This article is based on findings contained in the author’s research paper titled “Statistical Process Control and Target Setting” which he presented (placing 5th overall) at the National Finals of the 1997 edition of the Nigerian Institute of Management’s Young Managers’ Competition in Lagos, Nigeria.

(Published Online: 1st March 2007)

Statistics Is Easier To Understand When You Apply It(In Conjunction With Logical Reasoning) To Activities That ALREADY Make Sense To You

Not many people are comfortable with mathematics, the sciences or statistics. For a number of us, these subjects are just too tasking, mentally, to be enjoyable enough for us to willingly study or apply. However, ever since I met a Maths teacher who took me in a remedial class after my secondary school education, and helped me score a distinction in the General Certificate of Education exams, I have - as he told me - discovered that it all depends on the mental attitude with which you approach any subject or activity. If you THINK it will be too difficult for you to learn, then it will be. If you believe it will require too much effort to understand, then it will! But if you tell yourself in your mind that you will learn and master that area or subject, then you will - so long as you stay committed to doing just what you have resolved to do, and do what is necessary to get the tight kind of support to achieve your purpose.

The above is the attitude with which I have always approached the subject of statistics. Rather than see it as complex and esoteric, I choose to see it as providing me access, at least on the basic level I prefer to use it, to a variety of simple, easy-to-use tools that can help me MORE accurately measure how effective and efficient I am in doing anything I am doing. I have always been keen on measuring performance with a view to exploring ways to get more done in less time, with less effort and at less cost. Evidence that this is true can be readily seen in the accounts - contained in many of my past articles - of my past achievements in and out of paid employment.

Probably the most recent proof that I am a Performance Measurement/Enhancement advocate is contained in my article titled "Write Great Articles That Get Read, By Computing The Readers’ Interest Index (RII)™ For Your Articles As A Guide! ". In it I share with the reader the formula for a performance measurement index - which I call the "Article Readers' Interest Index(RII)" - that i developed to help me measure - more precisely - how well my individual articles are doing with respect to "getting" Readers' Interest. I use the RII to more intelligently decide on what topics, themes or areas to write on next. Of course, I do not follow the lead of this index EVERY single time. I consider it against a number of other factors, apply some logical reasoning, and THEN decide what to do. What is important is that at the end of the day, the ability to measure the RII for my articles helps me take better informed decisions to improve my my articles writing/marketing strategy.

So, read this article with the above in mind. Don't expect any heavy technical stuff or complex jargon. I like to make use of statistics on the simple, quick and easy level. To facilitate this even further, I use my familiarity with spreadsheets(and spreadsheets programming) to enhance derivation of the measures I develop. That way I avoid getting bogged down with the technicalities of computing the statistics, so that I am able to stay focussed on making sense of the results I get.

How Does One Measure Performance?(The American NBA's Model)

Measuring performance is not about just using ANY statistical tool that catches your fancy, or which you've seen someone else use. You need to carefully study and select the MOST appropriate statistic(s) to help you evaluate the performance you are interested in. To explain what I mean by this, I have found it useful to use what I like to call the American NBA's Player/Team Performance Analysis Model.

Have you noticed that while watching an NBA game, the commentators(supported by footage on the TV screen), continually supply an almost endless array of player, match, game and team statistics aimed at enhancing the viewer's ability to reasonably evaluate the performance of the teams and player(s) in action relative to their counterparts on the court, and in the entire league?

Take a few examples. Field goals percentages, rebounds, three point shots, blocked shots, turnovers - statistics are available on a minute by minute basis thereby allowing viewers easily compare performances of competing teams, and players, from a number of different perspectives. A point guard may not do much scoring from match to match, but could be discovered to be behind his teams ability to score well whenever they play because he knows how to create chances for others to score. So his stats in assists, rebounds and blocked shots could be higher than those of other players while he might be poor when it comes to shooting three pointers and taking free throws.

