As an engineer, we love numbers. We love them so much that sometimes we think too highly of them. We have become so familiar with our statistics and calculations, that we forget that not all of the numbers we have been provided with are useful. So we greedily input every number, and as a result, we obtain incorrect information unknowingly and furthermore, we are surprised by the inaccuracy. How have we become so blind to the measurement of the uncertainty that we love so much? It is because the numbers are not enough. This idea is described in an article by Brandon Theiss called “Numbers are not enough” found in volume 43, number 9 of Industrial Engineer magazine. The dilemma is described in the author’s words in the excerpt below:
"For data to be considered 'quality,' it needs to be accurate, timely and complete. 'Quality' is being used in the context of Joseph Juran’s definition 'fitness for intended use' and W. Edwards Deming’s 'meeting or exceeding customer expectations.' To collect quality data, an organization must maintain a critical level of process intelligence at its most fundamental levels (the manufacturing floor, restaurant dining room, brokerage sales desk, etc). This intelligence can be manifested in the human capital of the operators or the physical capital tools and systems of production. The critical yet often sub optimally implemented — or completely omitted — tool of production is the data collection system."
The article begins by describing how technology has gifted us with the ability to obtain large amounts of timely data. The data is timely in the sense that the lapse of the time that occurred between the action taking place and the recorded data from that action being collected is an amazingly short amount of time. In the past, companies have relied on making improvements to processes based on data from days or weeks in the past. The further in the past an event is, the more irrelevant the data from that event is to current processes. Applying old data to analyze current processes is essentially a misapplication of industrial engineering tools that of which can be explained in the definition of the word quality. The author goes on to enlighten the reader to numerous instances of commonplace and widely practiced faulty data collection systems.
Because I love numbers too, I want to leave you with a little advice: The next time you find yourself in love with the numbers you’ve obtained, be sure to do a bit of investigation to become certain that the data you are putting in won’t give you the garbage that you don’t want out.