A
vital methodology” DMAIC Practice” also ensures completeness of
testing. During the Define phase, knowledge is gained about the
business problem/request. In the Measure phase, the testing
magnitude/requirements are determined. The test cases and plan are
developed during the Analyze phase. In the Improve phase, testing is
completed to provide a business-as-usual environment. In the control phase ongoing monitoring is undertaken to ensure the defined and implemented process continues to reap maximum benefits in terms of overall test coverage as well as effectiveness in the execution cycle.
Defined testing roles are used to execute the software testing process.
Each phase of the DMAIC process is initiated by defined inputs and
outputs. Each phase goes through a tollgate process to determine if the
next phase should be initiated. Our DMAIC approach allows us to address
the variability of the testing process with a prescription for success.
Improvement is done in five steps:
Define
The basic steps to the Define phase of DMAIC are as follow:
 |
Identify or refine the problems in your process that must be solved in
order to meet or exceed the customer's specifications or expectations |
|
Identify and quantify customer requirements |
|
Identify and quantify the process and defects that fall short of these requirements and create a problem statement |
|
State the project goal, which also must be a clear and measurable goal, and include a time limit for the project's completion |
|
Determine the few vital factors that are Critical to Quality, which need to be measured, analyzed, improved and controlled |
Measure
Select the Critical to Quality characteristics in your process.
| |
Define what that process output should be, which is done by looking at the customer requirements and the project goal |
|
Define
defect for the current project at hand. Remember, a defect is an output
that falls outside the limits of customer's requirements or
expectations and must be measurable |
|
Find the inputs to the process that contribute to defects |
|
Define the exact dollar impact of eliminating the defects in terms of increased profitability and/or cost savings |
|
Measure the defects that affect the Critical to Quality characteristics as well as any related factors |
|
Incorporate Measurement Systems Analysis - a method to make sure the defects are being measured properly |
Analyze
| |
First, develop hypotheses about the causes of the defects |
|
Then, analyze the data and process by statistical and non-statistical methods |
|
Finally, prove or disprove the hypothesis |
|
If the hypothesis is correct, add it to the short-list of causes. If not, refine it |
Improve
In this phase of DMAIC, we will:
| |
Confirm the key process inputs that affect the process outputs, causing defects |
|
Identify the acceptable range of each input so that the Critical to Quality output stays within the specified limits |
|
Adjust the process as needed |
|
Implement the changes |
|
Install and validate a measurement system for the improved process |
|
Verify the new process is working |
Control
The final phase of the DMAIC process is control. There are many factors
that could affect the adjusted inputs and, thus, the output; so ongoing
monitoring of the process to make sure it stays "in control" is
critical. In most cases, this is done for a limited amount of time by
the Six Sigma team or the Black Belt and then handed off to the process
owner. The Control process involves quality and statistical concepts
that have existed for decades. However, the advent of quality control
software makes the process very disciplined for the tester community
to adopt and implement
The phases allow for:
| |
Resource identification, knowledge building and availability |
|
Processes to test and ensure that the customer request is fulfilled |
|
Channel
implications: The overall measurement has allowed deliveries of
business-as-usual environments that meet the requirement as defined by
the customer and place the metrics in everyday environment for ongoing
measurement, improvement and continuous accuracy |
|
