h1

Can we change?

February 22, 2013

Over the past few years, process improvement in general and CMMI® in particular have been discredited on a regular basis; in the past year, we have had a big change, which could mark a renewal, or not. So will this year, bearing the favourite number of all superstitious people (13) be lucky or unlucky for our dear old model?

First, I believe that the move from the SEI is a very positive move. Not only does it remove the model from US government intervention (they still have oversight of the SEI, even though they no longer fund the CMMI), but, more important, it moves us away from the “S” of the Software Engineering Institute. CMMI is still largely seen as a software industry product, even though it has proven its value in many different engineering and service based organizations. But that is not the main issue with the model.

We have been plagued for a long time with consultants, lead appraisers and users who have mistaken the concepts of “maturity” with that of “maturity level”. I have heard senior managers who have told me with a straight face that if they just do everything the model says, they will produce high-quality products – as if the CMMI was a recipe rather than a model.

Maturity is basic understanding of your own limitations and abilities. A maturity level is only one way of measuring that. Aiming for a maturity level rather than for true maturity is a little like a child looking forward to the day he turns ten, because he will be a big boy then: it is, in itself, a demonstration of a lack of maturity. Of course, there are always lead appraisers who are willing to avoid seeing the evidence, who will be satisfied that you are producing a lot of documentation, that you have copied over great chunks of the model and labeled them “policy” to allow themselves to be convinced that you have achieved something. They are happy with a “quality assurance” team which does not assure the quality of the process, but merely controls compliance to model terminology; a large collection of numbers will satisfy them that, if you are measuring, you are probably analysing as well (they even saw a graph as to how late the project was last month, surely that’s analysis, isn’t it?)

My wish for this year, which I would hope you could share with me, is that we start taking the model as a process improvement tool and no longer as a rubber stamp. I wish the people using the model, or not using it, would understand that we are actually trying to improve your business rather than satisfy some theory redacted in a distant university. As such, I would like to see:

  • Policies that reflect senior management’s expectations in terms of outcome and results rather than tasks;
  • Measurements that reflect the organizational objectives, including customer satisfaction, number of defects, time to market, over-time, personnel turnover, efficient use of resources, cost of quality, and so on;
  • Risk management which is based on a quantitative understanding of the value and cost of risks;
  • Planning which is based on an analysis of what is truly needed by the organization in order to deliver;
  • People taking responsibility for the quality of their work;
  • An effort to learn from mistakes (which requires acknowledging them) and successes – and near-misses – at every level of the organization (including sales!);
  • A sharing of knowledge throughout and across teams;
  • The acknowledgment that CMMI, ISO, ITIL, Prince2, Agile, Lean, TickITplus, and all the others, are only tools and not objectives.

If we could identify this, and communicate, whenever necessary, that fraud has been committed in the name of the process improvement, then, I believe we can truly start looking at being a useful industry rather than just a consulting industry which seeks to sell itself for as much money as possible.

Getting a Maturity Level or a certificate in any standard does not lead to customer satisfaction. If you cannot deliver, you lose your customers, you lose your reputation and you damage the reputation of the standard you are claiming to have used. If you provide unwarranted maturity levels, you are killing your own reputation and that of your industry.

I wish you the joy of looking back and seeing that you have made a difference for good in someone’s life.

7 comments

  1. This is a breath of fresh air Peter. I have left CMMI in my dim and distant past as I had lost faith in its usefulness and became totally disillusioned so I certainly fully support your sentiments. Agile seems to be going the same way i.e. being ‘criticised’ by its implementation and inappropriate usage as opposed to its benefits. As some bright spark once said ‘All models are wrong. Some are useful’


  2. I love CMMI just like the sun flower loves the sun. Even in the darkness of PMO, I still try to search the lights from QMO wherever it could be.


  3. CMMI is string of string of notes, it replies on the composers to make them into works, the performers to play them…


  4. Sometimes I think if there are SEI authorized quality managers and process managers deployed into those CMMI level companies, maybe they will help keep the reputation of CMMI along with theirs. I don’t understand why PMP is still so famous while so many projects are not even considered to be delivered on time by those PMP’s.


    • Hello Betty – there is actually talk (and has been for some time) of having “certified process professionals”. People who are not lead appraisers, but are certified by the CMMI Institute (no longer SEI) for consultancy, quality, etc.


  5. Many thruths are in this article, Peter.

    Weak spot: It also indicates that many managers did not understand what CMMI is all about alas. Training without putting knowledge to the test is risky. The proof of the pudding is in the eating …

    Strong points: I can confirm, it is clear CMMI is not just about software and an SEPG TEAM, it can be any product, so it is about engineering and productEPG in CMMI DEV and about any offered service in the CMMI for services …

    Great article, thanks


  6. Thanks Peter for putting things into their perspective. There is much wisdom in the CMMI (as there is many model), but it’s a model and applying it in a sensible way is what brings benefits.

    Working already for quite some years with Agile, Scrum and Lean I see similar patterns that I have seen in the past (and sometimes recently) with TQM, CMM(I), People-CMM and many other models. Real change comes from understanding why things are needed, and by communicating this to those involved. By empowering people with experience and skills. And not from just deploying a model (and certainly not if that is done “by the book”).

    The model is a language that can help to decide what can help you, and sometimes how to do it. But models don’t tell you why you should do it, you should know that before you use any model.



Leave a reply to Paul Hookham Cancel reply