Hi, I am working as a business analyst in my organisation and was having a discussion with my manager related to productivity of teams using fucntion point analysis. The problem that we were discussing is given below - Lets say we have a vendor software package(third party software suite) which needs to be deployed at multiple sites ( say various regions in this case). Function point provides measure of the system functionality from a user's point of view (logical view). Let's assume this vendor package( like a software suite) has got 100 function points. The management decides to deploy this package to 3 different regions say Asia, Europe and Australia. When this package is deployed in Asia, the team working there took out 20 functionalities out of the package since they were not suiting local regulatory acts. In addition, they also developed 20 new functionalities and it was added to the existing package. ( overall effort - 80 man hours) When this package was deployed in Europe, they also custom tailored the existing vendor package in order to suit their needs. They added 20 new funtionalities and began using the application.( overall effort - 60 man hours) Australian team decided to DEPLOY only 90 fucntionalities from the original package and did not add/modify/delete any of the functionality. (overall effort -50 man hours) My management wants to come up with productivity figures for each of the above regions. Productivity in terms of function point is calculated as Function points delivered divided by actual efforts. Each of these team had to deploy the overall vendor package first and then only they could make any changes in that. If we look from a user's point of view, scope of the application has changed for all of the above three examples. There would be a variance in productivity figures for each region. As per function point counting guidelines, vendor package deployed in Australia would contribute very less function point ( perhaps zero) as compared to other two regions. But my management is not able to understand the mismatch in actual efforts put for deployments versus low function point count. Can anyone please tell me what approach/ sizing method should be used for these kind of projects? Thanks, Anshul
Hi,
I am working as a business analyst in my organisation and was having a discussion with my manager related to productivity of teams using fucntion point analysis. The problem that we were discussing is given below -
Lets say we have a vendor software package(third party software suite) which needs to be deployed at multiple sites ( say various regions in this case). Function point provides measure of the system functionality from a user's point of view (logical view).
Let's assume this vendor package( like a software suite) has got 100 function points. The management decides to deploy this package to 3 different regions say Asia, Europe and Australia.
When this package is deployed in Asia, the team working there took out 20 functionalities out of the package since they were not suiting local regulatory acts. In addition, they also developed 20 new functionalities and it was added to the existing package. ( overall effort - 80 man hours)
When this package was deployed in Europe, they also custom tailored the existing vendor package in order to suit their needs. They added 20 new funtionalities and began using the application.( overall effort - 60 man hours)
Australian team decided to DEPLOY only 90 fucntionalities from the original package and did not add/modify/delete any of the functionality. (overall effort -50 man hours)
My management wants to come up with productivity figures for each of the above regions. Productivity in terms of function point is calculated as Function points delivered divided by actual efforts. Each of these team had to deploy the overall vendor package first and then only they could make any changes in that.
If we look from a user's point of view, scope of the application has changed for all of the above three examples. There would be a variance in productivity figures for each region. As per function point counting guidelines, vendor package deployed in Australia would contribute very less function point ( perhaps zero) as compared to other two regions. But my management is not able to understand the mismatch in actual efforts put for deployments versus low function point count.
Can anyone please tell me what approach/ sizing method should be used for these kind of projects?
Thanks,
Anshul
Anshul,
Function points are a good way to assess the size of the project in "effort" only. Now various teams would perform differently in delivering these function points. A real ace "French speaking" team would do much better than an English team trying to "Frenchivy" your software for the European edition.
I think I shared with this "group" that I once was fired because of functions points. I told my management that they grossly underestimated the function points and that we had the wrong mix of people to deliver the function points. My client eventually employed me 3 months later, once they realized that I was correct in my estimation of the project size and mix of people.
My point, you need the right mix of people to deliver the right number of function points.
warm regards,
K
I really appreciate your inputs on my business problem. However, if you suggest function point analysis would not be an ideal approach for calculating productivity where we don't have right blend of people, what should be the best way to size a project? What other sizing techniques could be useful in this scenario?
Your approach is OK.
The issue with function points is its an estimation of SIZE. At some point you have to have the right mix of people to work on these function points. Let say you've stimated 100 FP and your resource consumes these FP at a rate of 2 per day; so 50days later (+- factor) you deliver. However, if you have a resource that consumes these FP at a rate of 2.5 per day it takes about 40days to deliver. On the project that I referred to earlier, we had people with management information systems backgrounds (BAs who also wanted to program); what we needed in the end when we had to write communication protocols (a complete ISO stack), were computer science graduates. People who could write compilers, middleware and database gurus. Once these people were engaged our FP consumption rate improved significantly, and 2.5 years later with about 250 IT resources, we delivered.
Occassionally I go back to software estimation as a pet project and recently bought a book published by Infosys or the Infosys people were the main authors. Just cant get to it at the moment. But, the attached URL might shed some light on the issues that you may encounter.
http://www.infosys.com/Oracle/white-papers/Documents/estimating-size-package-points.pdf
All the best,
PS. I dont work for Infosys or have any relationship with Infosys.
brought to you by enabling practitioners & organizations to achieve their goals using: