"When they have a problem, they hit this button and our pager vibrates. We know when it vibrates, it means that in one of our test cells, somebody pushed that button."
That's how one medical device manufacturer described their Andon system in action. The concept is simple: operator needs help, presses button, responder gets notified. But behind that simplicity is a workflow that makes the difference between fast response and missed calls.
This article walks through how Andon systems actually work—from the moment someone presses a button to the data that gets captured afterward.
The Basic Andon Workflow
At its core, an Andon system follows a predictable sequence:
- Alert — Operator identifies a problem and triggers a call
- Notify — Responders receive the alert
- Locate — Alert indicates where help is needed
- Respond — Responder arrives and acknowledges
- Resolve — Problem gets fixed
- Close — Call is completed and logged
- Escalate — Automatic notification to backup or supervisor
Let's look at each step in detail.
Step 1: The Alert Trigger
The workflow starts when an operator needs help. In a traditional setup, this means pressing a physical call button mounted at their workstation.
Types of Alert Triggers
Manual buttons are the most common. Operators press a button when they encounter a problem they can't resolve themselves.
Many systems use multi-key buttons that allow operators to specify the type of help needed. As one manufacturer explained: "One through four can be configured as you need—technician, quality, supervisor, maintenance."
This differentiation matters because different call types often route to different responders.
Automated triggers can also initiate alerts. PLCs and machine sensors can send signals when equipment enters fault states, when cycle times exceed thresholds, or when production counts fall behind schedule.
What Gets Transmitted
When a button is pressed, the system captures:
- Which station initiated the call
- What type of call it is (if applicable)
- The timestamp
Step 2: Notification
The alert is only useful if responders receive it. Andon systems use multiple notification methods, often in combination.
Wearable Pagers and Watches
The most direct method: responders wear pagers or watches that vibrate when a call comes in. As that medical device manufacturer described, "they hit this button and our pager vibrates."
Wearables work well in noisy environments where audible alerts get lost, and they ensure the notification reaches the right person regardless of their location on the floor.
Visual Displays
Dashboards and display boards provide real-time status visibility. As one manufacturer described it: "You'll see everything in green. Once anything gets triggered, it'll go to red and you'll know that that area needs attention."
Displays can be shown on:
- Dedicated monitors on the production floor
- TVs mounted in common areas
- Computer screens in supervisor offices
- Tablets and smartphones on the network
Audible Signals
Stack lights with buzzers provide localized audio-visual notification. When a call triggers, the light changes color and the buzzer sounds, drawing attention to the affected area.
Digital Notifications
Some systems can send alerts via SMS text message or email. This extends reach beyond the immediate floor—useful for supervisors who may be in meetings or managers working remotely.
Step 3: Location Identification
Knowing that help is needed is only half the equation. Responders need to know where.
"I just need my technicians to be notified—and preferably, which cell to go to."
Good Andon systems make location obvious. As one manufacturer explained the configuration: "If it's cell one, it's going to say cell one. If it's cell two, it's going to say cell two."
Station Naming
Stations can be labeled however makes sense for the facility:
- Cell numbers (Cell 1, Cell 2, Cell 3)
- Line names (Assembly Line A, Packaging Line 2)
- Machine identifiers (CNC-04, Press-12)
- Area descriptions (East Warehouse, Shipping Dock)
Step 4: Response and Acknowledgment
This is where many basic systems fall short. A responder receives the alert and heads to the station—but the system doesn't know they're on their way.
More capable systems include an acknowledgment step.
"The responding team gets to the station. They press the yellow button. What that does is it stops the initial timer to let you know how long they waited for the call."
Why Acknowledgment Matters
Acknowledgment does several things:
Stops the wait timer. Now the system knows how long the operator waited before help arrived.
Shows help has arrived. Supervisors watching the dashboard can see that someone is responding.
Creates accountability. There's a record of who responded to which call.
Prevents duplicate responses. Other responders know the call is being handled.
Some systems capture the responder's identity (which pager acknowledged), adding another layer of data.
Step 5: Resolution and Closure
After the problem is fixed, the call needs to be closed. This is typically done by pressing another button at the station.
"Once you're done, they press the green button, which is going to cancel everything overall."
Separate Tracking for Repair Time
Advanced systems track two distinct time periods:
- Wait time — From alert trigger to responder arrival
- Repair time — From arrival to problem resolution
This separation matters for continuous improvement. Long wait times suggest notification or staffing problems. Long repair times suggest training or equipment problems. Tracking them separately helps identify the root cause.
Escalation: When Response Doesn't Happen
What if no one responds? Without escalation, the call sits. The operator waits. Production stops.
Escalation is the safety net.
"For example, you call for cell two and it's been 5 minutes and no one has came to cell two—you can preprogram within the software at that five minute mark to send an automatic message."
Escalation Options
Systems can be configured with different escalation behaviors:
Reminder to original team. After a timeout, send another alert to the same group. Maybe they missed it the first time.
Escalation to backup team. If the primary responder doesn't acknowledge, notify a secondary group.
Escalation to supervisor. "If the call's been over 10 minutes, you can have it send multiple reminders or escalations to the pagers... send an escalation to a different grouping."
Configurable Timeouts
Escalation thresholds vary by situation. A safety call might escalate after two minutes. A materials request might allow ten minutes before escalation. Good systems let you configure these rules per call type.
Data and Reporting Basics
Every completed call generates data. Over time, this data reveals patterns.
What Gets Captured
A typical call record includes:
- Timestamp — When the call started
- Station — Where it originated
- Call type — What kind of help was requested
- Wait time — How long until someone responded
- Repair time — How long the fix took
- Responder — Who handled the call
- Escalation — Whether the call escalated
Using the Data
With enough history, patterns emerge:
- Which stations generate the most calls?
- Which call types take longest to resolve?
- Which shifts have the slowest response times?
- Which responders are most efficient?
Variations by Industry
While the basic workflow remains consistent, implementations vary by industry:
Medical Device Manufacturing
Clean room environments often use wireless systems because wired installations compromise sterility. Alerts may go to specialized technicians who support multiple test cells.Food Manufacturing
Large facilities with multiple production lines may have dedicated response teams per area. Shift handoffs require clear status visibility so incoming teams know what's active.Automotive and Tier-1 Suppliers
Quality and maintenance often operate as separate teams with different escalation paths. Line stoppages may automatically trigger alerts to multiple groups simultaneously.General Manufacturing
Multi-department facilities may have complex routing rules. A call from Station 12 might go to Team A during first shift and Team B during second shift.Frequently Asked Questions
How fast do alerts reach responders?
With wearable pagers, typically within seconds. The button press triggers immediate transmission. Network-based displays update in real-time.
What happens if the button is pressed by mistake?
Most systems include a cancel function. The operator presses a cancel button, and the alert is cleared before anyone responds. Some systems log these as canceled calls for tracking purposes.
Can different people get different types of calls?
Yes. This is one of the main advantages over broadcast systems like intercoms. A maintenance call can route to maintenance techs. A quality call can route to inspectors. Each group only sees their relevant alerts.
Does the system work without software?
Basic systems—button, transmitter, pagers—can work without software. The alert goes out, pagers receive it. But without software, there's no logging, no dashboards, no escalation rules, and no reporting.
Next Steps
Understanding how Andon systems work is the first step toward evaluating whether one fits your operation. The workflow is straightforward, but the details—notification methods, escalation rules, reporting capabilities—vary significantly between solutions.
Related reading: