Periodic, scheduled reviews of identified risks, risk responses, and risk priorities should occur during the project. The idea here is to monitor risks and their status and determine whether their consequences still have the same impact on the project objectives as when they were originally planned. Every status meeting should have a time set aside to discuss and review risks and response plans.

Risk identification and monitoring is an ongoing process throughout the life of the project. Risks can change, and previously identified risks might have greater impacts than originally thought as more facts are discovered. Reassessment of risks should be a regular activity performed by everyone involved on the project. Monitor the risk register, including those risks that have low scores, and risk triggers. You should also monitor the risk responses that have been implemented for their effectiveness in dealing with risk. You might have to revisit the Perform Qualitative and Perform Quantitative Risk Analysis processes when new risk consequences are discovered or risk impacts are found to be greater than what was originally planned.

Risk audits are carried out during the entire life of the project by risk auditors. Risk auditors are not typically project team members and are expertly trained in audit techniques and risk assessment. Risk audits are specifically interested in examining the implementation of response plans and their effectiveness at dealing with risks and their root causes.

This tool and technique compares the technical accomplishments of project milestones completed during the Executing processes to the technical milestones defined in the project Planning processes. Variances might indicate that a project risk is looming, and you'll want to analyze and prepare a response to it if appropriate. For example, a technical milestone for a new computer software project might require that the forms printed from a particular module include a bar code at the bottom of the page. If the bar code functionality does not work once the module is coded, a technical deviation exists, which means you should reexamine project risks. In this particular example, project scope is likely at risk.

The risk register should be updated under two conditions. The first condition is when a risk audit or risk reassessment concludes that some element of the original risk information has changed—for example, the impact or probability scores are updated to reflect new conditions, the priority of the risk has changed, or the response plan has been updated. The second condition for updating the risk register is when the risk needs to be closed. If a risk event occurs, you'll record that in the risk register along with the effectiveness of the response plan. This information becomes an input to the Close Project Process, which I'll cover in the next chapter.

Change requests must be processed through the Perform Integrated Change Control process. You might find that a workaround is needed when implementing a requested change or recommended corrective action. A workaround is an unplanned response to a negative risk event. It attempts to deal with the risk in a productive, efficient manner. If no risk response plan exists (this might be the case when you accept a risk event during the Planning process) or an unplanned risk occurs, workarounds are implemented to deal with the consequences of the risk.

You can accomplish performance measurement analysis using a technique called earned value measurement (EVM). Simply stated, EVM compares what you've received or produced to what you've spent.

The EVM continuously monitors the planned value, earned value, and actual costs expended to produce the work of the project (I'll cover the definition of these terms shortly). When variances that result in cost changes are discovered (including schedule variances and cost variances), those changes are managed using the project change control system. The primary function of this analysis technique is to determine and document the cause of the variance, to determine the impact of the variance, and to determine whether a corrective action should be implemented as a result. We'll walk through various examples of how to determine these variances later in this section.

EVM looks at schedule, cost, and scope project measurements together and compares them to the actual work completed to date. It is the most often used performance measurement method. EVM is performed on the work packages and the control accounts of the WBS. To perform the EVM calculations, you need to first gather the three measurements mentioned earlier: the planned value (PV), actual cost (AC), and earned value (EV).

Let's take a look at some definitions of these key measurements before diving into the actual calculations:

According to the earlier definition, EV is the sum of the cumulative budgeted costs for completed work for all activities that have been accomplished as of the measurement date. For example, if your total budget is $1,000 and 50 percent of the work has been completed as of the measurement date, your EV would equal $500. You can plot all the PV, AC, and EV measurements graphically to show the variances between them. If there are no variances in the measurements, all the lines on the graph remain the same, which means the project is progressing as planned. Figure 11.1 shows an example that plots these three measurements.

Figure 11.1. Earned value

All of these measurements include a cost component. Costs are displayed in an S curve because spending is minimal in the beginning of the project, picks up steam toward the middle, and then tapers off at the end of the project. This means your earned value measurements will also take on the S curve shape.

Now you can calculate whether the project is progressing as planned or if variances exist in the approved baseline by using a variety of formulas discussed in the following sections. Use Figure 11.1 as your example for the formulas that follow. The Figure 11.1 totals are as follows: PV = 400, EV = 375, AC = 325.

