ACOSM 2010: KISIS - Keeping it Simple is Stupid

KISIS - Keeping it Simple is Stupid (How Trying to Keep it Too Simple can Lead to Difficulty)

John Ogilvie, IBM Global Business Services

Abstract

Based on many years of involvement across many outsourcing contracts using Function Points, this presentation describes the pitfalls that can be experienced if one takes too simplistic a view of the meaning and use of function point data and suggests ways in which they may be avoided.

Starting with a typical outsourcing scenario, and using ISBSG project data, techniques to improve the effectiveness of a Function Point program are demonstrated. Particular emphasis is made on the importance of setting baselines appropriate to the environment to be measured.

The use of statistical analysis to enable a more sophisticated and pragmatic interpretation of measurement data is recommended. The view that a little statistical analysis can actually uncover “lies and damn lies” is offered.

Finally, a template for design of a successful Function Point Analysis Program is presented.

Key Points

• Simplicity can cause a false sense of security
• A broader understanding, by all stakeholders, of what measurement data means is vital
• Statistical analysis can actually uncover lies and damn lies
• Factors to be considered in establishing performance baselines
• How to design a Function Point Analysis measurement regime that can work.

Biography

John Ogilvie is Function Point and Productivity Consultant with IBM Global Business Services.