Effectively evaluate efficiency
Feb. 1, 2014
by Jim Kline
Here we are, early in the New Year — and there’s no better time to look at how your bakery operations are performing. To someone from outside the baking industry, it would seem to be a simple enough task, one that could be easily accomplished. But dig a little deeper: What do the metrics that we recite for production line efficiency, process loss, and throughput represent? You’ll likely get many different answers. We really do not have an accepted standard for what these metrics represent, nor how they are calculated. And the variation in metrics used by different operations management systems and continuous improvement programs adds further complexity to determine how efficiently your process is running.
Let’s look at one of these metrics: efficiency. Many times the true efficiency of a production line is not known. This can be the result of many factors — a lack of appreciation for the value of metrics as production management tools, an incomplete assessment of contributing factors, inaccurate data collection or errors in the computational models in use.
Production line efficiency is a fairly easy concept to grasp; however, like so many things, the devil is in the details. Simply stated, it is the percentage of productive hours realized within a scheduled period of time; the percentage results from dividing productive time by scheduled time, such as the graphic here represents.
So where does the devil lurk?
First you must define “scheduled time.” Keep it simple; it is the hours the line is scheduled to run at the divider, depositor, sheeter, etc. Too often, changeover times are subtracted, as are scheduled breaks; count these for what they are: non-productive times.
Next, define “productive time.” Plain and simple, this is the time when the line is running and producing saleable, specification product. When changeovers are taking place, non-saleable product is made or when the line is idle for employee breaks or lunch — the time associated with these events needs to be subtracted from the productive hours.
Third, define “non-productive time.” Downtime, both mechanical and operational (changeovers, running out of dough, downtime associated with employee breaks and lunch, the time during which non-saleable product is produced, time spent reworking products, QA sampling, etc.), is to be included. Scrap dough and defective products made during scheduled production must be converted from units or weight to time and accounted for as such. Base non-productive hours on make-up time. Don’t get caught in the trap of confusing machine downtime with line downtime; there is a difference.
How you track your times, output, scrap or rework will vary with each location and the installed processes. What is important is properly training employees in gathering the operational information required to properly assess performance. Equally important is providing them with the tools to accurately record the events of their shift. The line lead or supervisor needs to review the information and ensure there is consistency between the various work stations along the line.
Finally, the devil lurks in computational error that results from a failure to recognize that the inefficiencies at each step in the process are compounded, not averaged or additive. Consider a five-step baking process.
Ingredient handling and mixing operates at 98% efficiency.
Dividing and forming operates at 97% efficiency.
Proofing operates at 99% efficiency.
Baking operates at 96% efficiency.
Cooling and packaging operates at 94% efficiency.
The overall efficiency that this process runs at is not 96.8% (the average of the individual efficiencies), nor 84% (100% less the sum of the inefficiencies). It is 84.9%, which is a multiple of the various efficiencies along the line.
If ingredient handling and mixing are operating at 98% efficiency, then the best the line can run is 98%. But then dividing and forming are only operating at 97% efficiency. The best the line can do is 97% of 98%, the basis for calculating line efficiency performance. So we get .98 x .97 x .99 x .96 x .94, resulting in 84.9% efficiency.
If this production line also had 0.9% in mechanical downtime and 2.4% operational downtime for changeovers, and process loss (scrap and QA sampling) of 3.8%, then 77.8% is the true efficiency of the line, despite the fact that no individual operating efficiency was below 94%.
Armed with this information, what can you do with it? In this example, the true efficiency for the line would seem to indicate some opportunity for improvement. I would suggest collecting and analyzing a few weeks’ worth of data, looking for trends.
Convert the non-productive time to potential volume and income. What is its potential value? This will provide a basis for evaluating future improvements.
Identify the top five contributors to non-productive time, and determine how can they be addressed and resolved. Do the same for the top five contributors to downtime. Keep in mind a similar approach can be used for process loss and labor productivity. Be productive, and get a jump start on a great New Year!