How to Scientifically Prove the Importance of a Process - CMMI Process Areas

P

patkim

I am in a tough situation where practitioners are outright rejecting processes. I have to implement various CMMI Process Areas but practitioners are not willing to commit at all.

In this case where there’s no past data to make some inference, how can I go about scientifically proving that a process say for example ‘Defect Prevention’ is really going to add a value to their existing software project work?

I can qualitatively talk about it but could there be any scientific way?
Thanks.
 

yodon

Leader
Super Moderator
Unless the team is getting paid for finding and correcting bugs (that they created), I can't imagine why a team would reject defect prevention out of hand.

Maybe it goes a bit deeper? Are they involved in establishing the (defect prevention) processes or are they being told "this is the way to do things because we said so"?

Do you have any data that indicates more defect prevention process is in order?

I don't meant to totally turn this around on you but the above are fairly common paths that lead to push-back.
 

Ninja

Looking for Reality
Trusted Information Resource
... how can I go about scientifically proving that a process say for example ‘Defect Prevention’ is really going to add a value to their existing software project work?

I can qualitatively talk about it but could there be any scientific way?
Thanks.

Are you sure that the defect prevention process is beneficial?
What makes you sure?
Plenty of examples out there where the cure is more painful than the disease.

When you are absolutely sure that the benefit outweighs the cost...distill WHY you are sure.
If it isn't numerical...then you aren't really sure, are you?
If it is numerical...you have your quantitative case.

{Hint: "The boss said so and we'll get fired if we don't" has a numerical answer...the sum of your paychecks...} It is usually pretty easy to turn things into a numerical case.

If you aren't sure enough to defend it quantitatively ... it is no surprise you are struggling to get buy-in. The buy-in should have been early in the process...long before they had a chance to accept or reject.:2cents:
 
P

patkim

Thanks for your responses. I have taken Defect Prevention as an example. It could be any other process as well. I am easy on them and not pushing very hard. I am focusing on 'effectiveness' rather than 'compliance'.

As far as buy-in is concerned, this is just the beginning for me. For them, they may be anywhere in the life cycle of projects as they did not follow CMMI processes before, however I am in early phases of implementation. Company is Adhoc mode of operation for years I guess.

I believe that humans don't like change, and this process change is going to be very painful for them. Hence they are looking for ways to turn it down come what may. It's more of a 'lack of willingness' issue rather than 'awareness'.

I can hit the management but I know management shall favor them and ask me to prove scientifically that CMMI processes help!
 

Ninja

Looking for Reality
Trusted Information Resource
I believe that humans don't like change, and this process change is going to be very painful for them. Hence they are looking for ways to turn it down come what may.

Me reading into your typed words...always a guess at best...

You seem to be approaching this as an Us vs Them situation and placing yourself outside of it.
Change happens easiest when you are part of the team, not the outsider waving your hands around with qualitative ideas.
Change does not "impose" well...as you are seeing. Change requires involved change agents building consensus prior to change beginning.

From your typed words, it looks like you are an outside change agent with no consensus looking for a bigger stick. That is not typically a great approach.
Consider moving inside and laying down the stick...you may find the road easier.

Are "they" looking to turn down the change...or are they looking to turn down you?
What can you do to make them want to follow you, even if it means change? ...then you can start having a team effort...and less barriers.
:2cents:
 

yodon

Leader
Super Moderator
I wouldn't expect any amount of scientific data would sway anyone. Look for champions within the organization. When they own the process and they drive the changes (with good guidance from you, of course), there's greater chance of adoption.

Since they've been ad hoc for years, as you say, I expect there's a perception of being successful this way so there's no motivation to change (if it ain't broke...).

Start by gathering data (number and types of bugs reported from the field, number and types of bugs found in test, number of changes to requirements specs after initial baseline, etc.). But be very careful to not allow the data to be used for punitive actions! That will kill any hopes of change. Ask the stakeholders for suggestions on how to improve the numbers. As Ninja implied, change comes from within.
 
P

patkim

Your responses are very much insightful. I am looking at a philosophical view.

I have attended several trainings on CMMI processes. Some conducted by CMMI experts. One thing I notice is that the entire training is based on the assumption that the model is effective.

But my question is, how one can prove scientifically in an initial adhoc stage of a company that CMMI process is really effective. There’s no past data to infer anything.
I guess one solution is to present some kind of hypothetical use cases. I am looking for your views.

There are countries & customs where processes and discipline is an integral part of life. What I have experienced is a chaotic place where indiscipline is persistent everywhere and there’s lack of management commitment to Quality. So was the intention of the post to seek views on possible ways to prove process effectiveness. If it’s in place, like a verdict, it becomes binding on everyone, more specifically the top management and they can't then run away from it.

Thanks.
 

Ninja

Looking for Reality
Trusted Information Resource
I have attended several trainings on CMMI processes. Some conducted by CMMI experts. One thing I notice is that the entire training is based on the assumption that the model is effective.
...of course...why else would anyone attend?
Just because that assumption is made...does not mean that it IS effective.
...and it also does not mean that it is ineffective either.

But my question is, how one can prove scientifically in an initial adhoc stage of a company that CMMI process is really effective.

How can you prove the effectiveness of something not done yet? You can't.
You can guide people to believe it, or fail to guide them to believe it. Proof needs data you don't have yet.

There are countries & customs where processes and discipline is an integral part of life. What I have experienced is a chaotic place where indiscipline is persistent everywhere and there’s lack of management commitment to Quality.

...sounds like you need to start with Management. If they haven't drunk the Kool-aid, why would anyone else?

So was the intention of the post to seek views on possible ways to prove process effectiveness. If it’s in place, like a verdict, it becomes binding on everyone, more specifically the top management and they can't then run away from it.

Sounds like a place I have no interest in working for. And a place Top management wouldn't want to work. What do you think the chances are of binding Top Management to something without escape? .... especially when you need their OK to move forward?

I've assumed from your command of written english that you are in the USA or similar.
If so...you are following a dictatorial process that is counter-cultural to the extreme. You may (will) want to change your approach.

People can always run away...they may run straight out the door.
Top management doesn't need to run away...they can send YOU out the door.

What is it you are trying to achieve with all of this?
I'm getting the feeling that you want to match the company to the courses you've taken, whether or not it makes everyone miserable or not. It doesn't sound healthy...
 
Top Bottom