Visual standard work and abnormality: detecting deviations

Why it matters

A part beside a fixture may be in the right location, but has it been inspected, is it waiting for processing or must it be segregated? The visual standard at the point of work should reduce this uncertainty. When expected location, quantity, sequence and acceptance status are visible together, the worker can compare them with the observed state. The visual's purpose is not to fill a wall or make work look good from a distance. The person doing the work must be able to understand the normal state from their location and know what to do when something differs. Visual standard work is therefore not a standalone poster replacing the work sequence; it is a way of explaining the current method that supports understanding and deviation detection.

Abnormality means more than a state that looks bad. It is a defined deviation from an expected condition. Examples include a missing part that should be in the fixture, work in progress above its limit or a product with an uncertain inspection result. Not every deviation has the same meaning or action. Some require a safe stop; some require a call for material replenishment; others require segregating the product from accepted output. A visual showing only red does not explain which option applies. Support color with text, shape, location and a clear action rule. The message must also remain understandable under different lighting conditions or for someone who has difficulty distinguishing colors.

Detecting a deviation earlier is useful, but it is not the same as resolving the problem. Event occurrence, detection, notification and safe action are separate recorded instants. A visual improvement may shorten the first interval while help still takes longer to arrive. Reporting detection time alone hides this delay. Also, actual detection time cannot be calculated if the event's occurrence time is unknown. It is wrong to turn such a record into an exact duration using an approximate start. State which event and time interval the measurement represents. Reliable visual management depends on a consistent event definition as well as a clear sign.

Assess the visual's effect using before-and-after observations within the same scope. If the product, point of work, deviation type or time boundaries change, averages cannot be compared directly. Improvement does not aim to select only easy events detected quickly. Also track false alarms, missed deviations and correct action. A sign that attracts everyone's attention but does not support the correct decision is insufficient. Test whether workers can explain the sign's meaning and demonstrate a safe response in the appropriate situation. Keep the visual aligned with the current method too. Even a highly readable old visual can become a source of deviation by teaching the wrong sequence.

Formula