An andon system cuts downtime by shrinking the gap between “something went wrong” and “the right person is fixing it.” That gap decides whether a problem costs you two minutes or two hours. Below, we explain how andon works, which andon system benefits actually matter on the floor, the limits most vendors skip, and how to prove the return in your own plant.
An andon system is a visual and audible alert system that lets operators and machines flag an abnormality the moment it happens, so supervisors, maintenance, or quality teams respond before the problem grows into a stoppage or a defect. The idea comes from the Toyota Production System, where it supports the principle of jidoka: stop, fix the problem at the source, and never pass a defect downstream. You can read Toyota’s own explanation of the Toyota Production System and the Lean Enterprise Institute’s definition of andon for the original context.
Every working andon system runs the same loop, and a weak link anywhere breaks it.
Most plants do the first three steps in some form. Almost none of them do the fourth one well, and the fourth step is where continuous improvement starts.
An andon system is a visual or audible alert system that helps manufacturing teams identify production abnormalities and notify the right people for a response. It can use operator buttons, pull cords, machine signals, or sensors to raise alerts. Digital andon systems can add automatic notifications, escalation rules, and event tracking to help teams respond faster and analyze recurring downtime.
| Layer | Typical options | What it does |
| Inputs | Push buttons, pull cords, machine signals, sensors, PLC outputs | Raises the alert |
| Outputs | Stack lights, andon boards, takt boards, buzzers, mobile alerts | Tells people where to go |
| Logic | Escalation rules, timers, priority levels | Decides who hears about it and when |
| Data | Event logs, downtime reasons, response times | Feeds analysis and improvement |
| Factor | Manual (light, cord, shouting) | Digital andon system |
| Who gets notified | Whoever sees the light | The person assigned to that alert type |
| Escalation | Depends on someone remembering | Automatic after a set time |
| Data captured | Almost none | Every call, response time, and duration |
| Multi-line visibility | Walk the floor | One board or screen per zone |
In short: An andon system turns a problem into a signal, routes it to the right person, tracks the fix, and records the event. Digital andon adds automatic routing, escalation, and data on top of the basic light or cord.
The main andon system benefits are faster response, less unplanned downtime, earlier defect detection, real-time visibility, and a data trail that supports continuous improvement. We see the same pattern across plants: the technology matters less than the habit it builds, which is that every abnormality becomes visible and gets a response.
| Benefit | What changes on the floor | Metric it moves |
| Faster response | Alerts reach the right person immediately | Response time |
| Less downtime | Problems get fixed while they are still small | Downtime minutes, MTTR |
| Earlier defect detection | Operators stop the line at the source | First-pass yield, scrap |
| Real-time visibility | Supervisors see line status without walking | Supervisor time on the floor |
| Continuous improvement data | Repeat stoppages become obvious | Top downtime reasons |
| Accountability | Every call has a time stamp and an owner | Escalation count |
In short: Andon benefits fall into three groups: speed (response and repair), quality (defects caught early), and learning (data on repeat problems). The metrics in the right-hand column are the ones to baseline before you start.
An andon system cuts downtime by shortening each stage of a stoppage: how long it takes to notice the problem, how long it takes someone to arrive, and how long the repair takes once they do. It also stops small issues from cascading into larger ones, and it shows you which stoppages keep coming back.
Without andon, a stopped station can sit unnoticed until a supervisor walks past or a downstream operator runs out of parts. With andon, the signal fires at the moment of the stop. That detection gap is usually the cheapest minutes to win back.
A technician who receives an alert that names the line, the station, and the fault type arrives knowing what to bring. Fewer return trips for tools or parts mean a shorter mean time to repair (MTTR).
Consider a typical scenario. A labelling station starts rejecting one pack in fifty because a sensor drifts out of alignment. With no alert, the operator keeps going, the rejects pile up, and a downstream station starves. With andon, the first repeated rejection triggers a call, a technician realigns the sensor, and the line never stops. The fix takes minutes instead of a lost shift.
