Andon Call, Response and Escalation

Why it matters

Andon is a call system that reports an abnormality or a need for help by light, sound or a visible record. The value of the system is not only that it produces a visible signal. Who will answer the call, in which situations support will be requested and what will be verified for closure must be defined. When K01 reports a problem, K01 wants to know the message was seen; but a message being seen does not show that help has come. Accepting the call, reaching the site and restoring the normal working condition are different events. Writing these events into the same time field hides which part of the service is late. Even if the technical tool is simple, when the response arrangement is set up clearly the records produce information that serves improvement.

The life of a call may begin the moment the abnormality is detected; the opening of the call may happen later. The response calculation in this lesson runs from the opening of the call to the arrival of competent support at the site. The delay between detection and call is kept separately. Intervention time is measured from arrival to the verification of the normal condition, and total recovery time from the call to this verification. This way the same interval is not counted twice. Another plant may use different definitions; starts and ends are equalized before any comparison. It is possible to record in seconds and show minutes in the report; the unit conversion is explained. A missing arrival time is not estimated and included in the average; the gap itself is kept visible.

A return to normal condition may not mean the permanent cause has been removed. With a temporary correction the operation can be continued safely while the cause investigation stays open. In this case it is tied clearly by which evidence the production call was closed and by which record the cause work continues. Switching off the light, removing the call from the screen or resetting the counter may create no change at all in the physical process. A closure measure based on the light going out can encourage records to be closed early under pressure for fast closure. For this reason the person who defines the closure status and the person who reviews the records use the same acceptance condition. Recurring calls are not deleted; under the same-cause assumption their links are kept, and the assumption is tested separately.

Escalation is not about punishing the person who gave first help; it is about bringing the needed authority and resources in on time. Different support paths are needed when no response comes, when the risk grows or when the problem goes beyond the authority given. Using a single threshold for all calls may not be suitable; the waiting condition for high-impact abnormalities is different. The numeric thresholds here belong only to the training scenario. Where human safety or the movement of unverified product is involved, a good average response time does not make the event acceptable. The team aims to reduce the causes of abnormalities that block normal work rather than to lower the number of calls. A low number of calls is not by itself evidence of few problems.

Punishing the worker who reports can delay calls; a low call count is therefore not evidence of few problems. The simultaneous workload and the backup of those giving support are kept visible. While priority is set by risk, product protection for waiting events is maintained. The aim is not only to switch lights off quickly but to bring the needed authority and resources to the site on time.

Formula