Mastering Metrics with Standard Deviation: Boost Efficiency and Consistency with Real-World Examples!
Mirko Vojnovic
Innovative Technical Program/Project Manager | Expert in Analog, Digital, and Mixed Hardware Systems | 21 Patents Holder
Introduction
The ability to analyze data accurately and predict outcomes is invaluable in program management. Key statistical techniques like mean, standard deviation, and variance are not just numbers, but vital tools that play crucial roles in achieving these goals. These metrics help program managers make informed decisions, assess risks, and ensure projects stay on track. Let's dive into how these techniques are applied, complete with detailed explanations and practical examples.?
Mean: The Simple Average
The mean μ, often referred to as the arithmetical average, is a fundamental statistical measure. It's the sum of all values in a dataset divided by the number of values. In program management, the mean can be used to calculate average costs, durations, or any other metric relevant to project performance.
Example 1: Calculating Average Project Duration
Suppose a program manager wants to calculate the average duration of five completed projects. The durations in months are:
10, 12, 8, 15, and 11.
To find the mean:
μ = (10+12+8+15+11) / 5 = 56 / 5 = 11.2 months
This average gives the program manager a baseline expectation for future project durations.?
Variance: The Spread of Data
Variance σ^2 provides a helpful intermediary step in calculating standard deviation σ. Variance is the average of the squared differences from the mean divided by the number of samples.
(10?11.2)^2 = 1.44
(12?11.2)^2 = 0.64
(8?11.2)^2 = 10.24
(15?11.2)^2 = 14.44
(11?11.2)^2 = 0.04
σ^2 = (1.44 + 0.64 + 10.24 + 14.44 + 0.04) / 5= 26.85 = 5.36?
Standard Deviation: Measuring Variability
Standard deviation σ quantifies the amount of variation or dispersion in a set of data values. It is calculated from variance as a square root of variance, sqrt(variance).
In any normal distribution with mean μ and standard deviation σ, approximately?68%?of the data fall within one standard deviation of the mean. Approximately 95% of the data fall within two standard deviations of the mean, and 99.7% fall within three standard deviations of the mean.
It is also known as Gaussian distribution.
A low standard deviation indicates that the data points are close to the mean. In contrast, a high standard deviation indicates a wide range of values and higher uncertainty.
Let's revisit the project duration example and calculate the standard deviation σ.
σ = sqrt(5.36) ≈ 2.32 months
A standard deviation of 2.32 months indicates that in 68% of cases, project duration deviates from the average duration of 11.2 months by 2.32 months, and in 95% of cases, by 4.64 months, or 41% from the mean of 11.2 months. That gives you a good insight into how much contingency you need to add to your schedule.
Example 2: Assessing Cost Variability
Consider a program manager analyzing the costs of five similar projects, with costs in thousands of dollars:
50, 55, 53, 60, and 52.
To find the standard deviation:
1.????? Calculate the mean cost:
μ = (50+55+53+60+52) / 5 = 270 / 5 = 54 thousand?dollars
2.????? Subtract the mean from each cost and square the result:
(50?54)^2 = 16
(55?54)^2 = 1
(53?54)^2 = 1
(60?54)^2 = 36
(52?54)^2 = 4
3.????? Find the average of these squared differences to find the variance σ^2:
σ^2 = (16 + 1 + 1 + 36 + 4) / 5 = 58 / 5 = 11.6
4.????? Take the square root of the variance to get the standard deviation σ:
σ =sqrt(11.6) ≈ 3.41 thousand dollars
A standard deviation of 3.41 thousand dollars indicates that in 68% of cases, project costs deviate from the average cost of 54 thousand dollars by only 3.41 thousand dollars, and in 95% of cases, by 6.82 thousand dollars, or not more than 12%.
?
Practical Application in Program Management
Risk Management
Understanding the variability in project costs and durations is crucial for risk management. For instance, a high standard deviation in cost estimates could signal potential budget overruns, prompting managers to allocate contingency funds. Conversely, a low standard deviation might indicate stable processes but could also suggest insufficient exploration of possible risks. In this way, the standard deviation is a vital indicator of the degree of uncertainty in project costs and durations, guiding risk management decisions.
Performance Evaluation
By comparing the mean and standard deviation of key performance indicators (KPIs) across different projects, program managers can identify outliers and investigate underlying causes. This comparison is crucial as it helps in benchmarking and setting realistic targets for future projects. For example, suppose one project has a significantly higher standard deviation in duration. In that case, it may indicate issues with planning or execution that need addressing.
