Solution
Vodafone Czech Republic unified fragmented on-premises data infrastructure by migrating all business performance, network, and campaign data to BigQuery, replacing siloed legacy systems with a single source of truth. Cloud Data Fusion (CDAP) orchestrates data integration, bringing operations in-house and eliminating external vendor dependencies. Data processing now completes by 9 AM, enabling same-day decision-making versus previous day-old reports. The serverless, pay-per-query model reduced operational costs 46% while improving team agility. Teams gained real-time visibility into trading positions, campaign performance, and network metrics, enabling commercial teams to respond to market dynamics within hours instead of days.
Data flow
CRM customer data, billing records, and network traffic metrics extract from source systems via Cloud Data Fusion ETL processes. Data consolidates into BigQuery with consistent schemas. Pre-calculated aggregations complete by 9 AM each morning. Business users query BigQuery directly for daily trend analysis, campaign ROI, and network performance dashboards. Reports and insights feed into decision-making for trading strategies and campaign optimization.
Solution architecture
2 components · 2 layers - Storage
- BigQuery Centralizes all CRM, billing, network traffic, and campaign data into a unified data warehouse with sub-second query performance for business analytics.
- Uncategorized
- Cloud Data Fusion Manages data integration and ETL pipelines, replacing external vendor integrations with in-house control over data flow from CRM, billing, and network sources to BigQuery.
- Automatic multi-region backup of data, replacing single-location on-premises servers
- BigQuery as the single warehouse holding CRM, billing and network traffic data
- CDAP Cloud Data Fusion (open source) for in-house data integration pipelines, deployed with help from Vodafone's central AI and Data team
- ML workloads being prepared for migration onto the same BigQuery foundation
- Per-query cost visibility used as a forcing function to write more efficient pipelines
- Serverless, pay-as-you-go BigQuery model replacing fixed-capacity on-premises infrastructure
46% reduction in operational costs by migrating to BigQuery
Data delivery improved from day-old to same-day by 9AM
Every query we run has a cost associated with it, so we're challenged to be better data engineers
First thing each day, all our data is pre-calculated and ready for users
We're now self-sufficient. We integrated all operations inside our teams with Google Cloud.
