This chapter is from the book
This chapter is from the book
This chapter is from the book
Step 2: Defining Measures
After defining the goals and initiatives, the next step is to define supporting measures. Each goal or initiative should be addressed individually to determine the measure or measures pertaining to that goal. One key question is,"What measure will show the status or progress of a particular goal or initiative?" It is appropriate to gain input from the groups that identified the goals, and the measurement definition is often completed in the initial workshop.
The facilitator might need to provide examples of common measures to aid the group. Lists of common measures are available in the International Function Point Users Group (IFPUG) manual, Guidelines to Software Measurement. In addition, various consulting companies provide training, onsite workshops, and resource material in this area.
Examples of measures for the previously defined goals are listed in Table 6-1. After the goals and measures have been identified, it is a good time to prioritize the list. You want to provide enough information to the organization to be helpful, but avoid overwhelming people with too much information. One measure may support multiple goals, enabling more goals to be addressed. In selecting which goals and measures to initially implement, consider the following:
Table 6-1
|
Goal/Initiative |
Measures |
|
Improve project productivity (Goal) |
Function points per hour |
|
Improve project quality (Goal) |
Delivered defects per function point |
|
Reduce project cost (Goal) |
Cost per function point |
|
Implement formal Defect inspections (Initiative) |
emoval efficiency rate (defects found prior to implementation divided by total defects found) Delivered defects per function point |
Measures for various groups: Because multiple groups participated in the definition process, selecting at least one goal and measure from each group may be appropriate. This helps people see that their input has made a difference.
Top-priority measures: Activities critical at the time of definition may dictate what measures to choose (for example, decisions being made about out-sourcing, low customer satisfaction with quality).
Quick payback: If an organization is concerned about acceptance of the measurement program, a helpful approach might be to select measures that can provide information and benefits quickly (for example, reduction of defects per function point due to formal inspections versus developing historical repository to improve estimating).