A control chart is a method for measuring change. If you have information about your business that you want to measure and analyze, such as manufacturing defects, patient wait times or how long customers take to pay, the control chart can map out the data over time. One of the advantages of control charts is that the chart makes it easy to see when your performance has run into problems.
The control chart purpose is to spot trends over time. The centerline consists of the historical average for the process you're studying. Mapping data around the center gives you a visual representation of trends – for example, if data points consistently trend upward above the average.
Any process in your business is going to vary, from manufacturing to customer service. Machines wear out or malfunction. People have good days and bad days. Control charts measure variation and show it to you graphically. One look can tell you if variation in the process with which you're concerned is staying within acceptable limits.
Control charts are popular with manufacturers because there are so many processes they can track: defects, production time, inventory on hand, cost per unit and other metrics. However, businesses can use charts to measure nonmanufacturing processes such as:
- Billing errors
- Hospital patients who acquire infections during a stay
- Missed appointments
- Customer support calls
- Past-due invoices
- Days between billing and payment
- Rate of donations after a fund-raising appeal
- Operating expenses
- Unplanned employee absences
- Time between posting a position and hiring an employee
- Average employee tenure
The control chart rules are simple.
- Define what you want to control or measure, such as customer satisfaction, employee productivity or how often your legal firm settles cases out of court
- Identify how you'll collect and measure the data.
- Create a control chart.
- Collect data and chart it.
Once you have the data mapped, you can decide whether there's a problem. If not, you're done. If you spot trouble, you can implement a fix.
A control chart has three elements besides the data:
- A centerline representing the average for the process, such as average patient wait time, average increase in donations after a fund-raising appeal, etc.
- An upper control limit three standard deviations above the center.
- A lower control limit three standard deviations below the center.
A standard deviation is a statistical measure for telling whether variation is random and meaningless or significant. If you get results more than three standard deviations from the mean, they're almost certainly significant. Figuring standard deviation requires some number crunching, but Excel spreadsheets can help with that.
For a control chart example, suppose you're tracking the time between entering a bill in accounts receivable and the customer remitting payment. You look at your accounts to find the average historical payment time. That gives you your centerline, after which you calculate the upper and lower control limits.
If, say, you want to review the next quarter's payments, enter each of them as a data point on the control chart. If the centerline represents payment on day 12, quicker payments would go below the centerline and slower ones above it. The advantage of a control chart is that this makes it easier to see trends or outliers than if you glance at a row of numbers.
It may be that when you look at the chart, you see nothing special. Payment times fluctuate randomly around the centerline but within the control limits. There are no obvious patterns and nothing extraordinary going on.
If, however, you see a string of six or more points trending steadily up or down, that indicates that something significant happened. A single data point either above or below the control limits indicates something significant too.
Spotting a trend is only the first step. You'll have to talk to your team and figure out why things are changing. If the change indicates a problem, you can take steps to fix it.