Please take a look at the following two minute video. It’s a good introduction to my stories about the use of data in the workplace.
Another experience I like to use as an example in measuring the process and improving it happened while I was employed by a manufacturing and distribution company in Nebraska. They also had several convenience/gas stations. One day the individual who was responsible for managing them came to me with a tale of woe. He felt there was a need to terminate one of the employees at the western most store. He was not sure if it was theft by the employee or just incompetence on their part that was causing the shortage in gas sales. It appeared to him that when this person worked more gas was pumped that was rung up on the cash register. He really wasn’t looking for me to solve his problem but just venting his frustrations before he went and fired the employee.
I offered my help if he could put off talking to the employee for a couple of days. He was open to it since the employee was on her days off but felt that he would need to do something because he felt his neck was on the line with his boss.
I went to the accounting department and asked for the daily financial records for that store going back over the past two years. Then I spent several hours putting the daily gas shortages for the past two years in a spread sheet then compared them with the work schedule during that same time period.
Once I had the data in the computer it didn’t take long to discover that the shortages seemed to happen to everybody working at that store, not just the one employee he was focusing on presently. Some days the gas quantity pumped matched the amount indicated on the cash register tapes and other days there were large discrepancies. This raised more questions for me so I asked the manager if I could go out to that store the next day and observe the process. He had no problem with it so I called the store staff and let them know I would be there bright and early the next morning to learn how their business worked. What I found the next morning still causes me to chuckle and shake my head.
When I arrived I explained to the counter person I was there only to observe and would not be involving myself in their process. If I was to help out I would be changing the process and it would not be working as before. I hoped they would not see me as a problem or threat but only as a person trying to learn.
Initially the traffic flow at the counter was light and well under control. I did notice when some gas pump sales were made the counter person would walk to the storefront window, look at the pump readings outside then return to the register and complete the transaction. This step happened every time a particular pump was used. When things slowed down I asked the clerk about the need to look at the pump readings. She said that the last pump in the series would not always give the correct readings inside so they had been told to go look at that pump’s readout and verify the correct amounts. She said that occasionally the readout inside the store would be lower than what was actually pumped.
Eventually the customer traffic picked up and a line formed at the counter. What I observed was that when a customer had used the pump with the bad readouts the counter person would not verify it by leaving the register and looking out the window. She would ask the customer if the readout on her machine was correct and they always said, yes.
Again when things slowed down, I asked her why she didn’t verify the amounts on the pump with the bad readout each time. She said that there just wasn’t time to do it when the customers in line were anxious and waiting to pay and get on the road. It made sense to me.
Just before lunchtime the lines started again so when she didn’t verify the pump reading on the last pump I casually walked over to the storefront and verified the pump readings. The amounts on that pump and the inside readout were the same.
I had several opportunities that noon to check out the pump with the bad readings and found that on two occasions the readouts inside were several dollars less than what the pump actually read. Those customers had told the counter person that the smaller numbers were correct, paid their bill and left. It looked like those two customers had figured out how to beat the system/process and buy gas at a cheaper price than others. I wondered how many other customers were beating the system but I did not have the time to hang around for several days to find out.
On Monday, the store manager and I sat down together and I described to him my findings and asked why he had not had the pump with the occasional erroneous readouts repaired. His response was that he had tried to get it fixed but the cause could never be located and the owners were not willing to spend the money to purchase a new pump. He understood that it was the process that was at fault. He worked with the store supervisor to have two people at the counter during the heavy traffic times so one of them would be available to verify the pump readings.
A couple of months later, I ask the manager how things were going. The smile on his face told me the whole story. Gas revenues were up at that store and the shortages had almost completely disappeared. He also had not needed to sit down with his good employee and tell her she was going to have to find a position with some other company.
An important aspect to learn about measuring a process is to understand that every process, no matter how rigid it appears to be, has some variation. Try to visualize a line graph of the stock market. Every day a measurement of the Dow Jones Averages is made and plotted on the Y-axis with the corresponding day of the week on the X-axis. As you have seen, when the points on the graph are connected, it appears as a very irregular squiggly line, going up and down and from left to right with the movement of the stock market. Lots of companies and individuals believe they know the reasons for all the up and down movement but there are so many factors involved in causing the daily variation that actually many of the reasons people give are really only simi educated guesses. Fortunately for most of us, when we try to measure our business process we are dealing with only a few or handful of variables.
Identifying a process’s variables is important to improving it, if that’s what is desired. Most business owners want to reduce the variation in their processes so that costs can be reduced and profits improved. To do this, the variables in the process must be identified and measured, so that when one of them is modified, the effect of that change can be identified and measured. The less variation a process has in it, the better it can be used to predict future performance.
Another event from the past that is a good example of measuring the process to help understand the present and predict the future involved the same convenience store and manager. One summer he approached me very upset and afraid for his job. Things had been going well at his stores until the May financials came out in June and he was going to have to explain the reason/s to his boss.
The spring numbers had been showing a nice increase and he was expecting May’s to look good also. When he received them they were off sharply and even down slightly from the previous year’s. He could not figure out why and knew he would need to be able to explain to his boss why they were down, so he came to me for help.
I retrieved the master financial book for the two stores and put four years of monthly numbers in a spreadsheet, then ran a line graph. What it showed us was that May had always been a down month and that this year was no exception. You could see the relief on his face as we looked at the trend over time. I tried to impress on him that now that he knew May was always a slow month (at least for the years we looked at), what was he going to do in the future to make it better? I’m not sure if he heard me or understood what I had said but he did take the figures into his boss and explain that May was expected to be a down month. His boss accepted the explanation and business went on as usual. I was not around the next spring so I don’t know if he ever discovered why May was down or tried to do anything different to improve sales.
