| Author | Message | ||
     Peter Hudson |
Can somebody refer me to websites (or other resources) where I can find good definitions for application maintainance productivity? I suppose that there is an indicator like number of full-time equivalents needed to maintain 1 FP in production per year? | ||
| Ashok Kumar |
Peter, Whether it's a development or maintenance, I use the following definition. Productivity = Size / Effort Examples: FPs per PersonMonth You can also use- FPs per PersonYear May be you already know it. Be cautious while comparing productivities between two applications in different domains. Ashok Creator of RASS Estimate www.rasstools.com | ||
| Peter Hudson |
Hi Ashok, Thanks for your input. I was not very clear by what I meant by definition. My case is for measuring application maintainance "productivity" - if we can call that. In that case, it is the size of the application compared to FTE. But the size of the application is ever changing, and so do we measure it once a year and keep it constant in the calculation or do we measure it each time the application evolves, or is there yet another approach? All ideas / thoughts are welcome. (Message edited by peter on August 31, 2009) | ||
| Ashok Kumar |
Peter, The approach to baseline the size may vary from situation to situation. I'll give consideration to the following. - Normally, maintenance should not change the size of the application. Change in size of the application is related to enhancements, re-engineering and re-configuration of sub applications or modules. I have no idea of your situation. - Ignoring whatever may be the reason for the change in size, if the change is within some limit say 20%, you may re-baseline the size at the beginning of the period i.e. say one year. If change is more than 20%, you may reduce the period to re-baseline to six month if it is practical and cost effective. The 20% figure is hypothetical. or - You may observe the variation of change in size over a period and calculate the mean and the standard deviation. Now instead of reducing the period, you may calculate the range of productivity by using the mean and standard deviation of the size change pattern. In this case Productivity may look like x±y%. Ashok Ashok Creator of RASS Estimate www.rasstools.com | ||
| Peter Hudson |
Thanks Ashok for your feedback. Just for info: the size change is certainly due to enhancements or projects. But once the enhancements are delivered in production, they come under the umbrella of application maintenance. That is what I referred to when I said that the size of the application is ever changing. I get your point, + reccos. Thanks. | ||
| Steve Neuendorf |
Hello Peter and Ashok, In maintenance, the use of "productivity" is a little bit sloppy. One of the common definitions is to express output as a function of input. FTE is clearly an input measure, but a application or portfolio size has a tenuous relationship, if any, to any sort of maintenance "output". Variation in size of the installed base over a period of time certainly would not have any relationship to variations in productivity. I just finished listening to an "Output Based Pricing" webinar put on by the Everest Group, where part of the discussion was around maintenance output measurement. You might find that useful, and you could probably get a recording and the materials. Here is the contact I have from the announcement: Kelly Stringer at: kstringer@everestgrp.com I would also recommend using a queuing model to measure maintenance productivity. I also wrote an article on this subject and you can find it at http://userwebs.serv.net/~steve/Cycle%20Time.htm#TOP. You will find the queuing model at the bottom of the article. The reporting method there has an interesting twist of showing backlog status as of when it was created, because more than once I worked with groups where they would encourage customers to cancel outstanding requests and rewrite them, just to meet the "time in system" goals. Also, there was a considerable amount of "backlog" retained in the system with no intent of ever working on it, but the intent of using it to avoid staff cuts and to justify increases. Of course, that was a long time ago, and I am sure no one operates like that now.  Good luck, Steve | ||
| Scott Goldfarb |
Hi Peter, We often call this Maintenance Assignment Scope (MAS), often expressed as FPs per FTE but can be expressed in hours, days etc. It is often stated on an annual basis but can be measured monthly or quarterly. Even when measuring it annually it is more accurate to take the FP count at the end of each month to determine the average application FP size for the year and divide by the effort. We have benchmark data related to maintenance productivity and more detailed definitions. Check out our website at qpmg.com or contact me directly. Regards, scott | ||
| Peter Hudson |
Hi Steve, Scott Thanks for your precious inputs. Steve: you do touch on an important point that I am also confronted with, the amount of backlog. The classic argument goes as follows: Yes i spent more time on "maintaining" my application this year compared to last year, but at the same time I cleared out more backlog then last year. It is indeed quite tricky, I will go through your article to gain more insight. Scott: I will check out your website. My argument against measuring monthly or quarterly is that it is the role of the manager to manage his application correctly against all external risks (changes = risks). Just to tell you where I stand. Thanks for your help. | ||
| Parvathy |
Hi All, Please go through the site below its really usefull to clarify ur doubts. http://it.toolbox.com/blogs/enterprise-solutions/software-maintenance-productivi ty-4527 | ||
| Peter Hudson Username: Peter Post Number: 122 Registered: 02-2004 |
Hi Parvathy thanks for the extra info. |
