Topic
  • 4 replies
  • Latest Post - ‏2009-09-17T19:02:38Z by Kruchten
Julian
Julian
7 Posts

Pinned topic Effort breakdown data for Risk/Value lifecycle?

‏2009-09-17T12:46:30Z |
To help in the refinement of my client's estimation models, does anyone have recent data for effort by phase?
i.e. What % of total effort is spent in each of the four phases.

Likewise, but less importantly, data of effort by classic RUP discipline, across the project, or by phase, would also be useful.

I will publish any analysis of the data I receive.

Regards,
Julian

Julian Holmes
Co-Founder
Unified Process Mentors

T. +44 (0) 7739 338322
E. Julian@UPMentors.com
W. www.UPMentors.com
Updated on 2009-09-17T19:02:38Z at 2009-09-17T19:02:38Z by Kruchten
  • SystemAdmin
    SystemAdmin
    3545 Posts

    Re: Effort breakdown data for Risk/Value lifecycle?

    ‏2009-09-17T16:20:32Z  
    Julian -

    According to Phillipe (circa 2004), I = 10%, E=30%, C=50%, and T=10%.

    From my experience it really depends on the organization. If project initiation is a formal bear, inception may be longer. Other organization have added a pre-inception or discovery phase for some of these activities too.

    If there is a lot of compliance and red tape to release into Production, transition may be longer.

    There is also some variance in how an organization defines the LCA milestone in terms of sound architectural stability. I guess it also depends on the project, but for big important ones I prefer to see it demonstrated in working code on a facsimile of the production environment with metrics to prove the non-functional requirements and any other SLAs are met. This may increase the % in elaboration....35-50.

    As far as a discipline breakdown goes...I seem to recall seeing some notes on this, but can't seem to place them. If only it was easy to calculate the surface areas of those humps in the classic chart and do the math, you might get some ideas around the original intent on effort per discipline per phase. I'd be interested in any stats around this.

    Carson
  • Mark.Lines
    Mark.Lines
    56 Posts

    Re: Effort breakdown data for Risk/Value lifecycle?

    ‏2009-09-17T16:34:16Z  
    Hi Carson,

    Your percentages are actually Schedule, not Effort. The Effort percentage "guidelines" are

    I=5%, E=20%,C=65% and T=10%

    The difference from the Schedule percentage arise from the assumption that you have variable resourcing throughout the project. i.e. a small team in Inception, increasing to a maximum size in Construction, and then reducing in Transition.

    I think we all realize that these are merely guidelines and can vary greatly depending on project specific factors such as risk, complexity, uncertainty etc.

    Mark

    Mark Lines
    RUP Discussion Facilitator
  • Mark.Lines
    Mark.Lines
    56 Posts

    Re: Effort breakdown data for Risk/Value lifecycle?

    ‏2009-09-17T16:46:53Z  
    Julian,

    I personally believe that estimation by Discipline is a bit archaic. I would prefer to resource the team with the classic mix of skills, looking for "generalizing specialists" (Ambler). Using a modern self-organizating teams approach makes estimating by discipline a waste of time.

    So use whatever high-level technique you wish to in Inception (doesn't really matter what technique,the numbers will be wrong anyway). Staff a team, make assumptions about scaling the team by Phase. Fund a couple of iterations of Elaboration, measure velocity, extrapolate based on experience and remainder of functionality to be implemented, and you then have a credible estimate at the end of Elaboration. You can then fix-price Construction if you like.

    See you in a couple of weeks at the IBM Rational UK conference :)

    Mark

    Mark Lines
    RUP Discussion Facilitator
  • Kruchten
    Kruchten
    7 Posts

    Re: Effort breakdown data for Risk/Value lifecycle?

    ‏2009-09-17T19:02:38Z  
    Julian -

    According to Phillipe (circa 2004), I = 10%, E=30%, C=50%, and T=10%.

    From my experience it really depends on the organization. If project initiation is a formal bear, inception may be longer. Other organization have added a pre-inception or discovery phase for some of these activities too.

    If there is a lot of compliance and red tape to release into Production, transition may be longer.

    There is also some variance in how an organization defines the LCA milestone in terms of sound architectural stability. I guess it also depends on the project, but for big important ones I prefer to see it demonstrated in working code on a facsimile of the production environment with metrics to prove the non-functional requirements and any other SLAs are met. This may increase the % in elaboration....35-50.

    As far as a discipline breakdown goes...I seem to recall seeing some notes on this, but can't seem to place them. If only it was easy to calculate the surface areas of those humps in the classic chart and do the math, you might get some ideas around the original intent on effort per discipline per phase. I'd be interested in any stats around this.

    Carson
    yes.
    The I = 10%, E=30%, C=50%, and T=10% was an observation, not a rule, not a constraint.
    This was for some middle of the road new projects. Use it as a baseline for your own reasoning on your own proejct in your own context.

    Elaboration may go down to zero, or transition in some cases.
    What really matters is: what do you want to have achieved at the major milestones.
    Then you need to plan your milestones appropriately.

    In short: this (I = 10%, E=30%, C=50%, and T=10%) is not useful for planning or estimating.

    Philippe Kruchten

    University of British Columbia
    Vancouver, BC, Canada
    pbk@ece.ubc.ca