- Introduction
- Step 1: Defining Goals and Initiatives
- Step 2: Defining Measures
- Step 3: Defining Data Collection
- Step 4: Defining Reports
- Step 5: Implementing the Measurement Process
- Summary
Step 5: Implementing the Measurement Process
This section describes implementation approaches in the areas of roll-out, staffing needs, process and methods development, and education.
The implementation of the measurement process requires planning and development, as well as decision-making in the areas of roll-out and staffing. The approaches taken will depend on the staffing levels and structure of the organization. Alternatives are presented for the implementation of roll-out and staffing.
Roll-out
The defined measurement process can be rolled out to the organization in various ways:
Across the entire organization at once
On selected projects or applications across several areas or departments
In selected areas or departments in phases until the process is incorporated throughout the organization
Tables 6-3 through 6-5 show pros and cons to each approach.
Table 6-3: Across Entire Organization
PROS |
CONS |
Everyone is involved right away. |
Consistency is difficult to control. |
All the training is completed at once. |
Any necessary adjustments involve everyone. |
Data is obtained on the entire organization's portfolio. |
Benefits may not be quickly visible with high volumes of data collection slowing analysis, management, and reporting. |
Positive outcomes are seen throughout the organization. |
Any negative impacts are seen throughout the organization. |
Table 6-4: Selected Projects or Applications Across Several Areas
PROS |
CONS |
Amount of data is initially manageable. |
Implementation is difficult as a repeatable process because it is seen as a one-time initiative. |
Initial impact on staff is minimal. |
Limited data across several areas makes trends difficult to see. |
Small focus allows for dedicated attention. |
Staff is trained on processes they may not use day-to-day. |
The decision on how to staff the measurement program may impact the roll-out approach and visa versa, so roll-out and staffing decisions should be considered concurrently.
Staffing Needs
The developers need to be responsible for some of the data being collected (for example, effort, cost, and defects). Developers will also need to have methods for recording project data accurately and consistently. However, the overall measurement activities, such as function point counting and data analysis, might require more time than the developers have.
Table 6-5: Phased Implementation by Selected Areas or Departments
PROS |
CONS |
Can train staff that will be utilizing the skills on a regular basis. |
Not all areas are initially involved. |
Can collect enough data points to see trends to provide opportunities for process improvement. |
Highlighting chosen departments may cause concern for some staff. |
Small focus allows process to be adjusted as necessary with small impact on staff. |
Time is longer before data becomes available on entire organization. |
Staff can see the entire measurement process from impacts to benefits. |
|
Consistency of data collection is greater. |
|
The following staffing alternatives can be used for counting function points, maintaining the data repository, and completing the analyses and reporting activities.
- Assign a resource within each department.
- Utilize external resources.
- Establish a central metrics group.
Tables 6-6 through 6-8 show the pros and cons for each approach. Depending on the resources within the organization, different companies select different options. Combinations of these options may also be necessary. For example, it is imperative that measurement knowledge and expertise be a part of the initial setup and implementation of the measurement program. This may require external resources up front to educate staff, but after the knowledge has been transferred, an internal metrics group can be established.
After the staffing decisions have been made, the detail process and methods and educational materials can be developed and presented.
Process and Methods Development
Establishing the measurement process and methods is a required activity regardless of the roll-out and staffing decisions.
Table 6-6: Resource Within Each Department
PROS |
CONS |
System expert and function point expert may be the same person. |
Reviewing processes to ensure consistency is more involved. |
Ownership of the data is stronger. |
The selected individual may become overwhelmed with FP counts, data analysis, and reporting. |
Scheduling function point counting sessions may be more timely. |
Having resources focused on their individual areas may make data analysis regarding organizational trends more difficult. |
|
Staff turnover would require identification of replacements. |
Table 6-7: External Resources
PROS |
CONS |
No learning curve for expert data analysis and function point counting. |
Additional resource costs are required |
Staff turnover would not be an issue. |
Scheduling of function point counts may not be timely. |
Accuracy and consistency are strong from the start. |
Knowledge transfer to the organization has to be defined as part of the measurement program. |
Prior to implementing the measurement program, the process and methods for data collection and reporting must be defined and documented. This involves integrating into the development life cycle the specific measurement activities that must occur, when they should occur, and how they should occur. All forms, tools, and report formats should be defined to ensure the process is implemented consistently, accurately, and effectively for all those involved. This can be a time-consuming and not always exciting activity for the planners, but a necessary one for implementing a standard, integrated, and repeatable process.
Table 6-8: Central Metrics Group
PROS |
CONS |
Consistency and accuracy is developed more quickly with a focused metrics team. |
Measurement and function point knowledge is centralized and ownership of the process is limited unless appropriate education is undertaken. |
Analysis and reporting throughout entire organization is more easily completed with a central data collection point. |
Scheduling of function point counts in a timely manner requires good communication. |
Development staff is impacted for system or project expertise only during the counts and not in reviews and documentation. |
Without appropriate education, the data analysis and reporting process may not appear owned by the development staff. |
Education
Appropriate education is necessary for all those involved in the process, based on their assigned activities or use of the measurement information. Table 6-9 shows examples of the types of education and training necessary for various audiences.
Table 6-9: Education and Training
Audience |
Sample Training |
Metrics Group |
Function Point Training |
|
Analysis and Reporting Education |
|
Software Project Estimating |
|
Measurement Methods |
|
Function Point Repository and Data Analysis Tools |
Developers |
Education on the processes and developer involvement |
|
Function Point Training |
Managers |
Education on the analysis and use of the data |