Putting Big Data in the Right Context

What can you do with ‘big data’? It’s a hot topic, but the discussion mainly revolves around management applications. One thing that receives much less attention is the availability of process improvement software: analytical tools, techniques and software that can be used by operators, focused on the production process itself. 

Over the past 90 years there has been a huge digital revolution which was triggered by major advancements in information technology. Compared with a century ago, the most noticeable change has been that far fewer people are involved in today’s production processes. However, the few operators that do remain are being given ever-more responsibilities. What do they need to be able to do their jobs as successfully as possible? That’s the question that Aad Eikelenboom is striving to answer and for which the company QiSOFT offers a solution. But first, let’s take a step back in time.

Background

In 1924 the Control Chart (also known as the Process Behaviour Chart) was developed by Dr Walter E. Shewhart. It was based on the scientific theory that reducing variance in the production process plays an important role in a company’s survival. “Variance is created by repeatedly making adjustments to a continuous production process without placing them in the right context. And that’s highly undesirable!” comments Aad. “The Control Chart became the heart of statistical process control (SPC) since it enables the operator or manager to assess whether a situation significantly deviates from the normal situation. It is a graph showing the measurement of a quality criterion over time along the horizontal axis and the value of that quality criterion along the vertical axis. By monitoring the quality criterion over time, we can decide whether a process is running in line with the statistics or not.”

Notably, Shewhart’s teachings from almost a century ago are also applicable to modern-day big data. Data has no meaning whatsoever without context. Another finding from that period that remains relevant today is that, within data, the noise must be kept separate from the signal. Aad: “So what we chiefly want to do is obtain information and context from the jumble of big data. But how can you access information that is relevant, directly available in the production area and specifically intended for the operator?”

‘How can you ensure that operators can directly respond to process changes?’

Core issue remains unchanged

The paper-based Control Chart has since been replaced by a paperless overview on a computer screen. But despite all the ultramodern information systems and big data, the core issue in the manufacturing industry remains unchanged: ‘What contribution have you made to your business continuity today?’ “It’s an important question at every layer within the organisation, from the shop floor right up to the management,” believes Aad. “It comes down to the fact that big data only adds something to the process if you put it to good use,” he states. “The data must be placed in the right context. The internal specialists – or often, in the case of smaller companies, the external specialists – with the expertise to apply 6-sigma, lean, TQM, SPC Zero Defect, ISO9000 and other such principles are generally not available 24/7. It’s unrealistic to expect operators to be able to manage, apply and monitor all that themselves, round the clock, not least because they don’t have time on top of all their other activities. Therefore, many companies struggle to answer this question: ‘How can you create a situation in which the operators can directly respond to undesired changes in the process based on real-time data?’” 

What does an operator need?

A lot is expected of operators: they must deliver high quality, often in line with specific targets, and comply with all manner of legal requirements and standards. Furthermore, they must make a transparent and measurable contribution to the company’s success which is why key performance indicators (KPIs) are so widely used. However, KPIs alone are not enough. What production operators need is data that facilitates direct action and continuous improvement. Therefore, the million-dollar question is: ‘How can you reduce the mountain of big data down to relevant information for the operator?’ “Joseph Juran provided the answer,” explains Aad. “Many people regard him as one of the founders of the quality movement and of discussing management involvement in the quality improvement process. He stated that ‘capability precedes empowerment’. In other words: train the employee so that he is able to do his job properly, ensure that he has the right knowledge and is authorised to make sound decisions.” 

Profit

“To achieve that, the focus must not be on complex statistical methodologies,” continues Aad, “but rather on making statistically sound information immediately available for the operator. Process improvement software provides that information. The resulting visible variance and the insight into its effect on KPIs enable targeted and effective management on the shop floor, both in the short term and the medium/long term. If companies can achieve that, they will improve not only their processes but undoubtedly also their profit.”
 

www.qisoft.nl