When two dashboards disagree, the instinct is to debug the data. The cause is usually a definition.
It is one of the most common meetings in any data-mature organization: two teams, two dashboards, one KPI, two numbers. The meeting produces an action to “fix the data” — and three months later, the same meeting happens again.
It recurs because in most cases nothing is broken. Each dashboard is faithfully reporting exactly what it was built to report. The disagreement lives upstream of the data — in language, scope and ownership.
Two dashboards disagree because two teams use the same word for different measures — and nobody owns which one is right.
Five reasons the numbers diverge
Same name, different definition
“Headcount” with and without contractors. “Revenue” recognized versus billed. “Attrition” voluntary versus total. Each team's definition is defensible; the collision only appears when the numbers sit side by side in front of a decision-maker.
Same definition, different scope
One dashboard filters out a company code, a test plant or a probation population — usually for a good historical reason no one wrote down. The filter outlives its author and becomes an invisible fact about the number.
Same scope, different moment
A live view and a month-end snapshot will disagree by design. Unless the timing basis is printed next to the number, the disagreement reads as an error rather than a property.
Same moment, different source
One number comes straight from the system of record; the other has passed through an extract, a warehouse and an enrichment step — each with its own refresh rhythm and its own opinions.
Nobody owns the measure
The root condition beneath the other four: there is no named person whose definition wins. Where measure ownership is real, divergences surface as change requests. Where it is not, they surface as meetings.
Why reconciliation projects fail
The classic response is a reconciliation exercise: trace both pipelines, find every difference, patch them level. It works — briefly. Because the exercise fixed the numbers without fixing the condition, the next new dashboard, filter or extract restarts the drift.
Reconciliation treats the symptom at a point in time. Ownership treats the cause permanently.
What actually fixes it
- One named owner per measure — a person, not a committee, whose definition is the tiebreaker.
- A published definition catalogue — scope, filters, timing basis and source, written where every dashboard builder will find it.
- One declared source of truth per domain — other copies may exist for performance, never for authority.
- Change control on definitions — a definition change is a business event with an effective date, not a quiet edit.
Name the measure→Agree the definition→Declare the source→Publish it once
Within an SAP landscape this is the working boundary between the system of record and the analytics layer — measures defined once and consumed consistently, whether through SAP Analytics Cloud or embedded analytics, rather than re-derived per dashboard.
VISCAP perspective
The dashboard is never the problem. The missing owner is.
When we are asked to reconcile disagreeing reports, the technical trace is the short half of the work. The lasting half is organizational: naming owners, publishing definitions and retiring the unofficial copies. It is unglamorous — and it is the difference between numbers that were made to agree once and numbers that stay in agreement.
The next disagreement is optional
Two dashboards will always be built by two teams with two purposes. Whether they can disagree about the same KPI is a governance choice. Make the choice once, and the recurring meeting disappears from the calendar.