Forecasting uses the information you've gathered to date and estimates the future conditions or performance of the project based on what you know when the calculation is performed. Forecasts are based on work performance information (an output from the Executing process group) and your predictions of future performance.

The forecasting formulas you'll see later in this section are called estimate at completion (EAC). EAC estimates (or forecasts) the expected total cost of a work component, a schedule activity, or the project at its completion. This is the probable final value for the work component (or project). EAC is most often calculated by using actual costs incurred to date plus a bottom-up ETC estimate. The formula for the most typical EAC looks like this:

The bottom-up estimate to complete (ETC) is an estimate provided by the members of the project team actually working on the project activities. They provide the project manager with an estimate of the amount of effort remaining (and therefore the cost of the effort) based on the activities they have completed to date and what they believe will occur in the future. Their estimates are summed to come up with a total ETC. We'll look at two other ETC formulas at the end of this section.

There are three EAC forecasting formulas outlined in the PMBOK® Guide that we'll look at next. They reference a new parameter you haven't seen yet called budget at completion (BAC). BAC is the sum of all the budgets established for all the work in the work package, control account, schedule activity, or project. It's the total planned value for the work component or project.

The first EAC formula is called "EAC forecast for ETC work performed at the budgeted rate." I know that's a mouthful. Here's what you should know. This formula calculates EAC based on the actual costs to date and the assumption that ETC work will be completed at the budgeted rate. The formula looks like this:

Let's assume your AC to date is $800, BAC is $1,200, and EV is $600. EAC, assuming work ETC will be completed at the budgeted rate, is as follows:

In English, you'll spend $1,400 to complete this work component, assuming the remaining work is performed at the budgeted rate.

The next EAC formula is called "EAC forecast for ETC work performed at the present CPI." I didn't make up these titles! Here's what you need to know. This forecast assumes that future performance will be just like the past performance for the project. The formula looks like this:

For this example, let's assume that BAC is $2,200 and cumulative CPI is 1.2. The formula looks like this:

This formula predicts you will spend less than the originally budgeted amount for the project.

The last formula is called "EAC forecast for ETC work considering both SPI and CPI factors." This formula assumes two things: there is a negative cost performance to date and the project schedule dates must be met. The formula looks like this:

Let's assume AC is $1,000, BAC is $1500, EV is $900, cumulative CPI is .97, and cumulative SPI is 1.05. Here's the resulting EAC:

Based on the assumption that cost performance to date is negative and we must meet the project schedule date, EAC is $1,598.10.

In addition to the bottom-up ETC, there are two other formulas for calculating ETC that you should be aware of for the exam. They are discussed next.

When you believe that future cost variances will be similar to the types of variances you've seen to date, you'll use this formula to calculate ETC:

Assuming your cumulative earned value (cumulative EV) is 725, cumulative CPI value is 1.12, and BAC is 1,000, plug in the numbers:

Therefore, at the measurement date, you need $245.50 to complete all the remaining work of this work component (or project if you're using project totals), assuming variances in the future will be the same as they have been to date.

When you believe that future cost variances will not be similar to the types of variances you've seen to date, you'll use this formula to calculate ETC:

Now calculate your value:

In this case, you need $275 to complete all the remaining work of this work component, assuming variances in the future are different than they have been to date.

To-complete performance index (TCPI) is the projected performance level the remaining work of the project must achieve in order to meet the BAC or EAC. It's calculated by dividing the work that's remaining by the funds that are remaining.

The formula for TCPI when using the BAC is as follows:

Assume for this example that BAC is $1,000, EV is $700, and AC is $800.

This means you'll need to reach a CPI rate that's 1.5 times what you've experienced to date in order to meet the BAC goal. If the result is less than one, future work does not have to be performed as efficiently as past performance.

When the BAC is no longer attainable, the project manager should calculate a new EAC and this new estimate becomes the goal you'll work toward once it's approved by management. The TCPI formula when EAC is the goal you're aiming for is as follows:

We'll use the same assumptions as we did in the formula earlier and note that EAC is $1,200. The formula looks like this:

There is one last thing to note regarding this formula. If cumulative CPI falls below one, all future project work must be performed at the TCPI

Performance reviews compare cost performance over time and the estimates of funds needed to complete the remaining work. Three types of analyses are associated with performance reviews: variance analysis, trend analysis, and earned value performance.

