Easily ignored certain key points in CMMI implementation

Proper interpretation and understanding of a model is a vital part in any new model implementation in an organisation. During the course of action, we have a tendency to over simplify things and thus ignore certain key aspects. And in certain cases there happen some misconceptions too. This article tries to clarify such ignored aspects and the next article will talk more on misconceptions in CMMI implementation. There might be some overlapping between these two articles.

The below aspects are easily skipped, not because they are too tough to implement but because, they are too simple to implement.

  • GPs – Ignored
    • The Generic Practices (GP’s) represent the most important part of the model. Poor implementation of GPs will always lead to process failure. Usually we will try to map all specific Practices (SPs) of each PA and simply ignore GPs. In some cases GPs will be mapped automatically through SP implementation itself, but not in all cases. GPs are the most beautiful aspect of the CMMI model as it talks about institutionalising the process
  • CM – Oversimplification
    • The Process Area (PA), Configuration Management (CM) appears clearly straight forward. Because of its over simplicity, CM audits are found to be skipped or less focused. CM audits must be conducted on a periodical basis to ensure the integrity of work products.
  • MA – Avoid Parts or Overkill It
    • We try to collect the full set of data which are measurable. And when a question arises regarding the usage of the same, we search in dark.   In fact we are wrongly focusing on data collections which are measurable rather than thinking about what information we really need. In this case, the first half of this Process Area is entirely skipped – Defining information needs and measurement objectives
  • PP – Skipping on Size estimation and Risk Management
    • We normally skip the process of doing size estimation or don’t spend enough time finding a good use of size or attribute estimation (SP 1.2). And finally the project size is projected in terms of hours or man months instead of Lines Of code or FP estimate.
    • Project Planning PA talks about early risk identification and mitigation (SP 2.2). These risks are identified in the early stages of project life cycle, most probably in the business development stage. Usually these risks are not properly transferred from to the project team from Business team and finally it  leads to problems
  • PPQA- Lack of focus on Product audits
    • Independent Work Product Audit (SP 1.2) is equally important as that of Independent Process Audit. But normally it is seen that process audits are done on a regular basis, but product audits done very rarely or spend very little time on product audits.
  • GP and related PAs – Over simplification
    • There exist some recursive relationships between generic practices and their closely related process areas. For example, in the case of GP 2.5 for OT, we try to map it with some common OT related records instead of thinking it as ‘training on training”. But here it talks about training for trainers, training for training coordinators etc. For more information please Refer How does the Generic Practice recursively apply to its related Process Area(s)?
  • Units and granularity – Ignored
    • The last but not the least is about units and granularity of data. Every mathematical calculation is done to the perfection. But often forget to add units. Finally when there are some comparison tests executed on the two data sets (e.g. two sample T test) chances are there to have inappropriate results especially when the data sets are placed without units (for example, might be comparing time in seconds and that in hours). Similarly to reduce rounding off errors, data should be at least one level more detailed than the goal metric (e.g. if goal is 0.5 then data should be captured to 2 decimal places). On the other side, there is no value in increasing the decimal points when the source data itself is not error free.