Process Control in a Nutshell
There are many actions a small company can take to improve upon its operations and most companies will reap significant rewards by searching for new ideas that fit within the scope of a small company.
Companies that adopt new methods for controlling processes will find that they can increase productivity by monitoring, evaluating and modifying operations when there is a production slowdown. Process control also helps profits by reducing rejects and process variations.
Process control, while potentially a comprehensive discipline, doesn't have to be all-encompassing when used in a small plant. In fact, the first step in formalizing process control may add only a small amount of time, twice a day. Examining production and quality figures for five minutes in the morning and five minutes in the afternoon are all it takes to begin a formal process control procedure.
All plants have some process control even if the process control is spontaneous, however, if a plant is not being run by the numbers, the minor slowdowns will not be caught and the minor increases in rejects will go undiscovered until the problem becomes so obvious that it can't be ignored. Even a slowdown of only 5% can easily be missed by casual managers and that 5% will drop down to reduce profits. By catching and correcting the slowdowns early, the attentive manager who monitors his processes, will realize a cost advantage over his less careful associates. Thus, there is a cost advantage to process control when the plant runs at peak performance all the time. In addition, the cost advantage will grow significantly higher for companies that use process control to improve their plant.
At this point I may have lost a number of people. Some managers will say, "Process Control is a rigorous and time consuming discipline. This article seems to trivialize a study that is covered in depth in engineering colleges." Other people may say, "I have heard how time consuming process control can be. The utilization of statistical methodologies and process engineering are way beyond the capabilities of my small company."
But this simple form of process control is designed specifically with the small company in mind. It takes a common sense concept and begins with those aspects that can be of practical use without painstaking academics. The knowledge learned in an engineering college will never be applied to a small company because most small companies cannot afford to hire process engineers. Yet, the need and the benefit of an uncomplicated control system should be available to even the smallest companies. I do not mean to insult the intelligence of an engineer by suggesting that all his knowledge can be condensed into a short article. Nor do I want to trivialize the benefit that statistical analysis can have on larger plants, but I also want to convince an audience of small business owners that process control is a smart thing to do from a control perspective, a cost perspective and a customer perspective and that process control can be started in a company which can't achieve the sophistication of a major corporation.
Before getting into particulars, consider the following general concepts:
Process Control includes the monitoring, evaluation and modification of production.
The purpose of process controls is to maintain existing output on a regular basis and to identify positive changes in the operation to increase quality and quantity on a long range basis.
In an ideal operation, the managers know at each moment how their operation is functioning with regard to speed, quality and variability. In all companies the management must know how operations are functioning at least on a daily basis.
Monitoring can be accomplished by measuring output over a set time period, monitoring mistakes, plotting variations from norms on a graph, calculating actual profit versus expected profit and/or keeping a diary of operating conditions/results.
Evaluation is accomplished by drawing conclusions about the information received from monitoring.
Modification is accomplished by making changes to the equipment, processes, skill levels, material, work environment, soft ware and management practices.
Fine-tuning is a subset of modification that is done on a frequent basis to keep an operation at peak performance.
Process control can be used on individual processes or on a whole system of processes at once.
Process control can be applied to production, shipping, sales, design, clerical work and data processing to increase efficiency / output / quality in each of those functions.
Selecting Useful Statistics: Monitoring
The manager who wants to initiate a formal process control will first choose statistics that are meaningful, predictable and which give rise to the company's primary goal (which is usually profit). As an example, consider the possible statistics for production. These include (1) units produced per hour (or day, week, etc.), (2) units produced per man hour, (3) dollars produced per day and (4) profits produced per order. In each case, the manager seeks to optimize the selected statistic and, in doing so, he will optimize his profits for the time period under consideration.
The reader may readily understand how each of the four potential statistics are meaningful and how each give rise to profit but the requirement for predictability may not be obvious at first. This predictability requirement is necessary to be able to set a benchmark for the day's production and the benchmark is necessary to gage how close the production comes to expectations.
For a plant that produces a daily average number of units or a daily average dollar amount either of these two statistics are good possibilities. However, a plant that produces custom work may show unpredictable variations in units or dollar value so that daily production is not constant. In these cases the benchmark must come from the estimator's calculations either by adding up the hours from the estimator's worksheets or by using his estimated profit as the benchmark against actual production profit for each job. Be aware, however, that the statistic, profits produced per order, is not a function of time. This is a weakness that can be removed by restating the profit statistic into profit per hour; by dividing the total profits by the number of man hours spent to fulfill the order.
Process control includes the monitoring, evaluation and modification of production.
Before monitoring can be done a statistic must selected.
Profit (the primary objective) is an important consideration when selecting a statistic to monitor.
Predictability is an important consideration when selecting a statistic to monitor.
Time is an important consideration when selecting a statistic to monitor.
For production the possible choices that fulfill these requirements are (1) total units produced, (2) total dollars produced (3) units produced per man and (4) profits produced per order all for a set time period.
The following form shows a simple example of a monitoring chart:
Process Control: Comparison to expected production
Company / Process:
AM / PM Date# Units Produced (or dollars)# Units Expected (or dollars)% over (under) expectation12345
All processes can be controlled with well selected statistics. The example in this paper is production, but the concepts presented here can be adapted to other functions that give rise to revenue, expense or profit. This includes sales, clerical, design, shipping and quality functions and it includes service companies as well as manufacturing companies.
