In process improvement, teams often want to move quickly toward solutions. However, before changing a workflow, adjusting staffing, or redesigning a service step, you need a reliable answer to a fundamental question:
What is the process doing today?
The average, or mean, provides one of the clearest starting points. It condenses a set of observations into a single measure of central tendency and establishes a reference point for performance. In Lean Six Sigma, that reference point becomes part of the process baseline used to evaluate improvement.
Yet the mean is not the complete story. An average without variation context can create false confidence, hide instability, or encourage a team to respond to the wrong problem. To fully appreciate the value of the mean, you must interpret it alongside the spread of the data, control charts, and the Voice of the Process.
What Is the Average or Mean?
The arithmetic mean is calculated by adding all observed values and dividing the total by the number of observations:
[
\text{Mean} = \frac{\text{Sum of all observations}}{\text{Number of observations}}
]
For example, if a claims team records processing times of 42, 47, 45, 44 and 46 minutes, the mean is:
[
\frac{42+47+45+44+46}{5} = \frac{224}{5} = 44.8\text{ minutes}
]
This result gives the team a practical summary: under the observed conditions, the process typically completes a claim in approximately 44.8 minutes.
The mean is useful because it:
- Establishes a starting point for the Measure phase of DMAIC.
- Supports comparisons between current and future performance.
- Helps teams quantify improvement in familiar units.
- Provides the centre line for many statistical analyses and control charts.
- Translates process data into a metric leaders can understand.
In the realm of Lean Six Sigma, the mean is therefore more than a descriptive statistic. It is a baseline indicator that supports disciplined decision-making.
Why the Mean Matters in a Lean Six Sigma Baseline
A project baseline describes the current state before an intervention. Depending on the project, it may include:
- Average cycle time.
- Average cost per transaction.
- Average response time.
- Average defect count.
- Average delivery duration.
- Average output per shift.
For example, a logistics team might establish that the mean order-picking time is 12.6 minutes across 200 orders. After introducing a revised picking sequence, the team can compare the new mean with the original baseline.
If the new average is 10.4 minutes, the apparent improvement is:
[
12.6 – 10.4 = 2.2\text{ minutes per order}
]
At a volume of 4,000 orders per month, that represents:
[
2.2 \times 4,000 = 8,800\text{ minutes}
]
That is approximately 146.7 labour hours per month. The mean converts operational change into measurable impact.
However, the comparison is only credible if both datasets were collected using consistent definitions, sampling methods, and operating conditions. A mean calculated from a poorly defined metric is precise but not necessarily useful.
For a strong baseline, document:
- The operational definition of the metric.
- The population and sample size.
- The period of data collection.
- The process conditions under which measurements were taken.
- The measurement system used to collect the data.
- The variation observed within the sample.
Worked Example: How a Mean Establishes a Baseline
Consider an invoice approval process. The project team collects ten cycle-time observations, measured in minutes:
| Invoice | Cycle time |
|---|---|
| 1 | 42 |
| 2 | 47 |
| 3 | 45 |
| 4 | 44 |
| 5 | 46 |
| 6 | 43 |
| 7 | 49 |
| 8 | 44 |
| 9 | 45 |
| 10 | 95 |
The total is:
[
42+47+45+44+46+43+49+44+45+95 = 500
]
The mean is:
[
\frac{500}{10} = 50\text{ minutes}
]
The team might now report: “The baseline invoice approval time is 50 minutes.”
That statement is mathematically correct, but it is incomplete.
The first nine observations range from 42 to 49 minutes, while the tenth observation is 95 minutes. The first nine observations have a mean of:
[
\frac{405}{9}=45\text{ minutes}
]
The tenth observation has increased the overall mean from 45 to 50 minutes. It has also significantly increased the spread. The standard deviation of the first nine observations is approximately 2.1 minutes, while the sample standard deviation for all ten observations is approximately 15.9 minutes.
The mean alone makes the process appear broadly centred at 50 minutes with considerable variation. The data tells a more precise story:
- The process usually operates close to 45 minutes.
- One observation took 95 minutes.
- The 95-minute result may represent a special cause, such as a missing approval, system outage, incorrect routing, or an unusually complex invoice.
- The process baseline should not be interpreted until the exceptional observation has been investigated.
This does not mean the 95-minute observation should simply be deleted. It must be verified, explained, and retained in the project record. Removing data without a valid process reason creates a distorted baseline. The correct response is to investigate whether the observation reflects normal process behaviour or an identifiable special cause.
Why a Baseline Without Variation Context Misleads
Two processes can have the same average and deliver very different experiences.
Imagine two customer service teams, each with a mean response time of 30 minutes:
- Team A: response times range from 28 to 32 minutes.
- Team B: response times range from 5 to 55 minutes.
Their averages are identical, but their operational reliability is not. Team A provides a predictable experience. Team B creates uncertainty for customers, employees, and managers.
This is why a robust baseline includes both:
- Location: Where the data is centred, often represented by the mean.
- Spread: How widely the observations vary, represented by range, standard deviation, or control limits.
