Why drill down?
Boards and stories show calculations. When a number looks wrong or surprising—for example, an age band of 0–20, or a headcount that does not match your expectation—drill down shows who is behind the number and which raw records produced it.
Aragorn’s model is: Analytics runs on Ontology data. If the visual is wrong, inspect the people in the visual, then inspect Ontology / source systems if the roster itself is wrong.
Open drill down
Select the segment, KPI, or chart region you want to inspect (for example, the 0–20 age band on average salary by age band).
Open drill down (the control shown on the visual in your environment).
You see the list of people (records) included in that calculation, with a count of how many fall in the selection.
Inspect and enrich the list
Review who appears in the selection.
Bring in additional columns when needed (for example, date of birth) to verify band logic.
Page through results if the list spans multiple pages.
Example: For a 0–20 age band, dates of birth in 2006–2009 can confirm ages in the late teens to 20—not literal newborns—so the band label matches the data.
Validate the math yourself
From drill down, download the people in the selection when you want an independent check.
Recompute the calculation offline (for example, average the salaries and divide by the count).
Compare to the story value (for example, average salary shown as 39.22k).
This is how admins confirm Aragorn is showing the correct analytics for the loaded data.
Distinct people vs row counts
Drill-down row counts can be higher than a distinct headcount KPI if the same employee appears more than once in the underlying grain (for example, multiple positions).
Example: A board KPI shows 2,734 distinct employees; drill down lists 2,737 rows because some people repeat. A distinct count on employee ID aligns with 2,734.
When the business question is “how many human beings?”, prefer measures that count distinct employee IDs.
If the roster itself is wrong
Note the unexpected people or missing people in drill down.
Open the matching Ontology entity (for example, Employees) and inspect the raw loaded data.
Trace back to the integration or HRIS load that feeds Ontology.