From Legacy Data Pipelines to a Modern Snowflake Platform at Philip Morris
CPG
Enterprise
Manufacturing

From Legacy Data Pipelines to a Modern Snowflake Platform at Philip Morris

Technology used

Snowflake
Matillion ETL
dbt Cloud
AWS S3
SFTP Ingestion Pipelines
Power BI

Solution

Delivered a full migration from legacy ingestion and reporting systems to align with a global modern Snowflake-based Data Platform as a Service (DPaaS). The project rebuilt ingestion pipelines, simplified data lineage, and enabled seamless data sharing, all without disrupting existing reporting.

Results

PMI France gained full ownership of its data platform, eliminated reliance on global support teams, and transitioned to a scalable architecture aligned with global standards. The migration was completed with zero downtime, improved performance, and stronger foundations for future analytics and AI initiatives.

The Challenge
The Solution
The Results

The Challenge

PMI France faced increasing difficulty managing a complex and fragmented data landscape that limited both agility and performance. At the same time, the organisation was preparing to decommission key legacy components that supported ingestion and reporting. Together, these factors created both a technical and business-critical need to modernise the platform quickly, safely, and without impacting ongoing operations.

‍

01
Limited Local Autonomy: All data resided within PMI's global Snowflake environment ("Data Ocean"), which restricted local autonomy. Even simple changes required raising tickets with global support teams, often resulting in long delays or unresolved requests. This dependence reduced agility and made it difficult for local teams to respond quickly to changing business requirements.
02
Increasing Platform Complexity: Over time, the platform had evolved into a highly complex structure. Deeply layered data lineage made querying slow and debugging difficult, while redundant views and tables reduced efficiency. Complex transformation logic and inconsistent naming conventions made the platform increasingly difficult to understand, maintain, and extend.
03
Performance and Operational Risk: The growing complexity of the platform directly impacted performance, with Power BI reporting affected by slow query execution. At the same time, PMI was preparing to decommission its Enterprise Landing platform, which powered ingestion pipelines and reporting. With key components approaching end-of-support, the risk of operational disruption increased significantly, creating urgency to move to a modern, future-ready architecture.
By migrating to the new platform, the team can now take full ownership of their data, enabling them to make changes, build new pipelines, and resolve issues without relying on global support.
Marwa Kouriat
Marwa Kouriat
Manager, IT Data Platforms
France

The Solution

Snap Analytics delivered an end‑to‑end migration to PMI France’s new Snowflake DPaaS Geo Node, combining modern data engineering practices with close client collaboration.

‍

Discovery and Planning
A rapid discovery phase validated system dependencies and migration scope, allowing the team to move quickly into delivery. Tight deadlines, driven by the Enterprise Landing decommission, required efficient prioritisation and clear technical alignment.
Rebuilding the Data Foundation
The team rebuilt ingestion pipelines using Matillion ETL, extracting data from SFTP sources, staging it in AWS S3, and loading it into Snowflake. DPaaS ingestion frameworks were used to ensure global consistency and future reusability. Transformation logic was rebuilt in dbt Cloud, introducing a standardised and governed transformation layer across the platform. dbt enabled the implementation of modular data models, automated documentation, and testing frameworks that significantly improved data quality, traceability, and consistency within the Snowflake data warehouse. This modern transformation approach reduced technical debt, accelerated change delivery, and increased confidence in reporting outputs.
Seamless Migration with Zero Downtime
Using Snowflake’s data sharing capabilities, Snap enabled the new platform to feed existing environments. This allowed dbt models to be repointed behind the scenes while Power BI reports continued running uninterrupted in Data Ocean. As a result, the migration was effectively invisible to business users.
Simpler Lineage
The team streamlined data models, standardised naming conventions, and leveraged dbt Cloud to introduce a structured and governed transformation layer. This improved lineage visibility, documentation, and consistency across business logic, reducing debugging effort and increasing trust in reporting outputs.
Collaborating and Upskilling
Snap worked as an integrated team with PMI France, collaborating closely with internal stakeholders on a day-to-day basis throughout the project. This approach enabled faster issue resolution through real-time collaboration, while also supporting knowledge transfer and hands-on training in Matillion. By working side by side with the client team, Snap helped build the confidence, skills and independence needed to manage and extend the platform long term. A key outcome was empowering PMI France to take full ownership of their data pipelines and continue evolving them independently.

The Results

Zero Business Disruption
The migration was completed with no impact to Power BI reporting or day-to-day business operations, ensuring a seamless transition to the new platform.
Full Data Ownership
PMI France can now make changes, build new pipelines and resolve issues independently, removing reliance on global support teams.
Greater Agility and Faster Delivery
Simplified architecture and streamlined processes enable new tables, columns and pipelines to be created more quickly, helping teams respond faster to changing business needs.
Improved Performance and Simpler Operations
Reduced query complexity and clearer data lineage have improved platform performance while making troubleshooting and debugging faster and more efficient.
Reduced Risk and Technical Debt
The successful retirement of Enterprise Landing eliminated platform risk and removed legacy components that were becoming increasingly difficult to support.
More Trusted and Transparent Data
Standardised transformation logic, automated validation, improved documentation and clearer lineage have increased confidence in analytical datasets and reporting outputs.
Future-Ready Analytics Foundation
A modern Snowflake-based architecture now provides a scalable foundation for advanced analytics, self-service reporting and future AI initiatives.

Philip Morris International (PMI) is one of the world’s leading tobacco and nicotine companies, operating across more than 180 markets and employing over 79,000 people globally. The organisation is focused on transforming towards a smoke-free future, investing in innovation and advanced data capabilities to support consistent, compliant, and insight-driven operations at scale.

‍

Next case study

West Burton Energy Accelerates SAP Analytics with a Cloud-Based Data Platform

Caffè Nero Builds a Scalable Data Platform for Real-Time Decision Making

Vivo Energy Transforms SAP Data with a Modern Databricks-Powered Platform