What is a control chart? (Definition and how to make one)

By Indeed Editorial Team

Published 14 April 2022

The Indeed Editorial Team comprises a diverse and talented team of writers, researchers and subject matter experts equipped with Indeed's data and insights to deliver useful tips to help guide your career journey.

Tools like control charts can be very useful for analysing an organisation's processes to find insights for improvements. Other names for control charts include Shewhart charts, process-behaviour charts or SPC (statistical process control) charts. If you're interested in project management or similar roles, understanding control charts can be very useful. In this article, we explain what a control chart is, how they work, how you can make a control chart and their major benefits.

What is a control chart?

Understanding the answer to the question 'what is a control chart?' can help you to determine if this is the right tool for your needs. A control chart looks like a line graph, with points plotted and connected with straight lines. It also has a straight line that shows the average between these points. The x-axis is going to be something like time, such as one point for each month or day. The y-axis is typically used to represent the number of occurrences of a specific phenomenon that you're interested in. These occurrences can represent something negative or positive.

For example, you could have a control chart that shows the number of returned items per month at a clothing shop. The x-axis would show months whereas the y-axis would show the number of returned items. If the majority of the points are within the same range of the average, this means that the process is under control. Any points that are excessively distant from the average are out of control, although this can be a positive indicator in some instances. There's also the upper control limit (UCL) and lower control limit (LCL), which indicate the range of control.

Related: A guide to the different types of charts and graphs

How do control charts work?

Control charts are useful tools for finding the causes of unwanted variation and can even help you improve an organisation's processes. To help you achieve this, the control chart has several elements that can help you derive insights. These are explained below:

Control limits

In a control chart, control indicates predictability. Almost no process produces the exact same outputs every time. Instead, a process has some variation in what it produces. In most cases, this variation is a natural part of the process and nothing to become concerned about, but sometimes variation can be more extreme. A control chart helps you to identify this with the two limits, namely the UCL and LCL.

You can calculate these using standard deviation. This means that any points that lie between the two limits are within control limits. This simply means that they're representative of the normal variation in the process' outputs. Any point above the UCL or below the LCL indicates that something unusual happened and this could warrant further investigation.

Signals and noise

These two terms describe the nature of any point on the control chart. Noise represents normal variations in process outputs, whereas a signal describes abnormal variations. Therefore, noise is the entire area between the UCL and LCL and any point outside of this area is a signal. These two also have different causes. 'Common causes' produce noise, whereas 'special causes' produce signals. You may find that people use these terms interchangeably.

What this means is that any point that's noise and therefore has a common cause, is one that you can mostly ignore as it's not unusual. Any point with special cause (a signal) is unusual and might deserve investigation. For instance, in the example of a clothing shop tracking the number of returned items, if in a particular month the returns were abnormally high and above the UCL, managers might want to investigate why this happened. Perhaps they received a bad batch of items? Were some items mislabelled? Finding the cause can help managers avoid such occurrences in future.

Related: What is quantitative analysis? (With definitions and examples)

How to make a control chart

If you want to make a control chart yourself to analyse process performance, consider following the steps outlined below:

1. Decide on your parameters

The first step is to decide what you want to analyse using the control chart. It's important to remember that control limits are different to specification limits. The latter is what's considered acceptable and is something that an organisation's decision makers decide. So, a process could be out of control but still be within specification limits.

For your control chart, the availability of data might make it quite simple to choose your parameters. Other factors might be the nature of the sector in which you're working. Decide the frequency of the observations and the specific process indicator you're interested in exploring.

2. Plot it on a control chart

Once you've got the data you're interested in, you can plot it on a control chart. This is initially the same as a line graph, with each point joined to the next with a straight line. You can use spreadsheet software to do this quite easily.

3. Calculate the average

The next step is to add the average, which is going to give you the middle control line that's between the UCL and LCL. The easiest way to do this is with your spreadsheet software, as almost all of them have a feature that allows you to do this. If you're unsure, conduct a quick online search to find out how to do this for your specific software.

Related: Functional skills: definition, importance and examples

4. Calculate the standard deviation

This step is important for adding your upper and lower control limits. There are four steps for calculating the standard deviation for your chart:

  1. Find the mean: This is the simple average of all the observations you're plotting. Add them all together and then divide by the number of observations.

  2. Subtract the mean and square: For every observation that you're plotting, subtract the mean from this figure and then square the result.

  3. Find the squared mean: For all your squared results from the previous step, calculate their mean using the same process as before.

  4. Square root the mean: Calculate the square root of the mean you calculated in the previous step. This gives you one standard deviation.

5. Add the UCL and LCL

The upper and lower control limits are a certain distance above and below the average line. You can find these using the standard deviation you calculated previously. Depending on how strictly controlled you want the process to be, you can use between one and three standard deviations. Three is quite common, so that's a good choice unless you have a specific reason to choose a lower number.

This means multiplying the standard deviation by three. You then add a parallel line above the average line by three standard deviations. You then add another parallel line that's below the average line by three standard deviations. These are your UCL and LCL respectively.

6. Make note of any signals

Once you've plotted your control chart, it's going to show you where special causes have caused a signal. These are the points above the UCL and below the LCL. Make a note of these, as they're indicative of unusual events or causes and could help you find ways to improve the process' effectiveness.

Related: Analytical skills: definitions and examples

Benefits of using control charts

There are several advantages to using control charts, especially for understanding what your data represents. Below are some of the major benefits of using a control chart:

  • Know what to ignore: Since a control chart shows you which points on a graph are just noise and therefore completely normal, you're going to know which observations you can safely ignore. This could save you a lot of time that you would've spent investigating something completely normal.

  • Finding improvements: In some cases, points that are outside of the control limits can indicate something positive, such as much lower product returns than average. If this is a special cause observation, you can investigate what caused the lower returns and use this to improve the overall process.

  • Spotting patterns: Over time, your control chart might show you certain patterns in the process, such as signal points that regularly occur at specific times. This can be even more insightful than a single signal point and you can therefore find solutions or otherwise derive value from this information.

  • Predict future performance: Once you've got a control chart that's got a significant number of observations, you can more effectively predict what future performance is going to look like. This can also ensure that you know whether or not to react to any future variations in process performance as they occur.

Explore more articles