Monday, August 4, 2008

Understanding Causal Analysis and Resolution (CAR)

In the context of the CMMI, CAR is a Maturity Level 5 Process Area. You can perform CAR qualitatively at lower Maturity Levels, but you cannot realize the full benefits of CAR until you have stable processes and statistically understand the process capabilities. Let’s step through the practices one-by-one and talk about what they mean from a quantitative perspective;

SP 1.1 Select the defects and other problems for analysis.
You can perform this practice very simply and qualitatively just by having a gut feel that something is wrong and you want to investigate its causes. However, you should be using data and statistical tools such as Pareto analysis, histograms, or process capability analysis to determine which defects to analyze.

SP 1.2 Perform causal analysis of selected defects and other problems and propose actions to address them.
You can perform this practice qualitatively just by meeting with those people responsible for performing the process, using tools like cause-and-effect (fishbone) diagrams or check sheets. From a Maturity Level 5 perspective, you may have criteria that states you perform causal analysis when a stable process does not quantitatively meet its specified quality and process performance objective.

SP 2.1 Implement the selected action proposals that were developed in causal analysis.
Again it is possible to perform this practice qualitatively simply by implementing an action as the result of the cause-and-effect diagram. From a Maturity Level 5 perspective, you may be designing an experiment to see if the proposed action will have the desired quantitative results.

SP 2.2 Evaluate the effect of the changes on process performance.
This practice will be difficult to perform qualitatively. The intent of this practice is to measure the effects of the change on the process performance, which can only be done accurately if you have stable and capable processes in place and you are using quantitative methods to manage these processes.

SP 2.3 Record causal analysis and resolution data for use across the projects and organization.
If you are performing qualitative causal analysis you may not have any data to record other than results from the cause-and-effect diagram and rationale for the change. From a Maturity Level 5 perspective you might have data on the defects and other problems you analyzed and measures of the changes to the process performance resulting from the causal analysis.

6 comments:

rozz said...

We are trying for level 3 ,but for our assessment review a point was raised that we havent done any Defect Analysis for Reviews,Defects, and NC's.In which process area does this come in level 3? MA/Validation.Can we really do this in ML3?

Henry Schneider said...

Hi Rozz,
When you read the VER and VAL Process Areas you will discover that VAL SP 2.2 states "Analyze the results of the validation activities" and VER SP 3.2 states "Analyze the results of all verification activities." These two practices are for analyzing the results of reviews, inspections, tests, analyses, demonstrations, etc. But the model does not explicitly require these specific items. The specific items you analyze are a function of what your processes and procedures specify as well as the criticality of the project. One important point to note is that neither VAL nor VER require root cause analysis. These two process areas are focussed on the project and product meeting the requirements. Causal Analysis and Resolution is focussed on identifying and analyzing causes of defects and other problems and taking specific actions to remove the causes and prevent the occurrence of those types of defects and problems in the future

Anonymous said...

Can you please explain integration of OPP, QPM with CAR and OID - pictorially if possible or even a explanation would help. Thank you.

Anonymous said...

To satisfy Level 5, is it mandatory to show the removal
of specific causes and prevent the occurrence of those types of defects and problems in the future resulting in improved PPBs?

Henry Schneider said...

This is a question that is way beyond the scope of this blog. It cannot be answered in a picture or short paragraph. If you are truly interested in how the four HM PAs interact, I suggest that you take a class in High Maturity practices, which will be 2 to 3 days in duration.

In addition, the PAs you listed are the v1.2 PAs, which have changed for v1.3. The new list of PAs is OPP, QPM, OPM, and CAR. I also strongly recommend that you read the four PAs and then ask a specific question rather than this broad question.

Henry Schneider said...

Asking what it takes to satisfy Level 5 is not the correct question.
Instead you must answer for yourself, what is the purpose of Causal Analysis and Resolution?
In the context of the CMMI there are three components: Required, Expected, and Informative. The Required components are the Goals, the Expected components are the Practices, and the Informative components are everything else. If by mandatory, you mean required, there are no CAR goals that require removal of special causes and prevention of recurrence, resulting in improved PPBs. And what do you mean by special causes?
Do you mean special causes of process variation? Or are you talking about something else entirely?

To meet the intent of CAR, as a High Maturity Lead Appraiser, I would expect that you are regularly examining your list of defects, issues, process problems, and unexpected successes using your PPBs and PPMs to select those items that have the biggest impact on achieving your QPPOs. Performing causal analysis on these selected items and taking appropriate corrective actions to address the root causes. Then using your PPMs to predict the results of the corrective actions, comparing actual results to predicted results, and if they don't match, performing addition causal analysis to determine why.
I would then expect to see a change in the appropriate PPBs and PPMs, though the change may or may not result in an improved PPB.