It is common to hear in all types of organizations, concern about the fluctuation of income, sales, fluctuation in delivery times of goods, variation in the arrival hours of personnel, variations in financial budgets, variations in time service at a bank window, library or service institution. etc.
There is also concern as to why there are problems in the weight of the products, in the filling of containers, in the hardness or viscosity of the products, as well as in the dimensions. A teacher worries because there is variation in the grades that students get.
There is a permanent requirement that all salespeople reach the goals or the students the minimum marks. Budgets are expected to be as accurate as possible. Today data-driven management has become fashionable. The Balanced Scorecard recommends indicators, goals, results, continuous improvement, performance evaluation. ISO in its data analysis clause states that the organization must determine, collect and analyze the appropriate data to demonstrate the suitability and effectiveness of the quality management system and to evaluate where the continuous improvement of the quality management system can be carried out. quality.
On the other hand, after establishing indicators, what happens? There will be a lot of data, but little information. If the goals that are set for the indicators do not consider the variation of the processes, it will be of little help, rather the frustrations will begin.
In all processes we find variation, and that is annoying; Despite the fact that variation is implicit in absolutely everything that is done and what surrounds us, it can hardly be avoided. Neither two hairs of the same head are exactly the same, nor are the two eyes of the same person, much less two people even if they are identical twins.
Why if the enemy of all processes is variation, whether in the area of Sales, Production, Accounting, Operations, Finance, Education, etc. and etc., in the companies it is not heard that something is done to control it. Quality systems are implemented to achieve repeatability of processes, products and services, we talk about continuous improvement, but none of them talk about the control of variation.
There is a maxim, "the enemy of all process is variation." A successful manager is one who manages to control it. The theory of variability is one of the four that Dr. Deming proposed to the Japanese within his philosophy of Deep Knowledge, another theory that complements the previous one is the «theory of causality», where he states that every effect has a cause, every defect too. The control of the variation can only occur in its causes, mainly in the control of its root cause.
The problem found is that it is accepted that there are problems with variation, but there is no interest in studying why it occurs or how to measure this variability.
It is very likely that you have taken at least one descriptive statistics course and one inferential statistics course. How much of what was learned there is used? How much is the standard deviation formula used? How much is remembered? How many can calculate it? What interpretation is given to its results?
Example of variation in processes
This example will show the importance of the standard deviation in daily life: «Consider buying a delicious pizza, which you order at the pizzeria on your way home. There are two pizzerias of which the following information is available regarding preparation times (in minutes), for 10 pizzas:
Pizzeria ABC: 6.5 - 6.6 - 6.7 - 6.8 - 7.1 - 7.3 - 7.4 - 7.7 - 7.7 - 7.7
XYZ Pizzeria: 4.2 - 5.4 - 5.8 - 6.2 - 6.7 - 7.7 - 7.7 - 8.5 - 9.3 - 10.0
Using common statistical tools, such as the mean, median, and mode, the following results are obtained:
Pizzeria ABC: Mean = 7.15 - Median = 7.20 - Mode = 7.7
XYZ Pizzeria: Mean = 7.15 - Median = 7.20 - Mode = 7.7
From these results it can be seen that both pizzerias have the same measures of central tendency; that is, on average, the customers of both wait for their pizzas the same time. Based on these results, it is difficult to distinguish differences in both processes to make any decision about it. Looking back at the prep time data, a greater variation (or dispersion) can be seen in the times for Pizzeria XYZ. If all the other quality characteristics of both pizzerias are equal, it is likely that customers prefer to buy their pizzas on the ABC, due to its more consistent and less variable prep times. In the business world, something more precise and reliable is required to measure and quantify the variation in processes; The following measures are available for this:
Range and Standard Deviation
Why is the standard deviation important? Because it is one of the simplest ways to control variability, call it budgets, sales, products, attention times and for the whole new set of indicators that are in fashion.
By means of the standard deviation, customer surveys can be analyzed and determine how conclusive the answers are, the probability that the goals are reached, that the products are out of specification, that an employee is late, that a student can be inferred did not achieve the minimum grade etc.
The standard deviation is used to know if the processes have the capacity to meet the customer's requirements, through what is known as Cpk analysis, or Cp, the correct formula to determine their capacity.
With the standard deviation it is possible to calculate the sigma level, in what is now fashionable and known as "Six Sigma", sigma is the Greek word for standard deviation. It is used to compare two totally different processes, such as a process in the Human Resources department and another in the plant or in finance, by using the coefficient of variation.
The standard deviation helps to calculate the natural limits of variation of a process, to establish the fluctuation in the sales goals and even to give permission to the personnel to be absent, as we will see in the following anecdote: In a consulting process in a bakery company We were meeting in the plant manager's office when a subordinate came in to ask permission to leave. As he had problems with the weight of the bread, the manager asked him to show him the weights of the samples that had been taken. With a few calculations (standard deviation, average and probability), he confirmed that the possibility that some product was under specification was remote, he said, "you can go easy."
Perhaps it was done to impress the consultant, but today more than 15 years later, the anecdote is remembered.
In addition to the calculation of the standard deviation, more complicated operations such as analysis of variance, measuring the precision, accuracy, skewness and kurtosis of a process can be done, but these analyzes can be left to the experts.
When Kaouru Ishikawa said that 85% of problems in a process are the responsibility of management, the comment was not well received. His assertion is based on the fact that variations in a process are generally attributed to normal causes, according to its designed capacity, which is the responsibility of management. The operator acts within what the process allows.
This leads to one of the best benefits in controlling variation: defining when it is typical of the process, something normal, caused by normal or common causes and when it is due to abnormal or external causes. Understanding the concept of variability and measuring it through standard deviation can establish normal limits of variation. (Usually plus minus 3 standard deviations), once these limits have been set, the process can be handed over in complete peace to the hands of subordinates, so that they can control themselves, as recommended by the philosophy of "empowerment". If the collaborators do not understand how to control the variability of a process, there is no procedure or instruction that will save it, even if they are certified.Rewards and punishments are useless if a process is varying within its normal causes. The reward or punishment becomes an abnormal cause, outside the process, where after its effect passes, it returns to the previous state.
Some of the concepts, of course, are subject to statistical rigor, but if at least this leads to understanding why a process cannot be controlled, why goals are not reached, another would be the situation of many companies. It is worth remembering that while in other times in the West novels were heard on the radio, the Japanese listened to statistics. Of course, other times, other means, just the same problems.
Additionally, there are another series of statistics that could be used to control the variation, the correctly used average is important, as well as the mode and the median, as well as the range, but we must not be so ambitious. Today, when most companies are managed by averages, the anecdote of the character is remembered who drowned in a river with an average depth of five feet.
