Understanding of Metrics – Project Managers watch in a Testing Effort
In test development projects, software testing effort yields deliverables like test plans, test design, test development and test execution.
A good project manager is the one who applies the principles of metrics to plan, organize & control his/her project deliverables in quantifiable / measurable terms.
Here we are talking about two significant terms i.e. 1) Control & 2) Measurement
Control: The test manager is able to foresee the deviations from the plan, so that he/she gets sufficient breathing time to react & get lesser & lesser surprises.
Measurement: It refers to documenting past effects to predict the futuristic effects in
In the famous book titled “Controlling Software Projects” – Tom DeMarco gives following two wonderful quotes.
Quote -1: “You can�t control what you can�t measure.”
Quote -2: “The only unforgivable failure is the failure to learn from past failure.”
Firstly let us see � what is a metric?
A metric is a measurable indication of some quantitative aspect of the system. We have following three types of metrics.
1) Predictor Metric: It provides forewarning having a definite correlation with some result that we are likely to get in future.
2) Result Metric: It provides measurement of some event or process that has already been completed.
3) Derived Metric: It is derived out of some calculations or from graphical or analytical technique & may have involvement of more metrics.
Secondly let us see different characteristics of Metrics
1) It is measurable: By definition – a metric has to be measurable for being called a metric. If a process cannot be measured, there is no way to control it by the management tools.
2) It is Independent: Metrics is supposed to be independent of all types of human influence. Man should not be able to change the measurement except changing the activity that created the metric.
3) It is Accountable: The metric data itself is responsible for all types of analytical interpretations. Hence it is essential to carefully preserve every raw metric data and detail of audit of all analytical processes.
4) It is Precise: Precision refers to accuracy. The important aspect of precision is that a metric is especially documented as a part of the data collection process. In case a metric changes, it can be measured as a range or tolerance.
Next question comes – Why Test Metrics are collected?
Biggest reason is to make the testing process more effective. This is achieved by carefully analyzing the metric data and taking the appropriate action to correct problems. The process begins by defining the desired objectives of the metric. Few examples are:
a) Defect analysis: Every defect needs to be analyzed to provide answer to questions like the root causes, how detected, when detected, who detected, etc.
b) Test effectiveness: How much is the effectiveness of testing? e.g. How much is the ROI – Return on Investment?
c) Development effectiveness: How much is the effectiveness of defects fixing process?
d) Test automation: How much effort is needed on test automation?
e) Test cost: Quantum of resources consumed and time spent on testing?
f) Test status: This refers to status tracking, or identifying the current situation during the testing process?
g) User involvement: Understanding the level of user involvement in the testing?