Single Customer Views for British American Tobacco on Snowflake and Microsoft Fabric
MAF Group brought data from around 60 British American Tobacco markets, each with seven to eight data sources, into one model, first on Snowflake and then on a Microsoft Fabric platform built from scratch, and built single customer views on top.
About the organisation
British American Tobacco is a FTSE 100 consumer goods group operating in markets around the world. MAF Group worked with BAT from 2022 to 2026 across data platforms, analytics, marketing technology and AI.
The problem to solve
Customer and commercial data sat across around 60 markets, each with seven to eight data sources including e-commerce platforms, CRM and digital analytics tools, with no common model. Without one, there was no single view of a customer and no reliable way to reconcile figures between systems.
- !Around 60 markets, each with seven to eight data sources
- !E-commerce, CRM and digital analytics data held in separate platforms
- !No shared data model, definitions or documented joins
- !No single view of the customer
What we built
MAF Group started on BAT's existing Snowflake platform with Salesforce data, then designed and built the Microsoft Fabric platform from scratch, and acted as the subject matter expert for the data throughout: the modelling, the data dictionaries and how the sources join.
- Initial work on BAT's Snowflake platform, starting with Salesforce
- Microsoft Fabric platform designed and built from scratch
- Data brought in from around 60 markets, seven to eight sources each
- Sources including Magento, Shopify, Salesforce and digital analytics platforms
- Data modelling and joins across the sources
- Data dictionaries documenting the model
- Data reconciliation across systems
- Single customer views built on the unified model
Technology stack
Our engagement
MAF Group worked on the Snowflake platform, built the Microsoft Fabric platform from scratch, built the data models, wrote the data dictionaries and held the subject matter expert role for the data, working with BAT teams from 2022 to 2026.