The way statistics are used in the NBA there is often little room for doubt about who a team's most influential or valuable players are(It is important to note also how simple and easy to understand the stats used in the NBA are: By this I mean even if you're a newcomer to the game, you won't really need any major help to interprete them - they MAKE simple common sense!). The stats will often show where the strength lies. Sometimes too, the regular players may not do so well and a trend where players come off the bench to turn the team's fortunes around develops. Again, this when noticed helps narrow down to the "root cause". The coach, armed with the knowledge of each players strong points will decide based on the match situation who he fields and when, in order to increase the teams chances of coming tops.

An individual, group or organisation that wants to be able to constantly deliver optimal performances will find it useful to study the systems used by the NBA. I will now ask you to leave the basketball court and return with me to the confines of a business organisation that wishes to setup simple performance measurement and control systems to enhance its ability to deliver quality products and services to customers.

So, What Are Targets - And What Is Statistical Process Control?

I read somewhere that "Targets are a valuable instrument of management, indicating to the workforce what management wishes to achieve, and by when, in addition to conveying the measure of priority given by management to the activity to which they relate". I agree with that definition. That's why to make it more general, I have taken out the words, "management" and "workforce" and re-written it thus: "Targets are valuable instruments of performance management, indicating what is to be achieved, and by when, in addition to conveying the measure of priority given by those concerned or responsible, to the activity to which they relate".

Statistical Process Control(SPC) derives from a philosophy that acknowledges the need for any process operation to demonstrate a reasonable amount of stability (through control systems put in place) for it to be able to turn out products of a consistent quality.

I am not a statistics graduate. All I have is a B.Sc degree in Agricultural Extension Services(which required earning credits in a Research Methodology& Statistics Course in my third year), and an International Brewing certification. But I have always been passionate about putting anything I learn to use in real-life settings. That attitude led me to write/take a management research paper titled “Statistical Process Control and Target Setting(in manufacturing processes)” to a 5th overall placing at the 1997 NIM’s Young Manager’s competition. The paper was written based on extensive work I had done in establishing Process Control systems(in a brewery where I worked back then) and solving process problems using a few simple, traditional statistical analysis tools – including one that I developed myself.

In the research paper I presented at the competition, tools that I had used in my work in the brewery (especially when tasked to implement a project) were discussed. They included Schewart Charts, Coefficient Of Variation(COV), Standard Deviation and Mean, Regression Analysis, Correlation Analysis and CUSUM(Cumulative Sum Deviation) Charts - and Total Waste Unaccounted For(TWUF)™, which I derived in the process of carrying out a beer waste investigation on one of the brewery's product lines. (You can read my article titled "I Flopped Badly At The National Finals!(A True Story About How NOT To Prepare For/Deliver An Important Presentation)" to learn more).

How Do You Decide On Targets To Aim For?

Statistical analysis can be used arrive at a descriptive understanding of a particular process, thereby facilitating derivation of standards relevant to it. These standards would then become targets to be continually aimed for, periodically revised with a view to achieving continuous improvement.

Such targets can however not be arrived at arbitrarily. They must be based on good knowledge of the capability of the process, and a clear vision of the desired results you wish to obtain from it. It is in order to derive such realistic targets, that statistics become relevant. An example of how you can decide on targets to aim for is available in my article titled "Write Great Articles That Get Read, By Computing The Readers’ Interest Index (RII)™ For Your Articles As A Guide! "

But then even after deriving the targets, effectively monitoring - and controlling - process performance in relation to the set targets will also require use of statistics. Hence the term Statistical Process Control(SPC).

Lack of Performance Measurement And Process Control Can Drain Valuable Profits

I once worked in a manufacturing plant where huge product losses occurred on a particular bottling line due to bottle breakages at the filler machine after they had been filled with product. When the losses per day were computed, it became obvious that major financial losses were being incurred from just that point alone. A project team was set up to look into it. One key approach adopted was the use of simple statistical tools. See box below for a brief illustration (based on a real-life occurrence) that reveals the importance of Statistical Process Control, and how its neglect can lead to loss of huge profits.

