Art of Lean
Back to Reference
Quality & Jidoka

Abnormality Management

The systematic approach to detecting, signaling, responding to, and resolving deviations from the defined standard — the management discipline that makes jidoka and visual management operational on the shop floor.

Japanese

異常管理

ijo kanri

abnormality management; anomaly control

Also known as

Anomaly Management, Ijo Kanri, Exception Management

Definition

Abnormality management is the systematic practice of making deviations from the standard immediately visible and triggering a defined response. It connects three elements: a defined standard (so that deviations can be recognized), a detection and signaling mechanism (so that deviations are immediately visible), and a response system (so that deviations are addressed, not ignored).

At Toyota, abnormality management is the operational discipline that brings jidoka and visual management to life on the shop floor. The principle is that normal and abnormal conditions must be distinguishable at a glance, and any abnormality must trigger an immediate, defined response from the responsible person.

Japanese Origin

Ijo kanri (異常管理) combines 異常 (ijo, “abnormality” or “anomaly” — literally “different from normal”) with 管理 (kanri, “management” or “control”). The term is standard in Japanese manufacturing management. Toyota’s particular emphasis is on the speed and reliability of the response — not just detecting the abnormality, but having a system that ensures someone acts on it immediately.

How It Actually Works

Abnormality management operates at multiple levels simultaneously:

At the machine level — Jidoka devices detect process abnormalities (wrong dimension, missing component, excessive force) and stop the machine automatically. The abnormality is detected without human intervention. This is the most reliable form of abnormality management.

At the operator level — Workers follow standardized work and are trained to recognize when conditions deviate from the standard. When they encounter something abnormal — a part that does not fit correctly, a tool that feels different, a supply that is running low — they signal by pulling the andon cord or pressing the andon button. The expectation is that they signal immediately, not after trying to fix it alone.

At the line level — The andon board displays the status of every station. When a station signals an abnormality, the team leader is expected to respond within one takt time. If the problem cannot be resolved within the takt time, the line stops at a fixed position. This fixed-position stop gives the team leader a defined window to respond before the stop occurs.

At the management level — Daily management boards track abnormalities by type, frequency, and response time. Recurring abnormalities are escalated to problem-solving activities (A3, kaizen). Management’s role is not to respond to each individual abnormality but to ensure the response system is functioning and to drive permanent countermeasures for recurring problems.

The Four Steps

Toyota’s abnormality management follows a consistent pattern:

  1. Define the standard — Without a clear standard, you cannot distinguish normal from abnormal. Standard work, visual controls, and specification limits establish what “normal” looks like.
  2. Detect the deviation — Through machine sensors (jidoka), operator observation, visual management, or process monitoring. Detection must be immediate, not delayed.
  3. Signal and respond — The abnormality is made visible (andon, visual signal) and the responsible person responds within a defined timeframe. The response is triage: contain the immediate problem and prevent defects from flowing downstream.
  4. Investigate and countermeasure — After containment, determine the root cause and implement a permanent countermeasure so the abnormality does not recur. This is where PDCA and A3 thinking connect to abnormality management.

Common Mistakes

No defined standard to deviate from. Abnormality management is impossible without standards. If there is no defined normal condition, everything looks normal — or everything looks abnormal. Standard work, visual controls, and specifications must be in place first.

Detection without response. Installing sensors, alarms, and visual displays is pointless if no one responds when they signal a problem. The response system — who responds, how quickly, with what authority — is the most critical and most commonly missing element.

Responding without investigating. Restarting the machine, replacing the part, and moving on is not abnormality management — it is firefighting. Every abnormality is information about a weakness in the system. Without root cause investigation, the same abnormalities repeat indefinitely.

Punishing people for surfacing abnormalities. If workers who pull the andon cord are criticized for stopping production, they will stop pulling the cord. Toyota’s culture treats abnormality signals as valuable information, not as evidence of failure. The failure is in not signaling — allowing a defect to pass to the next process.