Planning Considerations For Loading Sources

In this section:

Working With Ratio and Percent Type Measures

Before creating a new measure, you need to find out:

What are the source data tables and DBMS for the measure?
What fields do you need from the tables to perform this load?
What will be the unit of measure?
What dimension linkages do you need to consider?
What temporary fields (if any) will be created, such as sort fields for measures and joined field definitions for defined dimensions.

Critical fields for measures are:

Target (real field or calculated from real field)
Actual (real field or calculated from real field)
Now date (real field or calculated from real field)
Owner
Dimensional links
Scorecard links
Time (will be linked using the Now date)

Working With Ratio and Percent Type Measures

If you have ever tried to aggregate percentages or other results of statistical methods, you may have noticed that it is not possible to accurately aggregate percent or other ratio type metrics. For example, if you have percentage results for defects per thousand for Charlotte at 20% and Toronto at 10%, you cannot simply add the percentages together to get an aggregate of 30%. To aggregate ratios or percentages correctly, PMF performs the following steps for you:

Separate the numerator and denominator components for each individual measure.
Add all numerators together.
Add all denominators together.
Apply the appropriate calculation method to the aggregated numbers.

Note: Percent and ratio type measures use separate numerator and denominator fields to let you store and correctly aggregate measures.