The challenge

A hospital network operating twelve patient-facing digital properties — appointment booking, patient portal, health information, and specialist referral portals — had no central analytics governance. Each property had been implemented by a different agency or internal team over eight years, using different eVar schemas, different event definitions, and in some cases different analytics platforms entirely. The marketing team could not report on digital appointment bookings accurately because the same action was tracked as three different events across different properties. Privacy compliance was an additional concern — nobody had a complete picture of what patient data was being collected and where it was going.

Implementation audit across all twelve properties

We conducted a full analytics audit across all twelve properties — reviewing every active tag, every data element, every event, and every eVar assignment. The audit was documented in a central implementation register that mapped current state against what each property was intended to measure.

The audit findings were significant: three properties were sending data to report suites that no longer existed in the analytics account. Two properties had tracking pixels from agencies whose contracts had been terminated. One property was capturing form field values — including appointment request details — in eVars, which was a privacy compliance issue that needed immediate remediation.

Governance framework design

We worked with the analytics, IT, and compliance teams to define a governance framework covering four areas: ownership (who is responsible for each property's implementation), standards (what the implementation must do and how), change management (how changes are reviewed and approved), and compliance (what data must not be collected).

The measurement standards document defined a canonical eVar and event schema that all twelve properties would adopt — with shared definitions for appointment booking, patient portal login, content engagement, and referral submission. Properties could extend the schema with property-specific variables in designated reserved ranges, but the core schema was non-negotiable.

  • Central variable registry — owner, definition, valid values for every eVar and event
  • Data layer standard — required objects and properties for every page type
  • Change management — no implementation changes without analytics team sign-off
  • Privacy checklist — mandatory review before any new data element goes live
  • Quarterly audit cycle — implementation reviewed against standard every quarter

Phased implementation and migration

Migrating twelve properties to the new standard simultaneously was not feasible. We prioritised by business impact — the appointment booking portal first, then the patient portal, then health information properties. Each property was migrated using a parallel tracking approach to validate the new implementation before decommissioning the legacy setup.

The privacy remediation — removing the form field data from eVars — was treated as an emergency track and completed within the first two weeks across all affected properties, independent of the broader migration timeline.

Outcome

Within six months, all twelve properties were operating under the unified governance framework with the canonical measurement schema. The appointment booking event — previously reported three different ways across different properties — now produces a single, consistent figure that marketing, operations, and finance all agree on. The privacy remediation eliminated the compliance risk before it became a reportable incident. The quarterly audit cycle is now embedded in the analytics team's operating rhythm.