If you’re a manufacturer looking to boost efficiency, a process control chart is a simple but effective device to use. A part of the Lean Six Sigma toolkit, control charts will help you monitor and improve your production processes
So what is a process control chart, why are they so useful for manufacturing businesses, and how can you start using them? We take a look.
What is a process control chart?
A process control chart – also known as a Shewart chart or process behaviour chart – is a time-series graph that helps monitor the acceptable limits of a particular process. It’s a way of using real-world analytical data to spot when a particular process is starting to get ‘out of control’, so that you can intervene in a strategic manner.
What is the main purpose of a process control chart?
Process control charts are used to understand variation – and when to react to change.
Using a control chart helps to indicate when an unusual event has taken place, causing an unexpected change in a process. The process itself can be anything, from a form being filled out to complex machining – and everything in between. In other words, you can use a process control chart for almost any process you want to control.
When we notice a variation in a particular process and we react to it, our constant meddling may actually start to create more variation, not less – and this observation was one of the reasons why control charts came about.
Dr. Walter A. Shewhart discovered this phenomenon, and invented control charts in the 1920s. He used data to show that not all variation should necessarily be meddled with, because some variation will always be inherent in a given process.
The two types of variation a control chart displays
Dr. Shewart defined two types of variation:
- Common cause: Inherent variation that will always be there. This is created by the intrinsic variation of the process elements. There isn’t necessarily a ‘cause’ of this type of variation, and improving it would mean changing the process elements.
- Special cause: An assignable cause of variation. An unexpected event has occurred and created statistically significant variation in the process. This cause can be investigated, found and dealt with.
We’ll talk more about these two types of variation below.
What are the attributes of a process control chart?
A typical process control chart is shown as a line graph over time, with three core components:
- A centreline, which represents the statistical average of the data
- An upper control limit
- A lower control limit
Ideally, your process will always fall within the two control limits. If they do, and the datapoints appear to be random, then any variation could be described as an inherent ‘common cause’ variation and would therefore be defined as ‘in control’. Or at least, within threshold limits.
But what if the data breaks out of one of the two control limits?
In this case, an assignable variation is likely to have occurred, and this will need to be investigated. This may also be the case if the data stays within the control limits but begins to adhere to a non-random pattern.
Benefits of a process control chart
Using a control chart brings with it several benefits for manufacturers:
- Always know when to react. If you’re using data to control your process variation, you’ll always know when to react to change – and when not to. This can help limit the disruption of inspections, changes and tinkering.
- Build data that you can use to optimise over time. Using data over a period of time allows you to create benchmarks that you can measure future performance against. You’ll know what is optimal for your processes and be able to see when your current performance wavers from that baseline.
- Use a single language to communicate performance. With good data, performance is no longer subjective. You can accurately track and measure it, and communicate goals and ideas with other key stakeholders.
How do process control charts improve manufacturing quality?
In manufacturing, control charts help to ensure the production process remains in a controlled state.
In manufacturing the ideal is for processes to be stable and predictable. You want to know what output you’re going to get from your inputs, so you can plan appropriately and meet the expectations of your customers and partners.
A process control chart is one means to achieve that end – its very nature is to find the centreline and ensure the process remains in a predictable state based on that desirable average.
Manufacturers can use process control charts in three ways:
- Real-time monitoring of processes to ensure predictability of the end product
- Root-cause analysis of assignable variations: with a control chart, you’ll know exactly when variations occurred, so you can investigate with clearer direction
- Gaining a sense of confidence that there is real data governing the control of process variation. This is something that could also ease any concerns investors have.
Tips for implementing process control charts in your business
Almost any business can use control charts for processes they aim to keep as consistent as possible. Here’s what you’ll need to do to start.
Five steps to implementing process control charts
Generally speaking, there are five agreed steps to creating any process control chart:
- Clearly define which process you intend to monitor and control
- Figure out how you will measure that process. What metrics can you use, and what system will supply the data reliably?
- Establish your centreline and control limits
- Collect the data and plot it on the chart
- Examine the chart and make decisions based on your analysis
How to find your control chart centreline
Finding a centreline is relatively straightforward once you begin capturing data. What you’ll need to do first is choose a time period to measure against.
Think about how your business operates, and try to measure over a few different cycles to get the most variable data. Perhaps you’ll measure for a quarter, or an entire year (e.g. if your business is seasonal). Or maybe you’ll create a control chart per season, if these are significantly different.
During this time period, you’re just plotting data – essentially making a run chart. It’ll give you an idea of trends over time but not much else. However, at the end of the time period you’ll be able to go back and calculate what the mean average was out of all the datapoints.
This average becomes your centreline.
How to find control limits
To understand how to find control limits, let’s use a very simple example: calculating how long it takes to walk your dog.
- Example: walking the dog. Every day you walk your dog, you measure it. After three months, you know the mean average time it takes to walk your dog. We then want to figure out the upper & lower control limits. To do so, we determine our standard deviation, and then use Six Sigma methodology to turn those into upper and lower limits.
How to calculate control limits
To calculate control limits, start by subtracting the mean average from each day’s result. Then, square each of those daily figures. Next, find the average of each of the squared results. This gives you the variance of your data set.
You then take the square root of that new figure (variance), and this is a single ‘standard deviation’.
