What information can be determined from the variance report that may accompany a Workforce Analytics data refresh?
SAP SuccessFactors Workforce Analytics
Reproducing the Issue
You may receive a report like this as part of your monthly refresh process, to assist you in approving the data on Preview before it is published to Production.
The variance report has a number of tabs, and the later tabs calculate based on the first few. You may wish to have one of your reports open as you read further.
- New: All the base measures (i.e. no derived or calculated measures) from the newest WFA cube, calculated for each month over the last several years (or however many periods are available). This may include restricted inputs and other measures which are not displayed on the site at all (e.g. Months So Far).
- Old: Same as the New tab, but taken from the previous cube. This may be from the last month's refresh or from a mid-cycle development cube. The timestamp in the cube name at the top (e.g. InstanceName201702020458) will tell you when each cube was processed.
- Difference: New minus Old = the mathematical difference between the two cubes. If a cell is 0.00, the value has not changed in between. For the latest month, which does not exist in Old, this is simply the value from New.
- Percentage Difference: Difference divided by Old = the percentage change in the value. If the value has decreased the percentage will be negative. As this is calculated automatically, some cells with no difference may show a blank or "Infinity" instead of 0.00.
- Percentage Growth: based on the New tab, the percentage change each month based on the previous month.
- EOP HC VarianceAccountable test: row Difference = Start of Period Headcount plus Hires minus Terminations minus End of Period Headcount. Checks that the number of people over time is changing in a logical way, e.g. that the right number of hires and terminations is reflected in the headcount.
Here are some guidelines for using the variance report.
- Non-zero values in the New tab at least indicate that the new cube contains data. You would expect this to be the case.
- If a cell reads "-100%" in Percentage Difference, there is no data in the new cube and it may not have been processed correctly, or data was missing from the extract.
- Some minor variances in several measures between cubes are common in the most recent few months, because this is where the source data is most likely to have been legitimately amended. Large differences in reporting periods several years ago are more likely to indicate a problem, or else the effects of any logic changes you have requested.
- The VarianceAccountable test may not come to exactly zero in every case. For example an employee may change Employment Status or org unit and become excluded from counts after a certain date, so the headcount decreases without a corresponding termination.
- If a calculated measure is critical for your approval, check the Input View for it on the site to determine which base measures are used to calculate it, then check these measures in the report.
- If you have metrics packs that are calculated in a separate cube, e.g. Survey data, it may not appear in this report or come in a separate report according to your arrangement with SuccessFactors.
WFA, refresh, variance, report, Workforce Analytics , KBA , LOD-SF-WFA-WA , Workforce Anaytics , How To