Visual Management and Abnormality Limits

Why it matters

Visual management lets workers read the same situation in a similar way and give the necessary response. A lit light or a painted box does not do this by itself. Next to a red box it should be clear which measurement, which limit and which decision are involved. Telling colors apart, seeing the screen from a distance and print quality vary from person to person. For this reason the status is also written in text, and supported by shape or symbol when needed. The person looking at the board should see not only that there is a problem but where help is requested and who holds the response. The same color should not carry contradictory meanings in different boxes.

The abnormality limit and the meaning of the target are determined in advance. In this lesson, 60 s is the operational monitoring limit of the cycle time. A value being at the limit and a value exceeding it are not the same situation. Nor can this number be used as a product quality acceptance measure or an equipment safety limit. Each limit governs its own decision. If the cycle is short but the protective device is unsuitable, the board does not give permission for normal operation. Conversely, when the cycle gets longer, that alone does not show that the part produced is defective. If the condition the indicator measures and the scope of the decision are not written together, the color acquires an authority it does not actually have.

The definition of the measurement matters as much as the threshold. If the start and end marks of the cycle change, the same number represents different work. That the unit of measurement is seconds does not explain whether the record belongs to every product or to a product group. A value whose refresh time is unknown may also not show the current status. Missing data cannot be counted as zero time or as a good result. The board must clearly carry the states no data and check required. That way the system does not silently turn green when the measurement stops. Monitoring data quality is a management task separate from monitoring the status of production, but tied to it.

The limit must not be turned into a score line that makes people change the result to avoid punishment. Nor does every exceedance have to require the same response. A small time overrun can be tied to an on-the-spot review, and a larger or repeated overrun to a defined call for support. But this distinction is agreed in advance; the limit is not widened after the result looks bad. Safety and critical quality conditions take precedence over the time tiers. For learning, both the alerts that occurred and the deviations that were missed are examined. A board giving few alerts does not by itself show that it is well designed; an arrangement that leaves real problems invisible can be quiet too.

The source of the operational monitoring limit must also be explained. The 60 s limit used in this lesson is a given training decision; it was not derived by looking only at the observed average. Limits based on the pace of demand, an agreed working arrangement or another management need answer different questions. Judging process behavior from the distribution of measurements requires a separate study. Representing these concepts with the same line can create false confidence. Stating on the board for what purpose the limit was set tells the user correctly what has changed. When a new product or working arrangement arrives, the validity of the limit is reassessed. During the transition the old and new conditions are kept apart; the conformance decision of the past is not silently redefined afterward.

Formula