- Tip: In a spreadsheet or a tool using similar formulas, you can automate this by using AVERAGE and STDEV calculations to work out the mean and standard deviation.
If we’re using Six Sigma thinking, we’re going to have three standard deviations above and below our mean average as our control limits. A well-controlled process fits within three standard deviations.
- Back to our dog-walking example: We find out that our mean average dog walking time is 26.8 minutes over one month. By going through the calculation outlined above, we find that our standard deviation here is about 8.1 – putting our upper control limit at just over 50 minutes, and our lower tolerance limit at just less than three minutes. That wouldn’t be an excellent walk for the dog!
Control chart example: maintaining the consistent manufacturing of beer
Let’s say you’re a brewer trying to control the flavour of our beer. You have a unique product with its own signature flavour, which you get from adding special ingredients to your brewing process.
You can use a process control chart to measure this.
Your process control chart might rank ‘flavour strength’ on a scale of 1 to 10. If you continuously track your quality control tests score on a chart over time, you’ll start to learn what taste is our desirable average. If, for instance, that is 8 out of 10, that’s your new benchmark.
Once you have that average locked in, you can use the same data to learn what your acceptable upper and lower limits are – the points at which you can no longer sell the finished brew, either because it’s too weak or too strong.
If we keep inputting data and measuring it against this benchmark, we’ll generate a trends graph. That graph will show us the moment our flavour strength eclipses one of our control limits, at which point we can investigate the problem and – perhaps – halt production of that batch.
Because we know exactly when the issue occurred and in what batch, we can trace the variation back through to its root cause and determine why it happened – and most importantly, avoid it in future.
Process control charts versus statistical process control (SPC)
Statistical process control (SPC) and process control charts are often mentioned together.
Process control charts are in fact one tool in the wider SPC toolbelt. SPC as a whole is the use of data to control a process, and it includes a range of tools and techniques. Control charts are a big one, but so are histograms, cause-and-effect diagrams, process flowcharts and more.
Process control charts versus run charts
Run charts and process control charts are very similar, but the basic distinction is that the latter is a more statistically complex – and useful – version of the former.
A run chart is like a process control chart without the centreline or control limits. It plots data over time, allowing you to see trends in your process at a glance. That means you can use a run chart to monitor data over time if you don’t require deeper insight. For example, health and safety officers might use it to check form completion rates over a year. If they see completion rates dropping, they would know to intervene and find out why.
A process control chart has those extra data points – the centreline and the limits – which makes it more useful for spotting variations and making strategic changes.
Process control charts in Six Sigma
Six Sigma is often mentioned alongside process control charts, as a process control chart is a valuable tool for anyone utilising the Six Sigma management ideology.
What is Six Sigma?
Six Sigma is a management method that uses qualitative and quantitative analysis to control and improve business processes. It is both a methodology and a certification that one can achieve, with training involving both the core principles of Six Sigma and a variety of tools (like control charts).
The philosophy of Six Sigma is that all work is a process that can be defined, measured, analysed, improved and controlled – collectively known as DMAIC. Under this philosophy, SPC, control charts and more can be deployed to achieve each of these steps.
The desired outcome is a process that can be defined as ‘Six Sigma quality’ – that is, variations are controlled within plus/minus three standard deviations from the centre line in the control chart. In mathematical terms, Six Sigma quality means a process will produce only 3.4 defects per million opportunities.
Do I need to adhere to Six Sigma to use control charts?
The short answer is no. Any business can use control charts to monitor processes, without adopting the whole suite of Six Sigma techniques. Process control charts can be deployed on their own and can still serve a valuable function within your business whether you follow the wider Six Sigma methodology or not.
How are process control charts useful for inventory management?
We’ve talked about how control charts are useful for manufacturing processes, but they also have a role to play in inventory management, which we drill down into here.
Inventory management is the use of tools and techniques – especially inventory software – to keep a clear view on all your inventory information. This includes a range of information, such as what you have in stock, where it’s being stored, ingredients for each item, and more.
In other words, it records the Who, What, When Where and Why of your inventory.
With good inventory management, you’ll have visibility across your entire supply chain, and be able to spot red flags as or before they become a problem.
- Learn more: Supply chain management explained
Process control charts can provide clarity for inventory management
Control charts are all about clarity, just like inventory management. So you can use them to hunt for process inefficiencies, variations and red flags within your supply chain, inventory handling, production or shipping processes. You might, for instance, track your rate of stockouts using a control chart, or defects in a particular production line.
This could lead to improving your cost efficiency as you reduce wasted time/money, especially if you’re trying to go lean. In this sense control charts go hand in hand with inventory management – in that they are both geared towards a reduction in waste and costs, and greater efficiency overall.
But good control charts need good data
One of the big challenges of deploying control charts for inventory management – or any process for that matter – is that it requires good-quality, consistent data, and preferably in real time. Without this one critical component, you may as well still be guessing – and your control chart will not be able to give you the insights you need.
So if you’re going to use control charts to increase your business efficiency, you’ll want robust systems to provide you with that data.
This is where manufacturing inventory management software is central for using control charts. With real-time data on key measures like waste, lead times, order fulfilment, and defects, you’ll be able to produce an accurate control chart for a range of inventory measures – and improve these when they go outside your control limits.