Introducing AI into your business doesn't mean heavy investment. Get started with BCN's AI Kickstarters today.
View Solution
layer 1 layer 2 layer 3 layer 4 layer 5 abstract shapes

How to use Statistical Process Control (SPC) to make better decisions in the Boardroom

Posted 2nd July 2024

With increasing amounts of process data to collate and understand, organisations need a mechanism for quickly spotting potential issues. In this blog post, BCN’s Quality Improvement expert looks at how SPC can help drive ongoing process improvements, service quality and cost enhancements…


Today, every business is under constant pressure to improve their performance. Whether in the service sector, manufacturing, healthcare or any other industry, at some point most businesses will experience failure due to poor process performance. But finding and minimising the causes of poor performance is often challenging, and takes both time and money. Having mechanisms in place that can help us recognise changes in process quickly is therefore essential and can help leaders and decision-makers take assurance, as well as identify where to focus their improvement efforts. In the longer term, this can help organisations deliver ongoing performance improvements, better quality outcomes and cost savings.

The post-pandemic data challenge

In 2024, organisations face several challenges in ensuring they have effective mechanisms in place to quickly recognise changes in process. This is especially true since the pandemic, as organisations increasingly operate within complex environments and generate increasing amounts of data. Without the right tools in place, staff, managers and leaders can find it difficult to make sense of what this data is telling them, and to understand how their systems and processes are performing.

To make matters more difficult, few of us ever receive any formal training on how to understand and interpret data. Basic statistics rarely features beyond school or specialist further education courses, yet board members, executives and healthcare managers are expected to produce reports and analyse data, before making important strategic decisions based on these analyses.

In this environment, and often with a longstanding culture of reporting within healthcare organisations, data visualisation and understanding has been simplified and presented to decision-makers in the form of summary statistics, rolling averages and comparisons between time periods. These statistics and comparisons are further simplified by being colour coded – red (bad), amber (ok) or green (good) – to help decision-makers quickly identify potential problem areas (red) so they can focus on the areas of the business that are not performing according to expected target levels and require urgent intervention. Read more the benefits of shifting from RAG tools to SPC charts.


RAG tools: Simplified, but often too simplistic

It’s easy to see the benefits of such a reporting mechanism. It helps those who produce the reports to summarise large amounts of data quickly, as well as to present it in a fashion that doesn’t overwhelm decision-makers. Further, it encourages those who are tasked with reading and understanding the report to focus on areas of the business that may need further support.

But the use of colour-coded performance metrics also has its limitations. For one thing, it can provide false reassurances to boards, executives and decision-makers by masking potential issues until they become problematic. Alternatively, it can also lead to over-reaction on small and fleeting issues, that requires additional work and interventions from managers and team members. Both situations can lead to wasted time, effort and money for the organisation. And this can result in poor customer satisfaction, loss of business, and in healthcare, harm to patients.

Take the below example. This was previously presented to an organisation’s board and executive team.

The accompanying narrative for the data table stated:

‘Data for March indicates that there has been a significant improvement in performance of complaint resolution. 94.4% of complaints were resolved within 3 weeks, exceeding the 90% standard. The indicator is now within the standard for the first time in a year and is the highest it has been for the past three years.’

The reaction of board members and executives was to congratulate the team for such an impressive improvement in performance. As the narrative suggests, performance that month was the best it had been for three years, and this was a real cause for applause.

Naturally in this situation, thoughts turned to how those levels of complaint resolution could be maintained, and one of the board members commented: ‘Performance improvements are fantastic, but how do we know if we can sustain this level of performance going forwards? Can we assure ourselves?’

The drawback of current reporting mechanisms in many organisations is that they try to summarise too much data for boards. And within those summaries, boards, executives and leaders are left to try and interpret performance based on several colour-coded data points that don’t always give the fullest or clearest picture.

What’s more, very few board and executive development programmes offer the sort of formal training or education around analysis of data and statistics that would empower effective use of these charts.

In these circumstances, organisations could consider using Statistical Process Control (SPC) charts to support more informed reporting and data-driven decision making.

A more nuanced method of data reporting

The first SPC chart was created 100 years ago. It was designed specifically to help decision-makers understand the variations and performance of systems and processes, and to better predict future performance in certain circumstances.

SPC charts have since been developed further, but they have remained grounded in statistical principles that enable us to distinguish between two types of process variation:

Common cause: The natural fluctuations that are inherent within any process

Special cause: Variations that arise from specific, identifiable factors, and which should be investigated. The identification of Special Cause variation is based on specific rules that can help in quickly identifying[1]:

  • Outliers – data points that are much higher or lower than would be expected according to the data itself (and not to targets set by management).
  • Trends – when 6 or more data points in a row either all increase or all decrease.
  • Shifts – when 8 or more points in a row fall either above or below the centreline.

