DMAIC Phase 4: Improve

In the first stage of Improve it is important to include the people who are involved in performing the process. Their input regarding potential improvements is critical, and this step should not be completed by the project team alone. In fact, it is wise to maintain communication with those who work on the process throughout any Six Sigma quality improvement project.

A variety of techniques are used to brainstorm potential solutions to counter the root cause(s) identified in Analyze. Encouraging participants to challenge rules and assumptions, ban excuses and think like small children can be very effective. For those who prefer a more structured brainstorming exercising, specific techniques are available, but often participants are more than able to produce a substantial list of ideas on their own.

It is important during this stage that ideas not be judged nor eliminated. Even an outlandish idea that couldn't possibly be implemented as first suggested may lead to a related idea that is an ideal solution. Similar to other aspects of a Six Sigma project, assumptions about what can or can not be accomplished should not be accepted without confirmation.

As in the prior step, it is a good idea to involve the people who work on the process that is being improved, be included in the decisions regarding which potential improvements to implement. With their help, the project team establishes criteria for evaluating the proposed improvements in an objective manner. Criteria usually include time line for implementation, financial cost, the extent to which root causes are likely to be countered and the overall ease of implementation. Some teams consider other factors such is the amount of buy-in that already exists for each possible change.

Not all criteria are created equal, so the team may want to assign weights to each criterion prior to evaluating the proposed solutions against each one. Tools to assist with the evaluation include a priority matrix and a Pugh matrix, both of which use basic calculations and ratings to compare the solutions against each other or against a standard. Occasionally a computer model or other simulation can be beneficial in the evaluation process.

The goal of this step is to determine the appropriate solutions to implement using objective means, rather than making a decision based on assumptions or preferences. This is a common theme throughout the Six Sigma methodology.

Planning the implementation is largely a matter of basic project management. The team needs to plan the budget and time line of the implementation, determine roles and responsibilities, and assign and track tasks. Tools for planning include Gantt charts, planning grids and flowcharts. A deployment flowchart can be created for the implementation process itself, as well as for the new process that will be followed as a result of the improvements being implemented.

A data collection plan should be created similar to the one used during Measure, and the same data should be collected. After the data is collected the team will compare the before and after data to determine if the key metrics show improvement.

It is often beneficial to use Failure Modes and Effects Analysis (FMEA) before implementing improvements to identify and address potential problems that may arise using the improved process. With this tool, the team lists risks and potential issues, and estimates the likelihood and severity of each one. Then the most critical are identified and the team establishes a plan for minimizing each risk.

One aspect of implementing improvements that is often overlooked is the impact of change on the people that are involved in and affected by the process. Basic change management procedures should be followed to smooth the way: communication, seeking input, and ensuring the necessary level of commitment from key players.

During the implementation itself, the team should be monitoring the process and act to address any issues that arise. In addition, the data should be reviewed periodically to ensure that appropriate data collection procedures are being followed.

For some DMAIC projects, it is appropriate to pilot the improvements before proceeding to a full roll out. The most common piloting options include either making changes only in one group or department or making changes for a limited time period. The benefit of a pilot test is that the project team can ensure the changes result in the desired improvements before a full roll out. In addition, the team can gain insights to allow a more effective implementation during the full roll out.

Whether evaluating the pilot results or the full roll out results, a variety of techniques are important for assessing the extent of improvement. Perhaps the most important is recalculating the process sigma, so that it can be compared to the baseline process sigma established earlier.

Also common are frequency plots or Pareto charts to show before and after data. And the hallmark tool of DMAIC improvement projects, the control chart, is often employed to show the reduction in variation and improvement in performance. Both before and after data are plotted on the same chart, and the control limits are calculated and depicted separately for the two stages. In all cases, statistical tests are typically used in addition to graphs and charts. The same tests that were used in Analyze can be used in Improve; ANOVA, regression and chi-square testing are common.

By the end of the Improve phase, the project team has demonstrated that the solutions implemented do in fact counter the identified root causes and thus result in substantial improvement in the CTQ metrics. The new process is in place and the team is ready to create a plan to maintain the gains and close out the project.