An important element in managing quality is ensuring that there is proper control of the processes involved in producing the product or providing the service. This requires set methods and procedures to ensure consistency.
Monitoring - collecting data - on these processes allows analysis to take place. Most businesses undertake some form of analysis already, even if it is only checking the monthly sales and costs figures. Collecting and analysing data can provide valuable information about the processes within your business; it has been said that if you can’t measure it, you can’t manage it.
Tally charts are merely forms by which instances of events can be recorded. A simple tally chart is shown below.
Remember to consider the data you wish to output when drawing up a tally chart. A useful site is on a display board or flip chart. The benefit of a tally chart is that it can be completed at intervals to suit the convenience of the user. In this instance, the typist merely places a mark in the appropriate column as each letter is completed.
Used in more sophisticated settings, tally charts can be designed to record data from production runs, frequency of faults, occurrences of deliveries or types of clients visited.
Bar charts are helpful in showing comparative data. The x-axis shows description. The y-axis shows the number of occurrences. A major benefit is the fact that it is a visual tool which displays clearly the information gathered. A simple example would be to compare typical production times (y-axis) for a range of activities (x-axis).
Example: Bar chart
A big issue these days, for many businesses, is the number of times the telephone rings before it gets answered. Many businesses are setting three rings as their target. You might, therefore, decide to monitor how quickly your staff are able to answer the telephone and then set out the results in a chart as shown below.
A histogram is a type of bar chart where it is the area, rather than the height, of each bar that is representative of its magnitude. The histogram provides a measure of spread. The bars always touch and the area of each bar reflects all the occurrences between the relevant scale points on the x-axis. An example will clarify this further.
Imagine, for example, that you make bread. The weight for a standard large loaf is 800 grams so you have decided that you will aim for every loaf you make to weigh 825 grams. If it is too low customers will complain - and if it is below the minimum requirement you might be prosecuted. On the other hand, if the weight is too high you will waste money. Although you have set a target weight, you cannot be absolutely certain that every loaf you make will weigh the same. You decide, therefore, to weigh every loaf for a certain period to monitor what is happening.
It is only one step from a histogram to producing a frequency distribution curve.
Many decisions in business depend on knowing the likely distribution over a population of, say, height or weight or length. Imagine that you run a bakers and want to know how many loaves of each size to stock. You can also use distributions to assist with setting up control mechanisms - say to monitor the weight of loaves of bread if you are a baker. Most distributions, and inevitably those occurring in nature, are 'normal' distributions.
We are not going to go into long explanations here of statistics, probabilities and sampling theory. Briefly, the average, or the 'mean' of a population with a normal distribution will be at the centre (see figure below). The standard deviation gives a measure of spread or 'dispersion'. A small standard deviation will result in a graph that is wide and squat. The area under a normal curve is always one (or, if you prefer, 100 per cent of the population or the sample population). One standard deviation on either side of the mean therefore covers 68.26% of the population, two standard deviations covers 95.45% and three covers 99.73%.
Example: Frequency distribution curve
Whilst the preceding sections may seem increasingly complicated, they are leading you towards the concept of statistical quality control. This is a practical use of sampling theory applied to a process - either to look at variables such as weight or length or to look at attributes such as satisfied or unsatisfied. Take a look at the case study for Bert’s Bakery.