One of the most widely discussed aspects of lean manufacturing is performance measurement, which makes a lot of sense because it is probably the most important element of manufacturing management. You get what you measure. Since most companies measure their performance by metrics based in traditional accounting principles, they get traditional – that is to say, not very lean – results.

The accounting system is usually built around standard costs with labor or machine hours as the driver, so labor efficiency and machine utilization metrics are the norm. Since the accounting system rewards production volume with ‘earned hours’, which in turn earn labor, material and overhead dollars in flexible budgeting schemes, it should come as no surprise that this approach typically achieves good direct labor efficiency, high machine utilization and a lot of production … and a lot of inventory, high overhead expense and not necessarily very good quality.

The flaw with many performance measurement schemes, including those advocated by quite a few ‘experts’, is a failure to differentiate between measuring overall process (or value stream) outcomes, and measuring specific activities or attributes of a portion of the value stream.

For example, in traditional manufacturing measurement schemes, a great deal of weight is often put on labor efficiency. In fact, most companies track and measure labor efficiency or productivity in great detail and monitor the numbers closely at the highest levels of the company. However, labor efficiency is a meaningless number by itself. It is treated as the primary driver of overall cost, and is measured as a surrogate for cost, but it is not the total cost. There can be no substitute or surrogate for measuring total cost.

Management focus must be on total cost, and metrics such as labor efficiency and machine utilization should only be subordinate data points that shop floor managers can use or not as they see fit to attain lowest total cost. The same is true with many quality metrics: First pass yield, rework percentages, and scrap rates are all potentially useful pieces of information for shop floor managers. None of them, however, can stand as substitutes for the final quality performance of the value stream. Only quality metrics from the eyes of the customer can do that.

Even the most basic manufacturing operation is extraordinarily complicated. Most factories have thousands, perhaps millions, of variables moving around at the same time. Just about every event has multiple drivers. Actions people take to optimize one variable often come to the detriment of another. Most performance metrics at the activity level can be traded off against other measures of performance: Labor efficiency can be increased to the detriment of quality; machine utilization can be maximized in the short term to the detriment of the life of machines; delivery performance can be increased to the detriment of inventory levels and overhead expenses.

Management cannot possibly measure all of those thousands of variables with equal attention and diligence. When one or two of them are elevated to the top level – treated as overall process outcome metrics, rather than the event metrics they are – a strong motivation to optimize performance to those few variables is created, usually to the detriment of offsetting variables that have not been elevated to such high level monitoring status.

In My Next Post: Part 2 of the 5 Golden Metrics.
Manufacturing performance is optimized when management recognizes that there are five golden metrics – five measures of value stream performance that really matter – and all other measurements are subordinate.