According to the PMBOK® Guide, trend analysis determines whether project performance is improving or worsening over time by periodically analyzing project results. These results are measured with mathematical formulas that attempt to forecast project outcomes based on historical information and results. You can use several formulas to predict project trends, but it's outside the scope of this book to go into them. For the exam, you're expected to understand the concept behind trend analysis, not the formulas used to calculate it. You'll want to remember that you can use the results you've analyzed using trend analysis formulas to predict future project behavior or trends.

I've already covered earned value performance and variance analysis is its own technique, so we'll look at it next.

Variance analysis in the Control Costs process examines the difference between the performance cost baseline and actual performance. One formula you can use with this technique is called variance at completion (VAC). This calculates the difference between the budget at completion and the estimate at completion. It looks like this:

If the result is a negative number, it means you're not doing as well with costs as you anticipated and that variance exists. Assuming your project performance is improving, as the project progresses, variances will become smaller.

Control charts measure the results of processes over time and display the results in graph form. Control charts are a way to measure variances to determine whether process variances are in control or out of control.

A control chart is based on sample variance measurements. From the samples chosen and measured, the mean and standard deviation are determined. Perform Quality Control is usually maintained—or said to be in control—within plus or minus three standard deviations. In other words, Perform Quality Control says that if the process is in control (that is, the measurements fall within the control limits), you know that 99.73 percent of the parts going through the process will fall within an acceptable range of the mean. If you discover a part outside of this range, you should investigate and determine whether corrective action is needed.

Figure 11.2 illustrates an example control chart.

Figure 11.2. Control chart

Let's assume you've determined from your sample measurements that 5 mm is the mean in the example control chart. One standard deviation equals 0.02. Three standard deviations on either side of the mean become your upper and lower control points on this chart. Therefore, if all control points fall within plus or minus three standard deviations on either side of the mean, the process is in control. If points fall outside the acceptable limits, the process is not in control and corrective action is needed.

Control charts are used most often in manufacturing settings where repetitive activities are easily monitored. For example, the process that produces widgets by the case lot must meet certain specifications and fall within certain variances to be considered in control. However, you aren't limited to using control charts only in the manufacturing industry. You can use them to monitor any output. You might consider using control charts to track and monitor project management processes. You could plot cost variances, schedule variances, frequency or number of scope changes, and so on to help monitor variances.

You have probably heard of the 80/20 rule. Vilfredo Pareto is the person credited with discovering this rule. He observed that 80 percent of the wealth and land ownership in Italy was held by 20 percent of the population. Over the years, others have shown that the 80/20 rule applies across many disciplines and areas. As an example, generally speaking, 80 percent of the deposits of any given financial institution are held by 20 percent of its customer base. Let's hope that rule doesn't apply to project managers, though, with 20 percent of the project managers out there doing 80 percent of the work!

The 80/20 rule as it applies to quality says that a small number of causes (20 percent) create the majority of the problems (80 percent). Have you ever noticed this with your project or department staff? It always seems that just a few people cause the biggest headaches. But I'm getting off track.

Pareto charts are displayed as histograms that rank-order the most important factors—such as delays, costs, and defects, for example—by their frequency over time. His theory is that you get the most benefit if you spend the majority of your time fixing the most important problems. The information shown in Table 11.1 is plotted on an example Pareto chart shown in Figure 11.3.

Figure 11.3. Pareto chart

The problems are rank-ordered according to their frequency and percentage of defects. The defect frequencies in this figure appear as black bars, and the cumulative percentages of defects are plotted as circles. The rank-ordering of these problems shows you where corrective action should be taken first. You can see in Figure 11.3 that problem A should receive priority attention because the most benefit will come from fixing this problem.

Run charts are used to show variations in the process over time or to show trends (such as improvements or lack of improvements) in the process. Differences in results will occur in processes because there is no such thing as a perfect process. When processes are considered in control, differences in results might occur because of common causes of variances or special-cause variances.

Common causes of variances come about as a result of circumstances or situations that are relatively common to the process you're using and are easily controlled at the operational level. Special-cause variances are variances that are not common to the process. For example, perhaps you have very detailed processes with specific procedures that must be followed in order to produce the output and a process gets missed. Or maybe your project requires the manufacturing of a certain part and a machine on the line has a problem and requires a special calibration. This is an easy set of terms to remember because their names logically imply their definitions.

