You are on page 1of 2

Question: What kind of metrics are typically collected by software development and

maintenance organizations?
The typical primary metrics collected by software organizations measure the organizations /
projects for:
Accuracy of planned/ committed schedule
Accuracy of estimated effort
Accuracy of cost estimates.
Productivity
Cycle Time.
Quality of Delivered Software
Customer Satisfaction.
Rework
The idea is that a software organization is successful if it estimates (schedule, effort and cost)
accurately, and achieves high productivity, low cycle time and high quality and need to
perform the minimal rework. In addition to the above metrics, there would be secondary
metrics that contribute to the above metrics. For example:
Metrics related to skill levels
Metrics related to extent of reviews
Metrics related to coverage of testing
Metrics related defects found during reviews and testing
The secondary metrics typically help the organization understand their processes in depth and
prioritize the areas for improvement.
What would be a quality metric for a maintenance project?
In a maintenance project we do not write much code, we just modify existing (often badly
structured) code. The percentage of bad fixes or bad deliveries can be used as a metric for
quality of the maintenance project.
Question: What are the most important things to track in a project?
Firstly, you have to ensure that the project has a plan sufficiently comprehensive and well
defined to track effectively against. Tracking needs to be done for all aspects of the plan, such
as
Actual activity start/ end dates need to be tracked against schedule to identify
slippage, specially critical slippage that require corrective action
Costs/ effort expended as against budget. This includes checking actual effort against
estimated for activities to see whether estimation for other activities needs to be
revisited because of some wrong productivity assumptions
Quality aspects whether products made are of acceptable quality done using
patterns of defects found in reviews and testing, extent of coverage executed in
reviews and tests, etc.
Scope changes scope creep is a major problem in most software projects and needs
to
be
guarded
against
The tracking is done using inputs such as timesheets, weekly/ monthly reports, review
logs, test logs, costing analysis, etc.

Other Set of Measurements for each level in CMM:


Level 2

Testingcost
Debuggingcost
CostindevelopingtestplansTestschedule
Numbersoftestcasesderivedfromtesttechniques:blackbox,whitebox

Level 3

NumbersofindependenttestersCostoftestorganization
Numbersoftestrelatedcoursesofferedperyear
Numbersofdefectsfoundbyphase
Scheduleslippage
Numbersoftestcasespassed
Numbersoftestcasesfailed
Numbersoftestedmodules
%statementcoverage
%branchcoverage
%basispathcoverage
Numbersofintegratedandtestedmodules

Level 4

Numbersofdefectsremovedbyreviews
Numbersofdefectsfoundedineachseverity
Numbersofdefectsfoundineachdefecttype
Testers'productivity
Correctness
Efficiency
Flexibility
Interoperability
Maintainability
Portability
Reliability
Reusability
Testability
Usability

Level 5

Totaleffortinfindingrootcausesofdefects
LCLandUCLfordefectdensity,revieweffortrate,testeffortrate,%statementcoverage,%branch
coverage,and%basispathcoverage
Totaleffortinevaluatingnewtestingtoolsandtechnologies
Totaleffortindefiningwhentostoptesting
Failurerategoal
Actualfailurerate

You might also like