The Value of the Lakehouse: How T-Mobile Articulated the Benefit of a Modern Data Platform

Summary

The T-Mobile team discusses their data strategy, focusing on updating their data platform to gain non-technical leadership support. They articulate their strategy through four pillars: data latency, mobility, performance, and economies of scale. The team migrated from Azure Data Warehouse to a lake house architecture, addressing concerns about time, cost, impact, and security. The migration resulted in reduced latency, increased data mobility, predictable performance, and significant cost savings, setting a foundation for future data strategy simplification.

Takeaways

• 📝 The presentation discusses the data strategy and lake house journey of T-Mobile's procurement and supply chain division, highlighting the importance of aligning data strategy with business needs.
• 💵 Ellen Schultz and Robert Thompson lead the Data Solutions and Analytics team within T-Mobile's Finance department, focusing on supporting back-office operations with data solutions.
• 📄 The team identified four key pain points in their data operations: data latency, data mobility, performance issues, and high cloud spend, which informed their data strategy.
• 📈 They articulated a data strategy with four pillars: low latency data, high data mobility, predictable performance, and economies of scale, aiming to address the identified pain points.
• 💲 The shift from Azure Data Warehouse to a lake house architecture was driven by the need to scale out instead of up, improve data mobility, and reduce costs.
• 📱 The migration to the lake house architecture resulted in significant improvements: a 75% reduction in data refresh times, a 30% increase in data connections, a 60% reduction in query failure rates, and a 50% decrease in Azure spend.
• 💵 The team managed stakeholder concerns by outlining the benefits of the migration, ensuring minimal user impact, and leveraging Azure Active Directory groups for security.
• 💳 The success of the lake house architecture has led to its adoption across T-Mobile, with a focus on shared governance tools and standardized ADF templates for data management.
• 📗 The team learned that real-time data provision is often limited by source systems rather than their lake house capabilities, emphasizing the need for robust data models for governance.
• 📵 Moving forward, the team is focusing on simplifying the lake house by introducing a data model, unifying the compute experience, and encouraging a global perspective on data usage across different business units.

Q & A

What was the main goal of updating T-Mobile's data platform?

The main goal was to articulate the value of updating their data platform to senior leadership, particularly non-technical leaders, to gain their buy-in for the change.

How did T-Mobile's data team map their data strategy to business units' pain points?

They derived their strategy from direct feedback received from the business units, focusing on four key pillars: data latency, data mobility, performance, and economies of scale.

What was the role of the data team within T-Mobile's Finance department?

The data team, led by Ellen Schultz, was part of the procurement and supply chain group within Finance, supporting a back-office heavy business unit.

What were the pain points experienced by the procurement and supply chain group before the lake house implementation?

The pain points included latency in data, performance issues with queries, high cloud spend, and the need for more data mobility.

How did the data team translate the pain points into a data strategy?

They aggregated the feedback into four pillars of data strategy: data latency, data mobility, performance, and economies of scale, which they then used to map out their strategy.

What was the significance of the lake house architecture for T-Mobile?

The lake house architecture addressed the identified pain points by providing a solution that allowed for workload isolation, data mobility, predictable performance, and scalable economics.

What were the metrics used by T-Mobile to measure the success of their data strategy?

The metrics included refresh cadence, ETL processing time, data mobility connections, data availability, query failure rate, and infrastructure spend.

What was the outcome of migrating to the lake house architecture?

T-Mobile saw significant improvements in data refresh cadence, data mobility, performance, and cost savings, with a reduction in Azure spend by $120k a month.

How did T-Mobile manage concerns from senior leadership about the migration?

They addressed concerns about time, cost, impact, and security by providing clear plans, showing incremental benefits, and using Azure Active Directory groups for security.

What are the current focuses for T-Mobile's data team after the migration to the lake house?

The current focuses include simplifying the data in the lake house, introducing a data model, unifying the compute experience, and driving towards a shared lake house architecture across the enterprise.

What challenges did T-Mobile face in getting users to adopt the new lake house architecture?

Old habits and a siloed approach to data management were challenges, with users hesitant to switch and a focus on local needs over global data perspectives.