For the exam, you should understand the three types of variances that make up common causes of variances:

Common cause variances that do not fall within the acceptable range are difficult to correct and usually require a reorganization of the process. This has the potential for significant impact, and decisions to change the process always require management approval.

Trend analysis is another technique that's carried out using run charts. Trend analysis in the Perform Quality Control process is a mathematical technique that uses historical results to predict future outcomes. Trend analysis often tracks variances in cost and schedule performance by monitoring the number of activities completed with significant variances within a certain time period. This information can then be used to forecast future performance. Trend analysis also tracks technical performance by determining the number of defects observed and the number of defects not corrected. Technical performance measurements compare the technical accomplishments of project milestones completed to the technical milestones defined in the project Planning process group.

Scatter diagrams use two variables, one called an independent variable, which is an input, and one called a dependent variable, which is an output. Scatter diagrams display the relationship between these two elements as points on a graph. This relationship is typically analyzed to prove or disprove cause-and-effect relationships. As an example, maybe your scatter diagram plots the ability of your employees to perform a certain task. The length of time (in months) they have performed this task is plotted as the independent variable on the X axis, and the accuracy they achieve in performing this task, which is expressed as a score—the dependent variable—is plotted on the Y axis. The scatter diagram can then help you determine whether cause-and-effect (in this case, increased experience over time versus accuracy) can be proved. Scatter diagrams can also help you look for and analyze root causes of problems.

The important point to remember about scatter diagrams is that they plot the dependent and independent variables, and the closer the points resemble a diagonal line, the closer these variables are related. Figure 11.4 shows a sample scatter diagram.

Figure 11.4. Scatter diagram

Statistical sampling involves taking a sample number of parts from the whole population and examining them to determine whether they fall within acceptable variances. The formula to calculate the correct sample size is beyond the scope of this book. However, Creative Research Systems has an online calculator and an explanation of statistical sampling that you might find useful. You can find them at http://www.surveysystem.com/sscalc.htm.

Perhaps you determine to statistically sample 25 parts out of a lot or run. The quality plan outlines that the lot will pass if four parts or fewer fall outside the allowable variance.

Statistical sampling might also involve determining the standard deviation for a process, as discussed in the control chart tool and technique. The quality plan determines whether plus or minus two standard deviations—95.44 percent of the population—is adequate or whether plus or minus three standard deviations—99.73 percent—is adequate.

Inspection involves physically looking at, measuring, or testing results to determine whether they conform to the requirements or quality standards. It's a tool used to gather information and improve results. Inspections might occur after the final product is produced or at intervals during the development of the product to examine individual components. Acceptance decisions are made when the work is inspected and the work is either accepted or rejected. When work is rejected, it might have to go back through the process for rework. Inspection is also known as reviews or peer reviews.

Inspection might take actual measurements of components to determine whether they meet requirements. Maybe a component part for your product must be exactly 5 mm in length. To pass inspection, the parts are measured and must meet the 5 mm length requirement. If they measure 5 mm, they pass; if they do not, they fail.

Measurements can vary even if the variances are not noticeable. Machines wear down, people make mistakes, the environment might cause variances, and so on. Measurements that fall within a specified range are called tolerable results. So, instead of 5 mm exactly, maybe a range between 4.98 mm and 5.02 mm is an acceptable or tolerable measurement for the component. If the samples that are measured fall within the tolerable range, they pass; otherwise, they fail inspection.

One inspection technique uses measurements called attributes. The measurements taken during attribute sampling determine whether they meet one of two options, conforming or nonconforming. In other words, the measurements conform or meet the requirement or they do not conform. This can also be considered a pass/fail or go/no-go decision.

Attribute conformity and inspections are not necessarily performed on every component part or every end product that's produced. That's time-consuming and inefficient when you're producing numerous components. Inspection in cases like this is usually performed on a sampling of parts or products where every x number of parts is tested for conformity or measurement specifics.

Inspection will tell you where problems exist and gives you the opportunity to correct them, thus leading to quality improvements. The other tools and techniques I'll talk about in these sections also lead to quality improvements in the product or process or in both.