One morning, many years ago, while pedaling away on a Schwinn air-dyne stationary bicycle at the YMCA, a thought came to me about how management responds to certain situations. After going over in my mind, a couple of examples I have seen in the workplace, the words, Knee Jerk Management, came to mine. The following are a couple more examples.
Each year in the small town where I live the enrollment figures for the local community college are published in the newspaper. If the full time equivalents are up over the previous year the paper reacts with very positive headlines but if the enrolment is down a few it sends out a very grim forecast about what might happen to the college; implying that the eighty-three year old institution might have to close its doors. The next day’s paper will have several articles criticizing the college administration and demanding to know what they are going to do to get the headcount back up where it belongs. So, the administration goes about trying to guess why the enrolment is down and what new solutions and programs it can try. No one seems to want to look at the enrolment figures over time. They just seem to focus on the current year’s data verses the previous year’s data. What they need to do is a good analysis of many years’ of data, of course, understanding that the old data may not be collected using the same methods how a full time equivalent was defined. Once the data is collected and the common causes of enrolment variation determined, then and only then, should a hypothesis be made and action taken. One of the mistakes many managers make is treating a common process cause of variation as if it is a special cause of variation and making some profound management decision to immediately fix the problem. And just cause more variation in the system.
Another example I thought of while pedaling away was that of a regional public power utility who reacted to their board of directors concern about their bottom line. The utilities biggest users of power were the farmer’s irrigation pumps during the spring and summer months and electric home heating during the winter. If there is a mild winter and wetter than usual spring and/or summer the farmer’s need for electricity is lower. If Mother Nature did not cooperate and go along with their forecast they might very well reserve too much or not enough electricity causing their costs to be much higher than expected.
Most of the utility board members are good old farmers from the district. They have no special training for the position except that they are farmers and well intentioned. During a board meeting they might express their concerns about the bottom line looking pretty grim and suggest that management lower its overhead by cutting staff. Management reacts by trying to consolidate job duties and letting several people go or offering those near retirement incentives to retire. What happens the next couple of years is that the climate returns to its normal dry conditions and the winters are cold and snowy. Now the utility can’t keep up with the demand for increases in service calls and office demand so the staff positions deleted two years before are filled once again. Have you started to see what is happening? It eventually becomes a vicious circle of climate changes, higher or lower electrical demand and fewer or more employees.
One final example that comes to mind is how the Bell System use to handle variation in their telephone system. By the time I joined Southwestern Bell in the fall of 1977 they had been delivering quality phone service throughout the country for many years. They had developed a very good understanding of measuring data and building phone offices and long distance networks to precise standards. They had a thorough grasp of variation (common and special causes) within their systems and how to build economical systems so that offices and long distance trunk groups were not engineered to peak demands, to do so would cause rates to be higher than necessary.
It was easy for the telephone companies to measure customer traffic going through their offices as well as what was traveling over their trunk groups between offices and long distance networks. The data they collected was used to determine how much equipment to put in an office and how many circuits to put between offices and different cities. I know from working within the company that it was very important to size an office or trunk group with just the right amount of equipment for a certain number of years.
Since it was not practical or economical to oversize a trunk group they were engineered to the busy hour of a busy day. Usually that was some hour on Monday if the office primarily served a business area. Where as if it was primarily a residential office it might be an evening hour. What the telephone companies tried to do was balance the traffic (business and residential) through an office or trunk group. They also offered lower long distance rates per minute if calls were made during the late evening hours or on the weekend. What they were trying to do here was level out the range of variation on long distance circuits. In other words, getting people to call on the weekends and at night when businesses normally were not open and making long distance calls.
What the phone companies did not do was put in enough circuits to handle all the calls for days like Mother’s Day, Christmas and the other holidays. If they had, many of the circuits would have been sitting idle most of the year except those few special days. Eventually as toll switches became computerized the different time zones across the nation and globe allowed calls during early busy hours on the east coast to be switched west then back east over a different trunk group. An example of this would be the traffic from New York City to Miami. At that trunk group got into its busiest telephone time period, say 9 to 10 am some of the calls might overflow to Denver then back to Dallas, Houston, New Orleans and finally Miami. Their goal was to keep as many of the circuits in a group busy and not deny (give a busy signal) to a revenue producing call.
One of the key things with the Old Bell System was that they were always trying to improve their services and costs, in today’s environment that would be called continuous process improvement. Of course they had good reason to want to improve their costs and service. Blocked calls caused customer complaints and cost increases forced them to file for higher rates with the utility commissions, neither a pleasant task. It seems that most people get set in their ways (rigid processes) and only seek change when things get so uncomfortable for them that they seek relief by looking for ways to improve their processes.
What I have been talking about in the above examples is called Statistical Process Control (SPC). The words really look complicated and may take you back to a course required in most schools of business called statistics. Actually it is a very small part of the course material but a very meaningful one, especially since it deals with everyday data that any business needs to be collecting and using. If you decide to look into using SPC then I suggest you go to the library, a good bookstore or the Internet and get more information on it and spend time learning. What I’ve given you here is just an introduction, something to wet your appetite, but don’t take it lightly because it can help you improve your business and your bottom line.
I’ve only mentioned SPC because the data I needed to help answer the manager’s questions in the above stories was all I had at the time. Measuring process is vital. Knowing what and where to measure can be shown by constructing a PROCESS FLOWCHART. Flowcharting will be the topic of my next blog.
The following is a short video showing how Deming got his start in Process Improvement.