Because the system captures every event, your top five stoppage reasons stop being opinions and become a ranked list. Maintenance can then target the repeat offenders instead of firefighting.
| Stage of a stoppage | What happens without andon | What andon changes |
| Detect | Someone notices eventually | Alert fires at the moment of the stop |
| Respond | Operator searches for help | Right person is notified automatically |
| Repair | Technician arrives without context | Technician knows the station and fault type |
| Learn | Cause is remembered, not recorded | Event is logged with time and duration |
In short: Andon attacks downtime at four points: detection, response, repair, and learning. The first two usually deliver the fastest wins because they need no new machinery, only a reliable signal and a clear owner.
An andon system speeds up response by sending each alert to a specific person or role and by escalating automatically when nobody acts. It removes the two biggest delays on a busy floor: hunting for the right person and waiting for someone to notice.
A quality issue should go to quality, a breakdown to maintenance, and a material shortage to the stores or logistics team. Mixed alerts that go to everyone teach people to ignore them.
Set a time limit at each level, so an unanswered call climbs the chain without anyone chasing it. The times below are an example only; set yours from your own baseline.
| Level | Who is alerted | Example trigger |
| 1 | Line lead | Alert raised |
| 2 | Maintenance or quality owner | Not acknowledged within a few minutes |
| 3 | Production manager | Not resolved within the agreed limit |
| 4 | Plant head | Repeated or extended stoppage |
Boards work for people on the floor, but your maintenance team is often elsewhere in the plant. Mobile and wearable alerts reach them wherever they stand. Our wearables range and our Mobilizing ERP solution both support taking shop-floor information to people instead of making people come to a screen.
In short: Faster response comes from three design choices: route by alert type, escalate on a timer, and deliver alerts where people actually are. Technology only supports those decisions.
Andon also improves quality, safety, and productivity because it gives operators the authority to stop and raise a problem instead of working around it. That cultural shift is the core of jidoka, which the Lean Enterprise Institute describes in its entry on jidoka.
When an operator can raise a call at the first sign of a defect, the problem stays at one station instead of reaching final inspection as a whole batch. Combine andon with work in progress tracking and you can also see which jobs sat behind the stoppage. Our post on how WIP tracking reduces production delays shows how the two work together.
A visible, audible call brings people to a station quickly when an operator is struggling with a jam or an unsafe condition. A rule of thumb we share with plants: if raising a call feels risky for the operator, the system will fail.
Every stoppage that lasts minutes instead of an hour keeps downstream stations working. Availability, the first factor in overall equipment effectiveness (OEE), is where most andon gains show up.
In short: Andon improves quality by stopping defects at the source, improves safety through faster help, and lifts productivity by protecting availability. All three depend on operators trusting that a call will be welcome.
A good board shows line status, open calls, and pace against target in a way that people can read from a distance while they move. A board that needs study is a dashboard, not an andon board.
The two tools do different jobs. An andon board shows problems as they happen; a takt board shows whether production is keeping pace with customer demand. Our Andon and Takt Board solution combines both, so the same screen tells you when something is wrong and whether the line is on pace.
Both boards can connect to existing ERP or production tracking systems, which keeps the data consistent with what planners already use.
In short: An andon board flags problems and a takt board tracks pace. Keep both readable at a glance, keep the layout consistent, and place them close to the lines they serve.
Andon systems fail for people reasons far more often than technical ones. The main risks are alarm fatigue, calls that nobody answers, a blame culture, and underestimating the effort of rollout. We would rather you hear this from us before you invest.
If every minor event fires an alert, people stop reacting. Start with a short list of alert types that matter and add more only when the first ones work.
A call with no named responder and no escalation rule becomes wallpaper. Assign an owner to every alert type before the first go-live.
If an operator who stops the line gets questioned instead of thanked, calls dry up and the system shows a falsely healthy line. Leadership has to reward early calls publicly.
