Beyond conflicting dashboards: how HRIS integration creates a single source of truth

Dashboards conflict when each tool holds its own copy of people data; integrating every tool with the HRIS as the system of record, using shared definitions, makes them all read the same data and report the same numbers.

When HR, finance and operations bring different headcounts to the same meeting, the argument is rarely about the business. It is about which copy of the data is right. Integration fixes the plumbing; agreed definitions fix the meaning.

A team leans over printed charts and graphs on a meeting table
Photo: Yan Krukau on Pexels

Why dashboards disagree

  • Different snapshots. One tool was refreshed last week, another last month.
  • Different definitions. One counts contractors, another does not; one measures turnover monthly, another annually.
  • Manual edits in a downstream copy that never reached the HRIS.
  • Different populations, such as one system excluding employees on leave.

Establish the system of record

Decide which system owns each field. The HRIS usually owns employment data: hire date, job, department, manager and status. Payroll owns pay. Downstream tools read from the owner and do not let users edit owned fields locally, so corrections flow back to the source.

Agree definitions before connecting

Definitions to agree and write down
MeasureDecide
HeadcountWho is included: full-time, part-time, contractors, leave
Turnover rateLeavers over average headcount, for which period; voluntary and involuntary separately
Span of controlDirect reports only, or including dotted lines
Open rolesApproved, posted or both

Federal statistics, such as the Bureau of Labor Statistics' JOLTS definitions of separations and quits, are a useful reference for consistent turnover terms.

Integration in practice

  1. Use the HRIS's API or a scheduled feed, not manual exports.
  2. Request only the fields needed, which also limits privacy exposure.
  3. Lock synced fields downstream so they are corrected at the source.
  4. Log every sync with what was skipped and why.
  5. Reconcile once at go-live and investigate any differences.

Organization Viewer keeps the org chart in sync from an HR system (BambooHR, HiBob, Personio, Factorial or Deel), from a CSV feed, or from Google Workspace or Microsoft 365, reading only named fields such as name, title, department and manager, with read-only directory access. Synced fields become read-only in Organization Viewer, so corrections are made in the source. For selecting or reviewing HR systems, see the Vendor Due Diligence project.

An example of a definition fix

A common case: HR reports headcount including employees on long-term leave, while finance excludes them because they are not on payroll that month. Both are correct for their purpose. The fix is not to pick a winner but to define both measures, name them differently, for example "employees" and "active payroll headcount," and label every report with the one it uses.

Integration go-live checklist

  1. Confirm field ownership and the definitions each report uses.
  2. Grant the integration the narrowest access the source system allows.
  3. Run the first sync and compare counts with the source.
  4. Review skipped records and fix them at the source.
  5. Switch reports to the integrated data and retire manual extracts.
  6. Schedule a monthly check that counts still match.

Key takeaways

  • Conflicting dashboards come from copies and from different definitions.
  • Make the HRIS the system of record for employment data and lock synced fields downstream.
  • Write down and label the definition each report uses.
  • Integrate through APIs or scheduled feeds with the narrowest access possible.

Questions

What is a single source of truth for HR data?

An arrangement where every tool reads people data from the system of record, using shared definitions, so reports agree.

Which system should be the source of truth?

Usually the HRIS for employment data and payroll for pay. Decide owners field by field.

Why do reports still disagree after integration?

Usually because of different definitions or populations. Agree and document them.

Should downstream tools allow edits?

Not to fields owned by the source system; corrections should be made at the source.

How do we handle contractors who are not in the HRIS?

Decide whether they belong in the measure, and if so, bring them in through a defined feed with their own type, so they can be included or excluded deliberately.

Do we need a data warehouse?

Not always. Smaller organizations can integrate tools directly with the HRIS; a warehouse helps when many sources must be combined.

How do we know the integration is working?

Compare headcount and key counts with the source system after each sync for the first few weeks, then monthly. Review the sync log for skipped records and investigate any sudden change in counts that has no known business cause.

What if two systems both claim to own a field?

Choose one owner per field and make the other read it. Where both are needed for different purposes, create two clearly named fields rather than letting them compete.

How long does an HRIS integration take?

The technical connection can be quick with a standard connector; agreeing definitions and cleaning data usually take longer and should be planned for.

Does integration replace data governance?

No. Integration moves data reliably; governance decides what it means and keeps it clean.

Sources

← All data infrastructure and governance