What’s all the fuss about SPC?

I remember discovering THE paper about statistical process control, or SPC as it is fondly known. It was advice to analysts…..

“Fondly” might be stretching it in the case of these particular authors.

It was quite a lengthy tome. Among the technical caveats and impressive vocabulary was the suggestion that SPC wasn’t appropriate for performance monitoring or performance management.

There was heteroscedasticity. Mention of “suboptimal statistics”. There were reminders that SPC wasn’t a replacement for formal tests of significance.

I didn’t need to look up every fancy word to get the impression that the authors weren’t fans.

But “suboptimal”? For what, exactly?

SPC helps us understand variation over time, recognise signals of change and decide what action is warranted. Those are precisely the questions people running services need help answering.

So dismissing SPC as unsuitable for performance monitoring or assurance is quite a leap. Its practical value lies in helping us understand what a service is delivering, what we can reasonably expect and when something needs our attention.

Before we go any further, SPC isn’t the latest management fad. Its roots go back to Walter Shewhart’s control chart in 1924. That makes it 102 years old in 2026.

It has had a reasonably good run for something supposedly unsuitable for monitoring performance.

Mike Davidge first introduced me to SPC well over 20 years ago, when I was a Service Improvement Manager. On my latest road trip, I caught up with the great man over coffee to talk about spuddling—and how to put a stop to it.

Anyway, back then, if you’d come across SPC in the NHS, there was a fair chance you were in a quality improvement team. You used it to understand whether a change was producing the improvement you hoped for.

And it’s brilliant for that.

An SPC chart helps you distinguish the normal ups and downs of a process from a signal that something has changed. It gives you a better basis for judging whether your improvement project is making a difference.

But that’s only part of the story.

When the Making Data Count programme began, one of the less understood benefits was how useful SPC could be as an assurance tool.

Improvement asks: “Is our change making a difference?”

Assurance asks: “What can we reasonably expect this service to deliver?”

SPC helps with both.

For a stable process, comparing its expected variation with a target helps us understand whether we should expect to achieve that target consistently, sometimes, or not without changing the process.

You may recognise the shorthand: “always, sometimes or never”.

These aren’t promises. They describe what the current process gives us reason to expect.

Now compare that with our old friend RAG reporting.

Target achieved? Green. Target missed? Red. Somewhere in between? Amber, courtesy of a calculation someone once agreed in a meeting.

The next month, red becomes green. We congratulate ourselves. Then green becomes red, and we demand an explanation and an action plan.

Yet the underlying process may not have changed at all.

We’ve simply reacted to its normal variation.

Meetings are held. Reports are written. People are very busy.

Spuddling flourishes.

The problem with a RAG status on its own is that it tells us which side of a threshold a result landed. It doesn’t tell us whether performance has changed, or whether the service can reliably deliver what we’re asking of it.

SPC gives us a much better basis for that conversation.

It can help us ask better questions, investigate meaningful changes and recognise when improvement requires redesigning the service.

So what’s special about SPC?

It’s a brilliant improvement and assurance tool.

And if it helps us spend less time explaining every wobble and more time improving services, that’s less spuddling.

Which, of course, is a very good thing.