RevOps · Marketing Ops · Data Governance
Reporting is rarely the real problem.
When two dashboards disagree, the most useful question is: where did their definitions, data, or logic stop being the same?
A practical guide to tracing discrepancies, explaining the difference, and building reporting your team can trust.
By Kate Khoury · 7-minute read
Reconcile meaning before reconciling numbers.
A reporting dispute can hold up a budget decision, undermine a forecast, or send a team into another round of spreadsheet cleanup. The dashboard is where the disagreement becomes visible. The cause may sit several steps upstream.
In Revenue Operations and Marketing Operations, familiar labels hide very different questions. A booked meeting is not necessarily a held meeting. An opportunity amount, an invoice, and recognized revenue describe different business events. An activity row is not necessarily a unique lead.
My starting point is to establish which decision the metric supports, then trace the evidence behind it. A broken formula still needs fixing. But a correct formula cannot resolve an undefined metric or an unowned business rule.
The goal is an explainable number, with a definition people share and an owner who can stand behind it.
The diagnostic
Six places to look when reports disagree.
Use the symptom to narrow the investigation. One discrepancy can have more than one cause.
01
Definition
The reports answer different questions: created leads versus qualified leads, or booked meetings versus held meetings.
First check: Write each metric in one sentence.
02
Timing
Refresh schedules, snapshots, date fields, or time zones put the same event into different reporting windows.
First check: Match the cutoff and the business date.
03
Grain
One record becomes several rows after a join. An opportunity with multiple line items can inflate a total.
First check: State what one row represents; test key uniqueness.
04
Scope & data quality
Different exclusions, duplicate entities, missing IDs, or test records change the population being counted.
First check: Compare included record IDs and exclusion rules.
05
Transformation
Filters, mappings, calculations, and overrides change the value between the source and the report.
First check: Find the first step where totals diverge.
06
Ownership & system behavior
Integrations overwrite fields, automations reclassify activity, or teams use conflicting business rules.
First check: Review field history and assign decision rights.
The method
Move from “the dashboard is wrong” to a testable explanation.
1. Define the decision and the metric.
Ask what action depends on the number. Then agree the business event, included population, time window, and counting rule. Choose the authoritative source for each field: CRM might govern opportunity stage, an executed contract might govern commercial terms, and Finance might govern invoiced amounts.
A usable metric definition
- Metric: Held sales meetings.
- Unit: One distinct meeting occurrence ID.
- Include: Held status and actual meeting date within the agreed period and time zone.
- Exclude: Cancellations, no-shows, test records, and superseded reschedules.
- Accountability: A named business owner, source, and refresh cutoff.
Deliverable: A metric definition that both report owners approve.
2. Reproduce the difference and trace the records.
Capture both outputs at a documented cutoff. Record their source, filters, refresh time, identifiers, and what one row represents. Compare totals, then isolate the records present in one output but missing from the other.
Follow the affected records through ingestion, joins, transformations, model relationships, measures, and report filters. Check both row counts and business totals; matching row counts alone can still conceal different records or duplicated amounts.
Deliverable: A reproducible comparison and the first point of divergence.
3. Separate valid differences from defects.
Classify each variance using the six checks above. A timing difference may be expected. Counting a contact once per campaign touch may be correct for activity reporting but wrong for a unique-lead metric. A join that multiplies a contract amount is a defect.
Confirm the explanation against source records and business rules. Keep unresolved cases visible rather than assigning a convenient reason just to close the gap.
Deliverable: Evidence for each explanation, plus a list of open questions.
4. Build a bridge between the totals.
Show how the first result becomes comparable to the second. Apply adjustments in a documented order so the same record is not removed twice. Track the affected IDs, cause, owner, status, and resolution for each exception.
A useful bridge can leave a residual variance. Label it unresolved, identify the decision it affects, and assign the next investigation step. Zero unexplained difference is an outcome to verify, not a number to manufacture.
Deliverable: Starting total, explained adjustments, comparable total, and any unresolved variance.
5. Turn the explanation into a control.
Fix the source rule, transformation, or report label responsible for the confusion. Add a control where the error first becomes possible: required IDs, controlled status values, deduplication checks, integration precedence, or an exception alert.
Give the control an owner and review cadence. Re-run the comparison after the next refresh, and document definition changes so historical results remain interpretable.
Deliverable: A verified correction or documented valid difference, with ongoing ownership.
Worked example
120 exported rows. 80 held meetings. What changed?
An activity export is presented as “meetings booked,” while the management dashboard reports meetings held. Before questioning rep performance, define the two measures and reconcile the underlying records.
Illustrative scenario only. All counts and explanations below are hypothetical.
120
Rows in the meeting export
Includes repeated rows and meetings with different outcomes.
80
Distinct meetings held
Confirmed held within the agreed reporting period.
| Reconciliation step | Adjustment | Remaining |
|---|---|---|
| Starting meeting export | — | 120 |
| Remove repeated rows for the same meeting ID | −15 | 105 |
| Remove cancelled or superseded meeting IDs | −12 | 93 |
| Remove meetings scheduled after the reporting cutoff | −8 | 85 |
| Remove meetings recorded as no-shows | −5 | 80 |
| Distinct meetings confirmed held in the period | — | 80 |
Apply the exclusions in the order shown. Each removed record is counted once; the remaining 80 distinct meeting IDs are confirmed held within the period. Both views use the same captured source snapshot.
The reporting decision: Keep booked and held meetings as separate, clearly defined measures. Correct the misleading label on the 120-row export, govern meeting IDs and outcome statuses, and expose the gap between scheduling activity and completed meetings.
The control: Flag duplicate IDs and missing outcomes before the next report refresh. A difference between booked and held meetings can then support a useful conversation about attendance and follow-through.
For the revenue version of this problem, see When Salesforce and Finance Disagree.
Put it to work
Before the next reporting review.
Start with one disputed metric and the people who own its definition, source, and use. Bring the two outputs and a small set of affected records. Leave the meeting with a documented explanation or a specific investigation owner.
- Decision: What action does this metric support?
- Definition: Do the unit, population, status, and time window match?
- Evidence: Can we reproduce the variance and trace it to records?
- Authority: Which source and person govern the disputed field?
- Resolution: Is this a defect, a valid difference, or still unresolved?
- Control: What will detect or prevent the same ambiguity next time?
The strongest reporting teams can explain how a number was produced, what it means, and who will act on it. That is what turns reporting into decision support.
CRM & Reporting Clarity Audit
Have two reports that should agree?
Bring one reporting problem to my free pilot audit. We’ll focus on the definitions, evidence, and likely causes, then identify what to fix first.
Prefer a self-guided start?
Turn the investigation into an action plan.
The CRM & Reporting Clarity Toolkit provides an editable Excel workbook, a completed fictional example, and a practical guide for reviewing CRM data, aligning KPI definitions, reconciling reports, and assigning the next steps.
Paid digital download. Self-guided files; consulting and implementation are separate.