Data quality determines whether marketing reports can be trusted. A dashboard can be visually clear yet misleading if events are missing, duplicated, misclassified or defined inconsistently. Data validation should be part of measurement operations, not a one-time launch task.
Common issues #
- Missing events: tags fail to load, triggers are wrong or consent settings prevent collection.
- Duplicate events: tags fire from multiple implementations or refreshes repeat a conversion.
- Incorrect parameters: values, currencies, campaign labels or identifiers are absent or malformed.
- Misclassified traffic: inconsistent UTMs, redirects or referrer behaviour distort channel reports.
- Scope mismatch: sessions, users, clicks and conversions are compared as if they were interchangeable.
- Timing differences: time zones, processing delays and attribution windows create apparent gaps.
- Filters and consent: internal traffic, privacy choices and platform modelling affect comparability.
- Broken joins: CRM, ad and analytics datasets use different identifiers or date logic.
Validation routine #
- Maintain a measurement plan and event dictionary.
- Test critical journeys after each release and tag change.
- Compare analytics totals with backend orders, form records or CRM outcomes.
- Inspect sudden changes by event, source, device and landing page.
- Check tag firing, parameter values, filters, consent and data refresh status.
- Record the defect, impact, owner, fix and validation evidence.
Investigating a discrepancy #
First confirm that the systems use the same date range, time zone, conversion definition and counting method. Then compare counts at each stage: source event, analytics event, attributed conversion and confirmed business outcome. This helps isolate whether the gap originates in collection, processing, attribution or downstream reconciliation.
Governance #
Assign owners for tags, metrics, dashboards and source integrations. Version measurement changes, limit access to sensitive data and document known limitations. Never silently “correct” a report by changing definitions mid-period; annotate the impact and preserve an audit trail.
Related: Events, Parameters and Conversions, UTM Parameters and Dashboards and Reports.
