Mobile Analytics Reliability Framework

Role: Digital Data Analyst
Scope: Tracking audit, event taxonomy, QA, BigQuery validation, consent monitoring

A mobile analytics reliability project focused on improving tracking quality across iOS and Android.

The project audits Firebase-style mobile events, standardizes the tracking taxonomy, validates critical business events in BigQuery, and adds consent monitoring for analytics collection.

---

TL;DR

• Reduced payment reconciliation gap from 34% to 1.2%
• Reduced Android transaction_id null rate from 12.4% to 0.0%
• Reduced Android claim_submitted duplication from 7.9% to 0.2%
• Improved P0/P1 cross-platform event parity from 63% to 96%
• Added daily validation checks for nulls, duplicates, naming, parity, and event sequence issues
• Added privacy checks to monitor pre-consent event collection

---

How to Read This Repo

Time	Start here
2 minutes	impact/results.md
Technical review	sql/validation_suite.sql and sql/funnel_claim_submission.sql
Tracking review	tracking_implementation/commanders_act_tracking.js
QA review	tracking_implementation/qa_debugging.md
Governance review	taxonomy/parameter_dictionary.md and data_quality_framework/validation_rules.md
Privacy review	privacy_consent/gdpr_compliance.md

---

The Problem

A mobile app running Firebase Analytics and BigQuery had several tracking reliability issues.

The dashboard looked stable, but the underlying event data had gaps.

Main issues found:

• Duplicate claim_submitted events on Android
• Missing transaction_id values on Android payments
• Missing funnel events on Android
• Inconsistent event names between iOS and Android
• Deprecated events still appearing in production
• Custom events firing before consent status was known

These issues affected reporting, CRM activation, funnel analysis, and privacy monitoring.

---

What Was Built

1. Event Audit

Created an event audit log covering:

• Event name
• Platform
• Severity
• Issue type
• Root cause
• Fix version

2. Tracking Taxonomy

Built a standardized taxonomy with:

• Approved event list
• Parameter dictionary
• Required fields
• Enum values
• Deprecated event mapping
• Cross-platform naming rules

3. Measurement Plan

Documented P0 event definitions for critical events such as:

• claim_submitted
• payment_completed

Each definition includes:

• Trigger logic
• Required parameters
• Platform implementation notes
• QA rules
• BigQuery validation notes

4. Tracking Implementation

Reworked client-side tracking logic with:

• Consent gating
• Required parameter validation
• Server-confirmed triggers
• Duplicate prevention
• Consistent payload structure

5. BigQuery Validation Suite

Built daily checks for:

• Null business IDs
• Duplicate conversions
• Naming compliance
• iOS and Android parity
• Event sequence integrity
• Debug events in production

6. Privacy Monitoring

Added consent checks to verify that:

• Analytics storage starts denied by default
• Custom events do not fire before consent
• Consent events are available in BigQuery
• Consent source is populated

---

Results

Area	Before	After
Payment reconciliation gap	34%	1.2%
Monthly revenue gap in analytics	€41 200	€1 480
transaction_id null rate on Android payments	12.4%	0.0%
claim_submitted duplication rate on Android	7.9%	0.2%
user_id null rate on post-auth P0 events	8.3%	0.1%
Cross-platform event parity for P0/P1 events	63%	96%
Event types firing before consent	4	0
Data engineering time spent on investigations	~40% of week	~8% of week
P0 issue resolution time	3–5 days	< 4 hours
Sales reporting escalations linked to data gaps	3 in Q3	0 in Q4

Note

The Android claim conversion rate moved from 71% to 64% after duplicate events were removed. This was a measurement correction, not a product regression.

---

Stack

Area	Tools
Tracking	Firebase Analytics, Commanders Act
Data	BigQuery, dbt
QA	Firebase DebugView, Analytics Debugger for Apps
Monitoring	BigQuery Scheduled Queries
Platforms	iOS, Android
Languages	SQL, JavaScript, YAML

---

Repository Structure

audit/
  event_audit_log.csv

taxonomy/
  approved_event_list.csv
  parameter_dictionary.md

measurement_plan/event_definitions/
  claim_submitted.yaml
  payment_completed.yaml

tracking_implementation/
  commanders_act_tracking.js
  qa_debugging.md

sql/
  funnel_claim_submission.sql
  anomaly_detection_rolling.sql
  validation_suite.sql

data_quality_framework/
  validation_rules.md

privacy_consent/
  gdpr_compliance.md

impact/
  results.md

assets/
  data_quality_before_after.png

---

Key Files

File	Purpose
impact/results.md	Before/after results and business impact
taxonomy/approved_event_list.csv	Approved event taxonomy
taxonomy/parameter_dictionary.md	Parameter rules, types, and enum values
measurement_plan/event_definitions/claim_submitted.yaml	P0 claim conversion event definition
measurement_plan/event_definitions/payment_completed.yaml	P0 payment event definition
tracking_implementation/commanders_act_tracking.js	Consent-gated event tracking logic
tracking_implementation/qa_debugging.md	QA process for real-time and post-release validation
sql/validation_suite.sql	Daily BigQuery data quality checks
data_quality_framework/validation_rules.md	Rule catalog with severity and thresholds
privacy_consent/gdpr_compliance.md	Consent implementation and validation logic

---

What This Project Demonstrates

• Mobile analytics tracking QA
• Firebase Analytics event governance
• BigQuery validation logic
• Funnel and conversion reliability
• Cross-platform event parity monitoring
• Consent-aware tracking design
• Data quality rules for business-critical events
• Clear documentation for Product, Engineering, CRM, and Analytics teams