Costs vary widely with line count, hardware, and integration depth, so get a scoped quote before you budget. Training time and the change effort on the floor often cost more than the hardware, and nobody should skip them.
In short: Most andon failures are about people and process: too many alerts, no owners, and a fear of raising calls. Fix those first and the technology works.
Implement andon by piloting one line, defining a small set of triggers, assigning owners and escalation times, training everyone, and then scaling. Resist the urge to roll out across the plant on day one.
| Step | Action | Output |
| 1 | Pick one pilot line with visible downtime | Scope and a baseline |
| 2 | Define 4–6 alert types (breakdown, quality, material, safety) | Alert list |
| 3 | Assign an owner and a response target for each | Responsibility map |
| 4 | Choose inputs and outputs (buttons, lights, boards, mobile) | Hardware plan |
| 5 | Connect to your ERP or production data | Integrated data flow |
| 6 | Train operators and supervisors; explain that calls are welcome | Trained team |
| 7 | Run the pilot, review weekly, adjust, then scale | Rollout plan |
Integration is where many pilots stall. If your data lives in SAP or another ERP, our hardware integration with SAP and other ERP systems service covers that connection, and our Industry 4.0 and IIoT solutions cover the machine-connectivity side. Andon also fits naturally across manufacturing and automotive environments, where takt and line balance matter most.
In short: Start small, define few alerts, name owners, train people, and connect your data. A one-line pilot with a clear baseline teaches you more than a plant-wide launch.
Measure andon ROI by capturing a baseline before the pilot and comparing four numbers afterwards: response time, MTTR, downtime minutes, and OEE availability. Without a baseline, any improvement you claim stays an opinion.
| KPI | How to baseline it | What to track after go-live |
| Response time | Time from stop to first person at station | Average and worst case per shift |
| MTTR | Time from arrival to line running | Average per fault type |
| Downtime minutes | Total stopped time per line per week | Trend over 30, 60, and 90 days |
| OEE availability | Run time divided by planned time | Change versus baseline |
| Repeat stoppages | Count of the same fault per month | Whether top reasons shrink |
Review the numbers at 30, 60, and 90 days. If your logs still live on paper, a digital record helps too; our E-Log Sheets solution replaces manual logbooks so the data feeds the same review.
In short: Baseline first, then track response time, MTTR, downtime minutes, OEE availability, and repeat faults at 30, 60, and 90 days. That evidence also makes the case for scaling to more lines.
The main purpose of an andon system is to make production problems visible the instant they occur so the right people can respond quickly. In practice, we see plants use it to protect two things: line uptime and first-time quality.
The benefits of andon in lean manufacturing include faster response, less downtime, earlier defect detection, better visibility, and data for continuous improvement. It supports the lean goal of removing waste because every minute a line waits for help is waste.
Yes, an andon system reduces downtime when response and repair processes back it up, but the light alone does nothing. Across the plants we work with, the gains come from clear owners, escalation timers, and a team that trusts the system.
An andon board is the display; an andon system includes the triggers, alert logic, escalation rules, and data capture behind it. A board without trigger logic is just a passive screen, which is why we design the full chain together.
No, you can start with simple buttons and signals at stations and add machine signals later. Many plants begin with operator-triggered calls and connect machine data once the habit forms. Our IoT range helps when you reach that stage.
No, andon suits any plant with repeating processes, including pharma, food and beverage, chemicals, and ceramics. Automotive adopted it first because of Toyota, but takt-paced lines in any sector benefit.
Cost depends on the number of lines, the type of inputs and outputs, and the depth of integration, so there is no honest single figure. The best approach is a one-line pilot quote, which also gives you a baseline for ROI.
Andon triggers a response to a problem in the moment, while OEE monitoring measures overall performance over time. They work well together: andon fixes the stoppage and OEE shows whether the fixes add up.
Ready to see how this works on your lines? Explore our Andon and Takt Board solution, review our case studies, or talk to our team.
Copyright @2024 Raj Barcode. All Rights Reserved.