No Two Processes/Activities Are Exactly Alike

One important note here. In the same company, no two processes - even when they produce the same kind of product - are exactly alike. Therefore what works in one may not necessarily work in the other. Just like humans, every process is unique – having its own original features and peculiarities. In order to be able to get reproducible and repeatable performances from an individual process therefore, we must gain a sound understanding of how best it works and why it works best that way. Just like for humans, we must understand a person’s psychology, to get him/her to give his/her best at work. The effective way to gain this understanding of a process is by using statistical analysis. If you understand the “psychology” of the process or operation you manage, you will know how it will “behave” under different circumstances. That knowledge will enable you decide how to manage it to produce the results you want.

TRUE LIFE CASE STUDY: Logical Reasoning In Evaluating Bottling Line Shifts Performances Reveals WRONG Winner Selection

In a bid to stimulate healthy rivalry amongst the 3 shifts operating on a bottling line, the bottling line managers decided to plot Volumes Sent To Filler(Hectolitres - HL) by each shift weekly using bar charts. A winner was declared by determining which shift had the "tallest" bar.

Ironically, it was the process operatives themselves who pointed out the inadequacy of this method of assessment.

Firstly, a shift(s) might end up doing more "hours" in a particular week than others, giving it an unfair advantage to chalk up MORE volume output; another might be beset with low attendance problems due to illness or special assignments of some members to other areas by management, putting it at a disadvantage. Other factors equally exist which have a bearing on a shift's performance, and which an absolute value like Hectolitres to filler would not reflect.

A better index of performance was therefore settled for: Hectolitres Per Man hour(hl/Mhr) obtainable by dividing the volume sent to filler by the product of the Total attendance and the Number of hours worked(8 hours). Compare Shift A (500hl with 28 men = 500hl/224 mhr = 2.2 hl/mhr) and Shift B (450hl with 22 men) = 450 hl/176 mhr = 2.5 hl/mhr. As you can see the shift with the lower absolute volume actually did better than the one with a higher volume output.

We could take it even further. In terms of the quality of work done, the quantity of low fills - which get counted as waste - generated by a particular shift for instance would also be used in assessment. If one shift achieves 500hls to filler, and in the process generates Low fills of 30hls for instance, it would effectively have actually sent only 470 (i.e. 500 - 30) hls to filler. Another way of looking at it is to express the low fills generated as a percentage of the volume sent to filler (or Full Beer Store in the case of the Packaging Department itself). This could be a useful measure of effectiveness of controls on (potentially wasteful) operations at the filler. In fact in the company where I worked this particular measure helped to temper the excesses of operatives who had a tendency to "tolerate" avoidable lowfills production in a bid to get as much volume as possible through the filler so as to "win" the competition.

Summary

People unconsciously develop fixed ways of thinking. When we get used to doing things in a particular way, we tend not to see a need to question the way we do them if only to be sure they cannot be done better i.e. assuming they are not being wrongly done. We need to move away from being subconsciously passive, and begin to challenge our ways of thinking and doing things more actively and constructively. The application of logic in what we do ensures that we avoid "bad reasoning" by drawing unwarranted conclusions from what is considered to be the evidence. Combine logical reasoning with intelligent use of simple statistical tools, and dramatic improvement in process(and people) performances can be achieved.

FINAL WORDS: (As Much As Possible) Keep Your Use Of Statistics Simple

Percentages, simple ratios, indices, charts of all types have always been most favoured, because they can be quickly and easily derived. They are also easily and quickly understood/interpreted by virtually anyone. Aim to use similar kinds of statistics for your performance evaluation and endeavour to apply logical reasoning at all times. You will find they will be readily adopted and put to use for the long term by all those concerned or responsible for the processes or operations. If you are a writer, and you would like to learn how you can compute the Readers' Interest Index for YOUR online articles, read my article which provides step-by-step instructions you can follow to set up a spreadsheet to do just that. It is titled "Write Great Articles That Get Read, By Computing The Readers’ Interest Index (RII)™ For Your Articles As A Guide! ".