Bill Gaw's Triple-Step, World Class, Lean Manufacturing Training Program
World Class Manufacturing

Increase the effectiveness of your
Lean Manufacturing Training Program
Lean Manufacturing Simulation Game 

Eight Requirements of an Effective Performance Measure

Drucker identifies eight requirements of an effective per­formance measurement system. These are critical compo­nents in building a model for developing and auditing a performance measurement system. Performance measure­ments should be:

• Economical: The cost of the measurement should not exceed the benefit derived. The bureaucratic need to measure things with precision can violate this require­ment, especially in trying to measure complex process effectiveness. This perceived need to measure pre­cisely things that cannot be defined precisely leads to the substitution of expensive, complex, static effi­ciency measurement system which provide little infor­mation and are easily manipulated by those being measured.
• Meaningful: Those being measured must clearly un­derstand the relationship between the measurements and the goals. Also, it is important that those being measured get continuing feedback on progress toward the goals. This means that arguments about confidenti­ality (No "need to know", etc.) have no place in a perfor­mance measurement system. Also, meaningfulness ar­gues for vector measures rather than static measures.
• Appropriate: The level of detail must be such to make the measurement convey useful information. Too much detail in the measurement (e.g. individual perfor­mance standards by individual operation) leads to the measurement of common cause noise in the process rather than actual process improvement or decline. Too little detail (e.g. financial statements) combines
and averages too many diverse factors and loses any understanding of the distribution of the measure. Overdetailing of direct labor reporting is an example of too little detail and summarizing of overhead reporting is an example of the latter.
• Congruent: Since most measures are indirect, it is important that the way in which the measure varies reflects the way in which performance is actually varying. A measure that moves in a contrary direction to actual performance, even part of the time, can motivate the people being measured to work to maxi­mize the measure to the detriment of actual system performance. For example, as companies move into just-in-time or total quality environments, traditional cost systems may indicate that overhead rates and product costs are actually increasing when this is not so. This will cause both managers and people being measured to resist changes and the organization ini­tiatives required to implement these new organization structures successfully.
• Timely: Information has time value. As it ages, its usefulness diminishes rapidly. Too many measure­ment systems don't get information to people being measured rapidly enough for them to adjust their performance to realign it with goals. Old measure­ments are only useful for fault finding and other negative activities. Given the ability of information systems to manipulate data, there is no reason for performance measures not to be reported in a timely manner. When this occurs it is usually the result of excessive human intervention (detailed reviews, mul­tiple approvals, etc.) delaying the reporting of perfor­mance measures to those who will use them.
• Simple: Performance measures must be understand­able to those who are expected to benefit from them. The right unit of measure (e.g., don't measure shop floor performance in dollars), the proper level of detail and a clear understanding on the part of those mea­sured of what the measure means and how it relates to goals are critical requirements of the performance measurement system.
• Operational: The measurement of accuracy using sta­tistical tools and the use of vector measurements both require continuous and reliably consistent data from which the performance measure is derived. Special studies and analysis, while providing valuable supple­mentary information, tend to be static measures, pro­viding no information on direction of performance or the statistical distribution of the information reported. Thus, any performance measurement must be based on readily and continuously available data and the mechanism must be put in place to provide this data for any performance measure desired.

To be Continued