Variation influences customer experience, resource planning, capacity requirements, and defect risk. A process with a reasonable mean may still fail customer expectations if its output is inconsistent.
In Lean Six Sigma, this distinction is central. Lean focuses attention on flow and value, while Six Sigma provides structured methods for understanding and reducing variation. The mean helps identify the process centre; variation explains how dependable that centre is.
Common Cause and Special Cause Variation
A baseline must also be interpreted through the distinction between common cause and special cause variation.
Common cause variation
Common cause variation is the natural fluctuation created by the ordinary design and operation of a process. It may result from:
- Routine differences in workload.
- Normal equipment or system performance.
- Standard differences in task complexity.
- Everyday variation in staffing or material conditions.
When a process is stable, common cause variation produces a predictable pattern over time. Improvement usually requires changing the process itself through better methods, standard work, technology, training, or system design.
Special cause variation
Special cause variation arises from an unusual, identifiable factor outside the normal process pattern. Examples include:
- A system interruption.
- A new employee using an incorrect procedure.
- A supplier delivery outside normal conditions.
- A one-time approval escalation.
- A machine fault or measurement error.
Special causes require investigation and appropriate corrective action. If they remain mixed into the baseline, the mean and standard deviation may overstate the process’s normal performance.
The objective is not to eliminate every fluctuation immediately. The objective is to understand which variation is routine, which is exceptional, and what type of action is appropriate.
Connecting the Mean to an X-Bar Chart
An X-bar chart monitors the average of successive subgroups rather than individual observations. For example, an invoice team might collect five cycle-time measurements every morning and calculate one subgroup mean per day.
The chart typically includes:
- A centre line, representing the overall process mean.
- An upper control limit.
- A lower control limit.
- Plotted subgroup means over time.
The X-bar chart shows whether process averages remain stable or whether the process is shifting. It is commonly interpreted alongside an R chart, which monitors the range within each subgroup. The X-bar chart reveals changes in the process centre, while the R chart provides insight into short-term variation.
A control chart is not the same as a specification chart. Customer or business requirements define specification limits. Control limits are calculated from process behaviour. They describe what the process is currently doing.
Teams can learn more about this application through Measure Phase: Creating Control Charts for Baseline Data in Lean Six Sigma.
The Mean as the Voice of the Process
The Voice of the Process is the evidence generated by actual process performance. It includes the process mean, variation, trends, patterns, defects, and stability signals.
The Voice of the Process should be considered alongside:
- Voice of the Customer: What customers require and value.
- Voice of the Business: What the organisation must achieve financially and strategically.
- Voice of the Process: What the process is currently capable of delivering.
For instance, customers may require invoice approval within 48 minutes. The business may want to reduce labour cost. The process data may show a mean of 45 minutes, but with occasional results above 90 minutes.
That means the process average appears acceptable, yet its long-tail performance may still create customer dissatisfaction and escalation work. The improvement opportunity may not be to reduce the mean dramatically. It may be to reduce variation and prevent exceptional delays.
You can explore this principle further in Voice of the Process: Why Your Data Is Already Telling You the Answer.
A Practical Protocol for Using the Mean Correctly
Use the following sequence when establishing a process baseline:
- Define the metric clearly. Specify exactly when measurement starts and ends.
- Collect representative data. Include normal shifts, operators, products, customers, or demand conditions.
- Calculate the mean. Use it as the initial measure of process location.
- Measure variation. Review range, standard deviation, histograms, or box plots.
- Plot data over time. Use a run chart or suitable control chart.
- Investigate unusual signals. Separate common cause behaviour from special causes.
- Confirm the baseline. Record the mean, variation, time period, exclusions, and assumptions.
- Compare improvement results fairly. Use consistent definitions and collection methods.
This disciplined approach prevents teams from treating a single average as the complete truth.
Build Stronger Statistical Thinking Through Training
Understanding the mean is an accessible entry point into broader Lean Six Sigma capability. As professionals progress, they learn how to combine averages with sampling, measurement system analysis, hypothesis testing, control charts, capability analysis, and structured DMAIC project management.
Practical Lean Six Sigma training is especially valuable because it connects formulas with realistic decisions. Through case studies, dummy data, worked examples, and end-to-end projects, learners can practise establishing baselines, interpreting variation, and converting the Voice of the Process into improvement priorities.
Lean 6 Sigma Hub offers CSSC-accredited online Lean Six Sigma training from White Belt through Master Black Belt. Learners can also review the end-to-end hypothetical project to see how data supports decisions throughout DMAIC.
Conclusion
The average, or mean, provides the baseline that launches a disciplined improvement project. It tells you where the process is centred and gives you a practical reference for evaluating change.
But the mean must never stand alone. Without variation context, it can conceal instability, exaggerate typical performance, or obscure special causes. When combined with standard deviation, time-based analysis, X-bar charts, and the Voice of the Process, it becomes a powerful foundation for understanding what the process is truly capable of delivering.
Start your Lean Six Sigma development today by pursuing structured training and professional certification, then use the mean, variation, and Voice of the Process to lead measurable improvement.
Kaizen. Kai-Care. Kai-Done. ( Lean Six Sigma)