No matter what functions the businessman selects to control, he should make substantial progress when compared to less formal approaches.
A manager that can change his perspective is a better manager than a manager who always follows the same old rules. A manager that can change his perspective will usually see more possibilities and more opportunities in any given situation then the manager that follows the same rules every day. A conventional manager may have blind spots that prevent him from seeing many varied options at any one time.
Often I encounter managers that want to continue doing what they have always done. Usually, they focus on getting the job done accurately and quickly. They have set up routines and schedules and plans which they believe can not be improved upon. Sometimes they say, "If I get everyone to do his job in the best way, then how will a new perspective bring about improvements? If our customers praise our quality and timeliness, what more can we do?"
While this sounds logical on the surface, the logic itself is narrow. On a broader sense a company wants to satisfy other goals besides customer satisfaction (although customer satisfaction is definitely an important goal).
A company wants to make a profit. In fact, profit is often the most important company goal. To maximize profit, the company may need to maximize revenue, minimize costs, increase efficiency, manage time, improve cash flow and be selective about orders. So, the manager with the perspective of keeping things running is missing a lot. He should take other perspectives to make sure that his function is serving all the company's interests.
What perspective should he take?
He should seek steady improvements. He should maximize the utilization of facilities, inventory, equipment and employees. He should maximize productivity. He should minimize rejects. He should minimize scrap. He should get peak performance from his operation every day and he should open his mind to creating changes in his operation.
Process control, while usually identified with engineers, is a discipline that lends itself to self study, on the job training and practical experience. By utilizing process control techniques, managers will raise their own value and the value of the operation they control.
In the first part of this article, I introduced the reader to concepts and wrote about monitoring. The reader should have come away from that section with a good idea about choosing meaningful statistics that relate to time, profit and predictability.
Now let's consider how to evaluate the numbers.
The first step is deciding what numbers are acceptable and what numbers should be investigated. A certain amount of variation should be expected and missing the expected output by 2% is probably not significant. On the other hand, missing expected output by 15% is certainly significant and should be investigated.
Each operation will be different and the manager should set a range of normal output that encompasses the average plus or minus some set percentage. (For the interested manager a range can be determined by applying statistics to calculate a mean, standard deviation and confidence interval.) Without Statistics knowledge, the manager can choose to make an educated guess about the acceptable range. Perhaps, the manager will set the range to encompass 90% of past data. (Setting the range too narrowly will cause many warnings and will result in investigating minor shortages. On the other hand, setting the range to widely will cause almost all the variations to go uninvestigated.) Once the range is selected, the manager will commit to evaluating every time period that falls below the bottom of the range.
As an example consider the following hypothetical set of daily outputs:
100 105 98 102 94 97 99 113 103 102 99 98 95 97 104 104 96 98 101 103
The sum of the 20 numbers is 2008 and the average is 100.4. Setting a range that would encompass 90% of the data could lead to a range of 94.4 to 106.4. Only 2 points, 94 and 113 would fall outside of the range and only 1 day (94) would require investigation because the manager would not investigate output on the high side All the other days would be within range and would therefore not require investigation.
On those few days when output fell below the bottom of the range the evaluation should be easy. If, for example, the power was out for 2 hours then the output would be understandable if it fell 25% below the mean. If a worker called in sick, then the output would be proportionally lower. And if the orders are below capacity then the output must decrease accordingly.
If, however, production fell below range for no apparent reason, the manager must investigate. Here's what he can do:
(1) Check each department (or worker) for the source of the shortage.
(2) Check Quality Control for an unusually high amount of rejects.
(3) Check resources (facilities, equipment, inventory and employees) for slowdowns or problems.
(4) Check for bottlenecks.
(5) Check for stumbling blocks (jobs that just won't run right)
(6) Check for missing material, tools or supplies.
He would record his findings on a (hypothetical) chart as follows:
Process Control: Comparison to expected output
Expected Output = 100.4 Average
Acceptable Range = 94.4 to 106.4
Date# Units Produced # Units Expected rangeInvestigate if under range1/10/0510294.4 to 106.4No1/11/0510894.4 to 106.4No1/12/059894.4 to 106.4No1/13/059794.4 to 106.4No1/14/059394.4 to 106.4Yes (See below)
Investigation of low production:
DateExplanations:1/14/05 Production came in at 93. Tooling department was backed up causing machine #2 to sit idle while waiting for a needed tool. Production would have been 7 units higher had machine #2 been operating normally. Each unit sells for $1000, so the company lost $7000 in production.
Over a period of time the production manager will build a history of shortages and explanations. The manager may make on-the-spot changes or may choose not to react to each problem. Over time, however, the information may suggest that changes should be made to resources or processing or training or work flow or vendors or inventory or inspections or paperwork or corrective action.
These changes fall under the third and final part of process control: Modifications. Now the process control system comes full circle and the modifications should begin to improve the company's record. As they do, the number of shortages should begin to get fewer and the acceptable range can get tighter. The overall results will be a more efficient operation with better information and less variation.
A few closing points:
1) Process Control can be applied to production, sales, quality control, paperwork processing, design and shipping. Each process control application will improve each function over time.
2) Process Control has an added benefit not covered by this paper: it can be used to test out new procedures, changes in processing and changes in resources. It can answer the question: What will happen if we change _____________?
3) Consultants that specialize in business improvements can help small companies with process control forms, procedures and training. Often a consultant will save time and needless trial and error.