The key is to use these charts to understand what type of variation exists, so that the appropriate management strategy can be employed. Neither type of variation is necessarily good or bad. However, when special cause variation is present, it generally suggests the system is unstable or unpredictable. The important thing is to investigate the cause of the variation. Trying to redesign the process without that understanding is unlikely to lead to any improvement, and in fact may result in worsening performance.


Listening to the ‘Voice of the Process’

When only common cause variation is present, the system can be described as stable and predictable. This means that if the process or system remains unchanged, it will continue to vary between the upper and lower control limits, known as the ‘Voice of the Process’.

If it is not performing at expected levels, it is essential to change the process with a programme of improvement.

Returning to the example above, the performance in March hit target for the first time in the preceding 12 months and was the best performance in the last three years. But why? If we switch things up and plot the data in a time-series format and SPC chart, we get much more information than we did from the original colour-coded chart.

[1] There are several sets of rules. For further information, consider reviewing other texts on Statistical Process Control

The chart above shows performance over the previous 12 months, along with the target (grey hashed line). It shows us that in the last month (March), performance was above target, but that over the year, data varies between about 43% and 94%.


SPC helps identify important changes over time

If we take data from the previous three years and plot it as an SPC chart, this reveals even more for the board and executives to consider.

There are a couple of things to note on this chart:

1. Seasonal change in performance

While not an indication of special cause, there is an increase in the performance in March 2018 and March 2019, suggesting an annual seasonal pattern. This may be coincidence, but it certainly warrants further investigation. For example, performance improvements are sometimes seen towards to end of a fiscal year, so could it be a budgetary injection that has improved performance?

2. Evidence of special cause

There are 8 data points below the centreline from July 2017 until February 2018. This is known as a ‘shift’ and is evidence of a special cause. The cause for this is currently unknown, but it is definitely something that warrants further investigation.

It’s sometimes important to recalculate control limits to account for special cause variations, and if we do that here, we end up with the following chart.

With the chart updated, there are a few things to consider:

  1. There is a clear change in the performance of complaint resolution after March 2018. Prior to this time, it was resolving on average around 60% of complaints within 3 weeks. However, the control limits are quite narrow and, at that time, the process was operating predictably and could produce performance anywhere between 43.3% and 74%.
  2. In March 2018, performance clearly improved significantly, before reducing again. There are no further signals, so the process is stable and predictable. Note, however, that the control limits are much wider. This means that with no further change or intervention, the performance of complaint resolution could be anywhere between 27% and 100% without triggering action. While there are clearly months where performance achieves or nearly meets the target, the amount of variation is considerably more.

So what does all this mean for the board and executives?

Deeper insights support better decision-making

Based on the updated SPC chart, the board and executives can’t be confident that performance will remain above the target level. They can, however, confidently predict that future performance will be somewhere between 27% and 100%.

In the chart below, we have added the performance data for the following two months, and we can see clearly that, as predicted, performance varied and went down.

Importantly, it is clear that something happened to in or around March 2018 that altered the system and massively impacted performance. While we can see that average performance improved, the amount of variation also increased enormously. What caused this change requires further investigation, because it has led to a massively more variable process, which is likely making it more difficult to manage, more difficult to predict, and more expensive to run.

Supporting the discussions that count

If we bring the original RAG visualisation back into play, we can predict that, if the board met in May and reviewed performance, they’d probably be asking for an action plan from the team to address the rather concerning drop of more than 38 percentage points within two months.

All the board would be responding to, in fact, is common cause variation, which cannot be seen within the table above (which only covers three months) but is clearly visible on the SPC chart (which covers the preceding 12 months). Redirecting resources and staging interventions on the back of the colour-coded chart will require work, time, effort and cost from the team managing complaints, but it won’t resolve the underlying issue.

By using SPC, the board and executives can have a better, more meaningful conversation around:

  • Helping the complaints team develop a thorough understanding of what changed performance in March 2018 and led to increased variation (for example, were there budget changes, new technology systems or even new management personnel?)
  • Developing an understanding of what support the complaints team needs in order to reduce variation and improve performance to consistently meet target levels.


Empowering boards to making ongoing improvements

Presenting data as SPC charts doesn’t tell us what’s happening or what needs to happen. It doesn’t give us the answers. But what it does do is prompt discussion on a different level and support conversations that lead to a deeper understanding of processes. While SPC charts can initially seem daunting, with the right training and support, the rules are easy to learn and quick to master. With SPC charts in place and supporting discussions, boards are empowered with better and more meaningful insights around what impacts processes and how we can improve them, consistently, in increasingly turbulent times.


EasySPC is a Microsoft certified Power BI custom visual

Simple, Powerful, and Compliant

About EasySPC down down down