DMAIC Phase 2: Measure

The goal of the Measure phase of a Six Sigma DMAIC project is to get as much information as possible on the current process so as to fully understand both how it works and how well it works. This entails three key tasks: creating a detailed process map, gathering baseline data and summarizing and analyzing the data. In some cases the process mapping is created first so that information gleaned from it can guide the data collection process. In other cases the general data needs are already known and the two pieces can be worked on simultaneously.

Depending on the type of process, a Six Sigma process map may be created using direct input from the individuals who participate in the process, by an observer who monitors and records information about the process, or a combination of the two. The most important aspect of this task it that the goal is to create a map of the existing process, good, bad or ugly. This is not the time to incorporate ideas for what can be done differently. Remember that the start and stop points of the process under study should have been determined as part of the Define phase, so the process mapping effort during Measure should have those same delimiters.

For many processes, tracking cycle time provides valuable information, especially when the problem being addressed is related to delays and process duration. If it is relatively simple to obtain basic data on the process time for key steps, that can be helpful at this stage. Also important is capturing any variation in the way the process is performed, for instance at different times, by different groups, or in special situations. Do not assume that the process is always performed exactly the same way.

Typically an activity flowchart is used to create a visual depiction of the process. Individual steps are shown in order, with decision points and feedback loops as needed, to describe what occurs. Also of benefit for many projects is a deployment flowchart, which specifically illustrates who is performing each step. An opportunity flowchart can be used to highlight steps that are truly necessary and add value to the outcome, and to separate them from steps that represent waste and inefficiency.

Once the process map has been compiled, the project team will review it to glean information about potential contributors to problems and inefficiencies. The team should watch for evidence of missing steps, extra steps, delays and bottlenecks, variation in how certain steps are performed, and anything else that could lead to defects, inefficiency and problems.

As with other aspects of a Six Sigma DMAIC project, it may be tempting to just jump in and start collecting data, under the assumption that the way to do it is obvious. As with the other steps in the process, the only way to ensure success is to do it systematically; with careful planning. For this reason, a data collection plan is usually the first step in the data collection process.

The focus of data collection should be in gathering data that helps to describe the problem, as well as uncovering any factors that provide clues about how, when, where, or in what circumstances the problem occurs or worsened. During planning, the team must not only determine what data to collect, but also ensure that the collection process is valid. In some cases this can be accomplish by establishing clear operational definitions for the measures being tracked; in others a more sophisticated technique such as a Gage R&R analysis is necessary.

Remember that during Define, the project team established one or more CTQs representing the customer expectation for quality. The CTQ measures and specifications will be used to calculate a process sigma score for the baseline process, so gathering data on those specific measures is a primary focus at this point in the DMAIC project.

Projects may require tracking of primary measures such as conversion rate, customer satisfaction rate or cycle time (process duration) or a plethora of other metrics. Typical related measures include the group performing the process (if multiple teams or groups are involved), the time period when the process is performed or the type of product being created or processed.

In most cases it is also important to obtain information on process performance over time. Often technology makes this an easy task, with data automatically being stored on an ongoing basis. In other cases it may take additional work, such as compiling historical reports.

Data analysis during the DMAIC Measure phase involves creating graphs and charts that provide a visual representation of the data, including trends over time. The type of data will determine the type of visual tool that is used.

A first step often involves graphing the data over time, by using a time series plot or control chart. The role of a time plot is to provide visual information about changes in a process over time, including trends and overall amount of variation. If a grouping factor such as department or product was tracked, it is also beneficial to create a stratified time plot, to see if the overall pattern over time differs for the different groups.

A control chart is one of the hallmark tools of Six Sigma. It is both a visual representation and a statistical tool, allowing a determination of which components of variation are inherent in the process (common cause) and which are due to external or specific factors (special cause).

For categorical data such as complaints or types of product defect, a Pareto chart can be used to visually display the distribution of data across categories and to determine whether the 80/20 rule applies. Frequency plots and stratified frequency plots also provide insights into the distribution of data and thus the process performance.

The Six Sigma methodology gets its name from the fact that, at a process sigma (a measure of process variation) of 6.0, only 3.4 defects occur per million opportunities (3.4 DPMO). Another way of looking at it is that 99.9997% of the output would meet customer specifications. Most DMAIC projects do not strive for this level of quality, due to the extensive costs and other tradeoffs involved, especially outside the manufacturing arena. However process sigma is a standard that allows comparison of process performance for different processes. It provides a means of gauging the baseline performance and, ultimately, the improvement that is accomplished through the project.

The percentage of process output that meets the customer specification is calculated (directly, or in some cases indirectly, based on normal theory) and used to look up the process sigma in a process sigma table.

At the end of the Measure phase, the project team should have a detailed process map, detailed baseline data and a calculation of baseline process sigma. Team members should have a clear understanding of how the process is currently done and of the specific problems with the process that may be related to the CTQs.