Traditional Performance Measures - What they do and why they don't improve performance
As previously mentioned the three primary traditional performance measures most companies and practitioners use are Utilization, Efficiency and Productivity. I'll do a quick review of what each measure provides and discuss the limitations and potential problems related to these measures.
Utilization measures how intensively a Resource was used to produce a Product or Service. In statistical terms, it measures the relationship between available time (capacity) of a particular resource relative to actual time utilized for producing output. For example, if a particular machine is available 16 hours day and is utilized for 15 hours in a particular day, the Utilization Rate would be 93.75% (15/16) - i.e. we only utilized 93.75% of the available capacity. First reaction many people have when faced with a Utilization Rate of 93.75% is that it is not good. But is it necessarily good or bad or just an indicator or capacity utilization?
Let's examine an example closer to production realities - assume 3 machines with Utilization Rates of:
Assume Labor Utilization Rates of:
Individual rates for 10 people varying from 60% to 100%
If you are the Production Supervisor, what do you really know about your manufacturing performance and more importantly how to improve it? I would contend that all we know from measuring Utilization are capacity utilization measures that are not necessarily related to production performance - more discussion on this later.
Another problem with Utilization is that it measures what happened previously - yesterday, last week, etc. - note the use of the verb was in my definition above. So how do production managers deal with this - it's called 'management by wandering around' - looking for inactivity. There is a fundamentally flawed assumption that Activity is a real-time indicator of utilization - the premise being that if a machine or production person is idle, then Utilization is negatively impacted or if busy, then Utilization must be good. What's the point of keeping every resource busy at maximum utilization?
Efficiency measures how well a Resource performed relative to an established standard. In statistical terms it measures Actual Output relative to Standard Output for a particular resource. For example, if the Standard Production Rate for a particular resource is 100 units per hour and it produces 740 units in an 8 hour day, the Efficiency Rate would be 92.5% (740/800). There are multiple ways to measure Efficiency - the previous example measures Actual vs. Standard Units per Time Period. Another alternative is to measure the Value (Dollar) Volume of Actual Output vs. Standard per Time Period. This alternative is essentially the same as the first example, except that it converts units to monetary value. The reason for doing this is to avoid skewing Efficiency by producing more low value products that are easy to make. A third alternative is Standard Hours Produced - for example, if a standard is established to produce 800 units in an 8 production day and it takes 8.5 hours to produce 800 units then the Standard Hours Produced Efficiency Rate is 94.12% (8/8.5).
So what's wrong with measuring Efficiency? In my experience, the biggest problem is establishing and defining the Standard. What is the standard based on, who set it, under what circumstances was it set, does it reflect production reality, does it consider quality criteria, what is it trying to accomplish? So, while it may superficially appear to be a reasonably good means to measure performance, you're measuring actual output relative to a standard that has a questionable basis.
If you always measure relative to average, you will always be average.
Here's another good example of what's wrong with measuring Efficiency - assume 2 work centers:
A: Standard Rate of 80 units per hour
B: Standard Rate of 100 units per hour
Efficiency Performance Considerations based on the above:
If B is downstream from A, can B achieve 100% efficiency?
If A is downstream from B, should B produce at 100% efficiency?
One of the biggest problems with measuring Efficiency is that it encourages a parochial view performance - i.e. production supervisors are being measured on the performance of their work centers, so why should they care about downline work centers - just achieve their efficiency rates and all's well. Where does quality fit in this - what if a work center is 100% efficient but producing out of tolerance parts causing downline problems? Some production supervisors play a game of 'bury you neighbor' with this - produce as much as possible to bury the next work center downline - while that work center's efficiency and utilization rates may be great, the downline work centers could be severely negatively impacted.
Productivity measures the overall Ability for producing a Product or Service. Statistically it measures actual production output compared to the actual input of resources. For example, if production costs (resource input) are $5,000 per hour and the value of production output is $6,000 then the Cost Productivity is 20% -- meaning that for every $1 of input (cost) we have $1.20 of output value. The most common measure of productivity is Employee Productivity - a company with $5m revenue and 50 employees has an Employee Productivity of $100,000 - i.e. each employee, on average, produces $100,000 of output.
Let's explore the above Employee Productivity example a little more:
» Why are all employees usually counted equally?
» What about direct vs. indirect production employees?
» What if the company adds more machines and reduce employees to 30:
» Productivity moves to $166,667 - is this absolutely better?
» What if costs for additional machines increase total costs and expenses to $5.1m? - we would have great productivity but run at a loss.
In my opinion, Productivity is a nebulous performance measure for manufacturing activity. It works fine for macro level comparisons across companies, industries and economies, but provides very limited indication of real performance at the factory level.
An overall problem with all three primary traditional performance measures is that they are retrospective - data are usually for the previous day or more likely the previous week or month. While historical data can provide useful trend identification, knowing what happened yesterday or last week may have limited relevance to what's happening today. Another major problem with all these measures is that they provide very little indication of how to improve performance. Having data around these performance measures might be interesting, but trying to improve performance is usually a trial and error process with delayed feedback on the results. Another characteristic of traditional performance measures is that they generally measure an actual state or rate compared to a target. The question is whether the targets are relevant overall and to each other.
What manufacturing companies are really trying to achieve is to produce the right product, in the right quantities, at the right time according to customer demand and at a price the customer is willing to pay. The connection between utilization, efficiency, productivity and other traditional manufacturing performance measures to these overall objectives is weak.
“There is nothing so useless as doing efficiently that which should not be done at all.”
Hopefully I've cast some doubt on the effectiveness of these traditional manufacturing performance measures, or at least raised some interest in considering alternatives which will be explored in the next section:Connecting manufacturing performance to business performance
with questions, comments or to contact the author.
© Copyright 2004- Ingistics LLC.