Hundreds of analytics defects were reaching production across Bell’s My Bell, Lucky, and Virgin apps, and inconsistent tagging meant the business was reporting on data it could not fully trust.
“Most analytics QA checks that an event fired. It rarely checks that the source of truth received what it should have.”
Bell owns three native apps, My Bell, Lucky, and Virgin, and analytics tagging was implemented inconsistently between them. Critical documents such as technical specifications, the Business Requirements Document, and the Solution Design Reference were disorganized, and a gap in technical knowledge meant QA was not performed completely, so defects kept rolling into production.
ML arteka took control of the Omniture data governance process end to end. Tagging was standardized by reviewing the Implementation Guide against Bell’s standards and aligning tags with historical data, critical documentation was rebuilt on Confluence, and because Adobe Analytics is the source of truth for other data sources, an extra validation layer was added at QA to confirm Omniture received the data as expected, where before only Charles logs were checked.
The result: consistent tagging across all three apps, significantly fewer defects reaching production, and a streamlined, well-documented analytics pipeline.
Business Outcomes
End-to-end analytics governance replaced app-by-app firefighting with a consistent, documented, validated pipeline.
Aligned to Bell’s standards and historical data.
Down from hundreds rolling into production.
Documented and easy to report on.
The Transformation
From inconsistent, poorly documented tagging to a governed analytics pipeline.
- 1
StandardizeMake the Tagging ConsistentReviewed the Implementation Guide so its authors followed Bell’s standards, and aligned tags with historical data for comparable reporting at the Omniture and BI level. - 2
DocumentRebuild the Documentation BackbonePublished the Solution Design Reference on Confluence so variable mappings and expected values were accessible across the organization. - 3
ValidateCheck the Source of Truth at QAAdded a validation layer to confirm Omniture received the expected data, where before only Charles logs were checked and platform reports were ignored. - 4
StreamlineEasy, Reliable ReportingConsistent tagging and clean documentation made analytics reporting reliable and repeatable across projects.
The Business Challenge
Reporting on Data You Cannot Trust
Three apps, inconsistent tagging, weak documentation, and incomplete QA meant defects kept reaching production.
Inconsistent tagging
Analytics tagging differed across My Bell, Lucky, and Virgin, breaking comparable reporting.
Disorganized documentation
Technical specs, BRD, and SDR were not well organized, though they are the backbone of any analytics implementation.
Incomplete QA
A technical-knowledge gap meant QA was not performed fully against analytics standards.
Defects in production
The result was hundreds of defects rolling into production on unreliable data.
Business Outcomes in Detail
What the Numbers Mean
Aligned to Bell’s standards and historical data.
Down from hundreds rolling into production.
Documented and easy to report on.
Technology Snapshot
Governing the Analytics Pipeline
Internal analytics deserve the same rigour as a customer-facing release. Validating the source of truth at QA, not just the network logs, is what stops defects no one can see.
Executive Questions and Answers
The questions leadership tends to ask when evaluating an approach like this.
Governance
Why govern analytics end to end instead of fixing defects as they appear?
Fixing defects app by app treats symptoms. Standardizing tagging, rebuilding documentation, and validating the source of truth together removes the causes, which is what makes the improvement hold across all three apps.
Data Quality
How do you know the analytics data can be trusted now?
A validation layer at QA confirms Adobe Analytics actually received the expected data, rather than relying only on Charles logs, which is what raises confidence in every downstream report.
Scale
How does this help future projects?
With the Solution Design Reference on Confluence and consistent tagging standards, teams can strategize analytics for new projects in advance from a single trusted reference.
