I have often seen organizations struggle trying to identify KPI's in the Analysis and design phase of Business Intelligence projects I have worked on.
Often, the Project Champion or Executive Sponsor has lofty goals for the project; however getting down to brass tacks becomes problematic
This is usually compounded by the desire to focus on the "sexy" metrics with no regard to the basic building blocks (Descriptive Statistics) that enable tracking these metrics.
The Goal Question Metric approach or GCM can be applied to help organizations translate business needs into actionable requirements which ultimately become KPI's inside a Business Intelligence Solution.
Here is a simplified outline of how to use this approach:
Step 1: Define the Goal
The Goal should be written in plain meaningful English and have the SMART attributes
Example: A college wants to increase student Completion Rates
Step 2: Define the Questions
What Information do we need to measure our progress to meeting this goal
What is the current completion rate for this Term?
What is the Trend? (increase or decrease for this year to last year)
Step 3: Define the metric
The Metric definition should include a complete description of what is included and excluded from the measurement.
You must also assessed the feasibility of actually obtaining the measure.
Completion Rate: The ratio of students who begin a term and are still enrolled at the end of the term irrespective of grade
only those students present at both snapshot are considered.
As you can see, applying this method often leads to a more complete analysis of the business domain at hand.
It also helps discover related area's of interest.
GCM Web Resources:
GCM Entry on Wikipedia http://en.wikipedia.org/wiki/GQM