Example 1: Using KPIs for Mean and Standard Deviation in a Customer Support Department
Let's consider a customer support department that tracks the average time to resolve a ticket (Mean Time to Resolution, or MTTR) and the variability in resolution times (Standard Deviation of Time to Resolution). These KPIs are crucial for assessing efficiency and consistency.
Data Collected
Over a month, the resolution times (in hours) for ten tickets are as follows:
2.5, 3.0, 2.0, 4.0, 3.5, 2.5, 3.0, 5.0, 4.5, 3.0
Calculating the Mean
First, we calculate the mean resolution time MTTR:
MTTR = (2.5+3.0+2.0+4.0+3.5+2.5+3.0+5.0+4.5+3.0) / 10 = 33 / 10 = 3.3 hours
Calculating the Standard Deviation
Next, we find the standard deviation to understand the variability in resolution times.
领英推荐
1.????? Calculate the squared differences from the mean:
(2.5?3.3)^2 = 0.64
(3.0?3.3)^2 = 0.09
(2.0?3.3)^2 = 1.69
(4.0?3.3)2=0.49
(3.5?3.3)^2 = 0.04
(2.5?3.3)^26=60.64
(3.0?3.3)^2 = 0.09
(5.0?3.3)^2 = 2.89
(4.5?3.3)^2 = 1.44
(3.0?3.3)^2 = 0.09
2.????? Find the variance σ^2 by averaging these squared differences:
σ^2 =( 0.64 + 0.09 + 1.69 + 0.49 + 0.04 + 0.64 + 0.09 + 2.89 +1.44 +
+ 0.09) / 10 = 8.1 / 10 = 0.81
3.????? Calculate the standard deviation σ :
σ = sqrt(0.81) ≈ 0.9 hours
Interpretation and Conclusions
A low standard deviation suggests that the support process is stable and predictable, which is desirable. If the standard deviation were higher, it might indicate inconsistent ticket handling, prompting further investigation into the causes of delays.
By continuously monitoring these KPIs, the customer support department can maintain high efficiency and identify deviations that may require process improvements.
Example 2: Using Mean and Standard Deviation in Project Management Risk Analysis
Let's consider a software development project where we want to analyze the risk associated with the project completion time. We have collected data on the time taken (in weeks) to complete similar projects in the past.
Data Collected
The project completion times (in weeks) for 10 previous projects are as follows:
15, 18, 20, 16, 22, 17, 19, 21, 23, 17
Calculating the Mean
First, we calculate the mean μ completion time:
μ = (15+18+20+16+22+17+19+21+23+17) / 10 = 188 / 10 = 18.8?weeks
Calculating the Standard Deviation
Next, we find the standard deviation to understand the variability in completion times.
(15?18.8) ^2 = 14.44
(18?18.8) ^2 = 0.64
(20?18.8) ^2 = 1.44
(16?18.8) ^2 = 7.84
(22?18.8) ^2 = 10.24
(17?18.8) ^2 = 3.24
(19?18.8) ^2 = 0.04
(21?18.8) ^2 = 4.84
(23?18.8) ^2 = 17.64
(17?18.8) ^2 = 3.24
2. Find the variance σ^2 by averaging these squared differences:
σ^2 = (14.44+0.64+1.44+7.84+10.24+3.24+0.04+4.84+17.64+3.24) / 10 = (63.6) / 10 = 6.36
3. Calculate the standard deviation σ:
σ = sqrt( 6.36) ≈ 2.52?weeks
Interpretation and Conclusions
Risk Analysis
To assess the risk, we can use the mean and standard deviation to create a range of probable completion times. For instance, assuming a normal distribution:
By understanding this range, project managers can better prepare for potential delays and allocate resources more effectively.
?
Conclusion
In program management, the mean, standard deviation, and variance are not just numbers; they are vital tools for predicting outcomes, managing risks, and ensuring project success. By understanding and applying these statistical techniques, program managers can navigate the complexities of their projects with greater confidence and precision. Whether balancing budgets or keeping projects on schedule, these metrics provide a clearer view of the road ahead, making it more manageable and ensuring project success.
Note:
This is a companion tutorial to my previous article, "Using Bayesian Techniques in Program Management: A Comprehensive Guide", since some Bayesian techniques use Standard Deviation as the basis for further analysis.
?
Sources: