2F12 Do You Use Data to Improve Performance?

© Alan C Clark 2017

© Alan C Clark 2017

Have you ever wondered why it seems that so many companies find it difficult to improve performance or do not even bother? Experts say all you need is the latest tech or app because they seem to want you to believe that is the problem. But the facts shows us that is not quite true.

Even in the early 2000’s when the tech was not that bad, research published by Cardiff University showed that only 5% of the things they did in manufacturing added value for customers! It was worse in information-based industries where only 1% of activity added value!! Consider then that in Toyota 20% of activity is value-adding! Wow! Just think what improving from even 5% to 20% will do to your company’s performance. The problem really is that many companies are unaware of this enormous potential that there is to improve their performance, whatever sector they are in.

Understanding what is actually happening in a company is the key to improving performance. You can get this understanding most powerfully by properly interpreting your performance data. Too often any performance data that is collected is used to ‘beat up’ people. The intimidation becomes worse when s-t-r-e-t-c-h targets are attached to performance data of, which is utter nonsense.

So if the key to performance improvement is using data to learn what is actually happening, how do you do that? Well you need a method or process that clarifies what is actually happening, which will help you to take the most effective improvement action.

First it may be helpful to realise that that there are different levels in the the learning that you can get from measuring performance. In his book, Re-Creating the Corporation: A Design of Organizations for the 21st Century, Russell L. Ackoff set measured performance data as the lowest of five levels for learning. In increasing power they are:

  1. Data is numbers and words representing objects, events and their properties.
  2. Information is contained in answers to questions of who, what, where, when and how many- it is processed data.
  3. Knowledge is contained in instructions, know-how – you might think of it as organised information.
  4. Understanding is contained in answers to why questions.
  5. Wisdom is being able to recognise and judge long-term possible consequences. Wisdom is also said to be knowing when and how to act effectively.

So what process can you use to increase your level of your learning and hence performance? Tabulating data will not do it because it is just presentation. Bar charts are again presentation, but certain types can provide some insights. Pie and doughnut charts… hmm, risky and may lack clarity. Personally I wouldn’t go there and prefer bar charts.

To use data to improve performance requires a process to analyse it. Data analysis authority Dr Donald J. Wheeler has long advocated what he calls the process behaviour chart (PB chart), which both presents and analyses data at the same time. Some people also known it as a control chart, though it cannot control anything, and also as a Six Sigma chart. The latter comes with some unnecessary complication.

Let’s look at an example of a PB chart, which keeps data in time order sequence and adds three decision lines. The x-axis is in time or date units and the y-axis in in the units of performance data.

The first ting to notice is that by keeping the data in time order, particular data points or patterns in the data can be related to real-world events. These may be a signal for corrective action and or of learning about the business system or process that leads to an improvement action. Visual presentation of the data makes this easier than tabulations. There is almost always point-to-point variation as this set of real-world data shows, i.e. some days are better than others. If your improvement action can reduce this variation you can improve your total customer experience by delivering more consistent products or services.

The green decision line is the average of the data. The two red dashed lines show the limit of variation if, and it is a big IF, the business process is stable and predictable. The data will go on randomly varying about the average and between the two dashed lines. If it is stable the chances of a point outside these lines is would be very, very low. Unless, that is, something changes in the environment or the process or you make a deliberate change to the process. In the chart above in week 30 the data point is actually on the line and so this could be a signal of a change and so an opportunity to learn.

The thing is, if you try to take action based on tabulated data you are missing all sorts of important learning and really could be just guessing. This is as true of food manufacture, as it is of airline services or sales processes. If you use a PB chart you will learn:

  1. If it is safe to take action on or to make a prediction from a set of performance data from a business process.
  2. You can separate signals for action from the noise or random variation in performance data.
  3. You will be able to tell whether any process change that you have made makes an improvement or not.

So you see just producing a table of data is misleading if you think you are really monitoring performance. By putting your data in PB chart you both present and analyse data providing an evidence base for performance improvement. You can get all sorts of specialist software to do this, but you can also use a spreadsheet. It is also so simple that you can even use good old pencil and graph paper. Good eh?

In the next blog we'll continue to expand on the 2F diagram by looking at processes for understanding the situation that you are currently in, drawing from my book Picture Your Business. This is available from me priced £12.00+P&P. The latest version of the companion 2F Worksheet is now available free in MS Word or PDF on request from me either via a Linkedin message or by emailing alanclark@keybiz.com.