work the same way. When only one person knows how to run payroll, every task that needs that skill must line up behind them. The problem becomes sharply visible when that person is sick or on leave and the whole process stops. Writing down the steps and teaching a second person is usually the cheapest remedy.

Shared resources become bottlenecks when demand clusters at the same times, such as just before a deadline. A meeting room may sit empty for much of the week and still be the limit at peak moments, which is why average use can mislead. Spreading the demand across the day can ease it without buying anything new.

How to Find the Bottleneck in Your Process

Methods for finding bottlenecks rely on queue information, how busy each step is, or both.10Production & Manufacturing Research. Throughput bottleneck detection in manufacturing: a systematic review of the literature on methods and operationalization modes For most people, watching where work waits and which step never stops is a sound start.

You can do this on paper. List every step in order, from the moment work arrives to the moment it is finished, and write down roughly how many items each step can handle per hour or day. Next, write down how many items arrive in the same time, since a step is only a problem if demand is close to or above what it can handle. Mark where piles, inboxes or waiting lists grow, because a growing queue shows where capacity falls short. The step with the lowest capacity compared to demand and the longest queue in front of it is your likely bottleneck, and it is worth checking over a few different days before acting.

Waiting times rise slowly as a step gets busier, then shoot up sharply past a certain point.11arXiv. QUEUEING THEORY AND MODELING The more uneven the arrivals and task times, the earlier that sharp rise comes.11arXiv. QUEUEING THEORY AND MODELING Larger operations with more servers can run closer to full before delays spike.11arXiv. QUEUEING THEORY AND MODELING A small team running near full capacity with unpredictable work is therefore especially likely to jam.

A few mistakes are common when hunting for the bottleneck. Complaints are a poor guide, because the loudest team may be the one that feels the delay most, not the one causing it. A step that sits idle may not be fast; it may simply be starved of work by an earlier bottleneck. The bottleneck can move when demand, staffing or the mix of work changes, so a finding from last season may no longer hold. And averages can hide a step that is only overloaded at peak times, so look at the busiest periods as well.

Digital records can reveal hidden bottlenecks. Process mining uses the time stamps systems already record to trace where work slows.12Proceedings of the 18th International Conference on e-Business. A Classification of Process Mining Bottleneck Analysis Techniques for Operational Support Combining data from several systems with factory-flow analysis can pinpoint constrained resources.13IOS Press. Bottleneck Detection Through Data Integration, Process Mining and Factory Physics-Based Analytics

How to Fix or Manage a Bottleneck

The Theory of Constraints is a management approach built on the idea that one constraint limits each system.1leanproduction.com. What Is a Bottleneck Operation? Definition and England Its five focusing steps guide improvement around that constraint: identify the bottleneck, get the most out of it, make everything else support it, add capacity if needed, then repeat for the next one.

It can work fast and cheaply. One doctor used these steps to raise the number of patients his eye-imaging practice handled by 64% in a few weeks.14PubMed Central. Achieving rapid and significant results in healthcare services by using the theory of constraints

Getting more out of the bottleneck step often costs nothing. Never let it wait for materials, files or approvals; prepare them ahead of time and keep a small supply ready. Stagger breaks and shift changes so the step keeps running while one person is away. Check quality before work reaches the bottleneck, because time spent on a defective item is time lost for the whole process. And move routine tasks such as data entry or cleaning off the bottleneck person or machine, so its time goes to the work only it can do.

Sharing skills spreads the load. Training people to cover more than one job lets work shift away from the most overloaded person.15International Journal of Production Research. Cross-training policies and team performance

Adding people, machines or hours costs money, so it comes after the cheaper steps of removing waiting and wasted time. If the bottleneck still limits output after those changes and demand really is higher than it can handle, extra capacity there is justified.

Once the limit is raised, the next-slowest step becomes the new limit, so output rises only until that step is reached. This shift is normal and shows that the fix worked.

Pacing Work to the Bottleneck With Drum-Buffer-Rope

Drum-buffer-rope is a method that releases new work only as fast as the bottleneck can finish it. The bottleneck’s pace is the drum that sets the rhythm for everything.16International Journal of Production Research. On the beat of the drum: improving the flow shop performance of the Drum–Buffer–Rope scheduling mechanism New jobs wait in a backlog and are released only at the bottleneck’s output rate.16International Journal of Production Research. On the beat of the drum: improving the flow shop performance of the Drum–Buffer–Rope scheduling mechanism The rope is a signal from the bottleneck back to the start of the process.17International Journal of Production Economics. Drum-Buffer-Rope and Workload Control in High-Variety Flow and Job Shops with Bottlenecks: An Assessment by Simulation A job is released only when the number waiting at the bottleneck falls below a set limit.17International Journal of Production Economics. Drum-Buffer-Rope and Workload Control in High-Variety Flow and Job Shops with Bottlenecks: An Assessment by Simulation

The buffer is a small queue of work placed in front of the bottleneck so that it never runs out of things to do. If it is too small, a hiccup upstream, such as a late delivery or an absent worker, leaves the bottleneck idle and costs output. If it is too big, work and inventory pile up, delays grow and problems that would otherwise show up early stay hidden. The right size is something to adjust over time by watching whether the bottleneck ever runs dry.

The idea works outside factories. A clinic with one scanner can book appointments only as fast as the scanner can see people, so patients do not crowd the waiting room. A kitchen can stop taking orders faster than the grill can cook them, so tickets do not stack up and food does not go cold. A software team can limit how many tasks sit waiting for review, and start new work only when that number falls.

Drum-buffer-rope beats other release methods when there is one strong, clear bottleneck.17International Journal of Production Economics. Drum-Buffer-Rope and Workload Control in High-Variety Flow and Job Shops with Bottlenecks: An Assessment by Simulation When the bottleneck is mild, other workload control methods perform better.17International Journal of Production Economics. Drum-Buffer-Rope and Workload Control in High-Variety Flow and Job Shops with Bottlenecks: An Assessment by Simulation The more spare capacity other steps have compared with the bottleneck, the less stockpiled work is needed.18International Journal of Production Research. Order review/release and lot splitting in drum-buffer-rope

What Bottlenecks Cost a Business

In a steel plant model, a warehouse bottleneck plus machine downtime raised total yearly cost about 4%.19Reliability: Theory & Applications. OPTIMIZING STEEL INDUSTRY PRODUCTION INVENTORY SYSTEM CONSIDERING WAREHOUSE BOTTLENECK AND MACHINE DOWNTIME In a supply chain model, easing bottlenecks cut total cost by about 20%.20Applied Sciences. Investigation of Bottlenecks in Supply Chain System for Minimizing Total Cost by Integrating Manufacturing Modelling Based on MINLP Approach Costs come both from lost output and from stock and delays piling up.

Many of the costs are easy to overlook. Overtime is often paid at the bottleneck while staff at other steps wait with nothing to do. Late deliveries lead to unhappy customers, and some of them do not come back. Backlog and extra stock tie up cash and space that could be used elsewhere. Staff at the overloaded step face constant pressure, which can lead to errors, burnout and people leaving.

An hour lost at the bottleneck is an hour lost for the whole process, because nothing can make up that output downstream. Spending to protect the bottleneck, for instance through maintenance, backup staff or avoiding stoppages, usually pays back best.