Tuesday, May 31, 2011

I have some questions about Requirement Development with the CMMI....


    Can we merge specific practices of REQM and RD in one process guide or it is required to clearly defined each process area separately?
·         Related to SP 1.1 of RD
o   What is the entry/input for SP 1.1 of RD?
o   what exact task has to be performed for this SP?
o   Is this SP conducted  once project is awarded or part of Proposal submission?  
·         What is the best practice, to define Customer Requirements and Product requirement in single document and get a sign off from customer as whole or first get customer commitment on Customer Requirements and  then proceed to Product requirement and its sign-off?
·         We don’t have any specific practice right now, so what is the best practice to go for?

All good questions.

Can we merge specific practices of REQM and RD in one process guide or it is required to clearly defined each process area separately?

  • of course.  The most important thing is that it reflects how you actually do business and provides a baseline process so you can improve it over time.

What is the entry/input for SP 1.1 of RD?

  • The entry point could be many things.  A Request for Proposal, a Statement of Work, an initial set of requirements, and "needs meeting" between principles.  It depends, again, on how you do business

What exact task has to be performed for this SP?

  • Hard to say because I don't have insight into your business, but the prior answer should help you out.  Whatever it takes to accept, review, and understand those things.

 Is this SP conducted  once project is awarded or part of Proposal submission?  

  • Again, it depends.  But it often happens early in the project, or even during the RFP phase.  If you're an agile/scrum shop it could happen early in each sprint and release.


What is the best practice, to define Customer Requirements and Product requirement in single document and get a sign off from customer as whole or first get customer commitment on Customer Requirements and  then proceed to Product requirement and its sign-off?

  • I prefer to see to two stages.  A work product that defines what the customer asked for ("their needs") and approved, and another one that says "this is what we plan to build" that also gets approved.

  We don’t have any specific practice right now, so what is the best practice to go for?

If you're asking what "methods" you should use for this, there is no single method that solves this for you.  These practices are very specific to your business, and you should be asking yourself "what is working today?" and start there.  Then you can use the CMMI model to improve it.  You might look into various methods to help you, like Scrum, Planning Poker, Wide-Band Delphi, JAD workshops, mind-mapping, or AgileCMMI as tools to help you.

Good luck!


http://www.broadswordsolutions.com


Saturday, May 7, 2011

We would like to become CMMI ML3, but my boss says it's only for big companies.

Dear Appraiser,


We would like to adopt CMMI and maybe even become ML3 someday.  My boss says he likes the idea, but that the CMMI is only for big, huge companies because of all the overhead.  What do you think?

One of the things I love about the CMMI is that it's a learning experience for everyone.  I often spend weeks educating the management team with a new client, but fortunately on this issue, we don't have to do that. We have data!

The SEI has just released their March 2011 Maturity Profile Report and one of the pages depicts CMMI appraisals by organizational size.

CMMI used to be the domain of big companies, but the data now tells us that over 50% of adopters have fewer than 100 employees, and that 18% have fewer than 25!

Oh, and by the way, CMMI doesn't add overhead - overzealous managers and process people do that.  But don't get me started on that rant!


Friday, May 6, 2011

What's the right amount of sub-processes to statistically manage for ML4?

Dear Appraiser


in order to move from ML3 to ML4 we have understood that processes selected for OPP must be all the processes the organization has to statistically measure in order to reach its business goals.  Is it right or there is a minimum set of processes to select for OPP that is judged adequate for a ML4?

Please another two questions:
  • do you think using Six-Sigma is necessary in order to move from ML3 to ML4?
  • do you think a 20-months PI project is too long in order to move from ML3 to ML4 for a large organization (1000)?
Hmmmm.....interesting question.  Since the inception of so-called "high-maturity" people have been struggling with this. How many is enough?  What are the right sub-processes to analyze?  What will a Lead Appraiser and the SEI "accept?"  The answer that has been broadcasted loud-and-clear by the SEI is "it depends!"

As frustrating as that is for most people, I've come to be comfortable with it (I have not always had this opinion btw).

This is one reason you should avoid hiring a cheap Lead Appraiser last minute for the minimum amount of days, because his/her perception of your organization's needs, goals, objectives, and behaviors is critical to helping to lead a team through that judgement call.

It really does come down to how you are running your business, and is the process you are using, and the data you are collecting, really credible.

I always tell my clients they should focus on "being a great company" and stop worrying about CMMI.  If they focus on that, they WILL reach their CMMI goals - but usually not without some honest self-evaluation and hard work.

If you build static web pages in a small-setting for a living, the number of sub-process that are being statistically managed might be small - say 3-4.  If you're building the Atlas rocket for NASA with 5000 engineers, you probably have a lot more - possibly 20 or 30.  It's all about context, goals, objectives, strategy, risk, and customer satisfaction.

Your best bet is to engage a Lead Appraiser who really understands your business model, knows you well, and who can help you make an informed choice.  

As to your other questions, Six Sigma certainly can HELP you in your quest to become a great company through the use of statistical analyses and methods, but it's not required.  Six Sigma brings some structure to the process that may not exist, but it's only one answer.

I can't really address the timing issue - 20 months seems credible, but without knowing more about your organization it's really hard to say.  Take a look at the "Time to Move Up" chart I posted a few days ago from the SEI.  It's an analysis of what other companies have done.

Good luck!

Wednesday, May 4, 2011

If you don't already follow us on Twitter . . .

Yes!  Ask the CMMI Appraiser is on Twitter!

If you're one of our hundreds of followers you're already getting our daily tweets packed with CMMI knowledge.

Follow us today at "CMMIAppraiser"

www.broadswordsolutions.com

Tuesday, May 3, 2011

How long does it take to move from CMMI Level 2 to CMMI Level 3?

Dear Appraiser,


We are a CMMI ML2 company that wants to move to CMMI ML3. How long does it take?

How long does it take?  How long does it TAKE?  Let me ask you this, how long does it take to cook a meal?  At McDonalds it takes 3 minutes.  At The Lark in West Bloomfield, MI, a 5-star dining establishment, it takes 2 hours..... but it's worth it!

But still.....inquiring minds want to know!  In the end it depends on many factors - where you are, where you are going, how you are going there, and so forth.  But it so happens that the SEI has people in the basement that compile this sort of data.  So here it is.



Good luck!

www.broadswordsolutions.com

Monday, May 2, 2011

Slowdown in CMMI Appraisals? I don't think so.

The SEI's data on the adoption of CMMI tells us that, even though there was a slowdown in 2008 due to the recession, the trajectory continues it's upward climb.  Download the entire report at http://www.broadswordsolutions.com.