The old adage, You can only manage what you can measure, applies
to product development. Metrics are needed to understand the level of process
performance, project performance and product performance. They are needed
to set goals and measure the trend and rate of improvement.
Proper metrics need to be selected. Improper metrics can optimize the performance
of a product development sub-process at the expense of global sub-optimization.
Improper metrics can require significant effort to collect data and develop
without providing meaningful information of any real benefit. Criteria for
effective metrics are:
- Keep them simple
- Keep them to a minimum
- Base them on business objectives and the business process - avoid
those that cause dysfunctional behavior
- Keep them practical - avoid metrics that
require significant additional data collection and effort
There are four basic types of metrics for product development:
1. Process metrics - short-term metrics that measure the effectiveness
of the product development process and can be used to predict program and
product performance
- Staffing (hours) vs. plan
- Turnover rate
- Errors per 1,000 lines of code (KSLOC)
2. Program/project metrics - medium-term metrics that measure effectiveness
in executing the development program/project
- Schedule performance
- Program/project cost performance
- Balanced team scorecard
3. Product metrics - medium-term metrics that measure effectiveness
in meeting product objectives - technical performance measures
- Weight
- Range
- Mean time between failure (MTBF)
- Unit
production costs
4. Enterprise metrics - longer term metrics that measure the effectiveness
of the enterprise in undertaking IR&D and developing new products
- Breakeven time
- Percent of revenue from products developed in last 4 years
- Proposal win %
- Development cycle time trend (normalized to program complexity)
See Product Development Metrics List for
other potential metrics
|