Situation
The business context
Multiple restaurant brands and channels generated large volumes of guest activity, but turning that data into a meaningful view of the guest was difficult and labor-intensive. Guest profiles had to be assembled on demand by a small number of specialized data analysts who joined data across multiple sources and performed ad hoc identity matching for each request. The process was slow, difficult to scale, and lacked a consistent approach to resolving guest identity across brands and channels. The underlying data integration processes were also slow and brittle, with failures requiring manual intervention. Together, these limitations made timely, reliable guest insight difficult to produce and even harder to scale across the enterprise.
Task
The leadership mandate
As a senior IT leader, I was responsible for developing and executing a multi-year data transformation strategy that would turn fragmented guest data into a strategic enterprise asset. We needed a reliable, scalable data capability that could provide a consistent view of the guest and make trusted, timely insights more broadly accessible across the enterprise.
Strategy
The strategic approach
My strategy was to build guest analytics as a sequence of expanding enterprise capabilities rather than approaching it as a single platform replacement. We would first resolve fragmented guest identity, improve the reliability of data movement, and progressively broaden access to trusted, timely insights. I deliberately avoided defining the long-term target architecture too early. We were several years away from needing it, and both the technology landscape and the needs of the business were evolving rapidly. Instead, platform discovery was built into the later stages of the roadmap, allowing us to make that decision closer to the point of need while the earlier stages continued to deliver business value and establish the capabilities required for what came next.
Action
How the work moved forward
- Sequenced the multi-year roadmap: Aligned technology investment to the company’s vision for becoming more guest-centric and data-driven, organizing delivery around progressive capabilities that created value while preparing the next stage. Platform discovery was intentionally scheduled later in the roadmap so the eventual target architecture could reflect then-current business needs and technology options.
- Improved the economics of the existing platform: Identified an opportunity to shift Azure Synapse compute from pay-as-you-go pricing to reservations, lowering the cost of the existing environment while the broader data transformation continued.
- Established a unified guest identity: Connected more than 500 million transactions through a tokenized identity-resolution framework, replacing ad hoc guest matching with consistent profiles that could support personalization, loyalty, and data-driven marketing.
- Stabilized data movement before scaling access: Replaced brittle PowerShell ETL processes with Azure Data Factory, creating a more reliable and scalable integration foundation with less dependence on manual intervention.
- Established and delivered the target architecture: Led platform discovery and selection as the roadmap reached the appropriate stage, incorporating evolving requirements for analytics, generative AI, and a composable CDP. Selected and led the migration from Azure Synapse Analytics to an Azure Databricks Lakehouse, improving platform economics and processing performance while creating a more extensible data foundation.
- Broadened governed access to trusted data: Used Azure Databricks Unity Catalog to provide appropriate users and groups with governed access to trusted data, reducing dependence on specialized analysts and making timely guest insight more broadly accessible across the enterprise.
- Turned guest feedback into actionable insight: Secured incremental funding to apply generative AI to unstructured guest feedback, translating large volumes of qualitative feedback into insights the business could use to identify opportunities to improve guest experience and operator performance.
Result
The measurable difference
- 500M+ transactions unified through identity resolution, creating consistent guest profiles across brands and channels.
- 95% reduction in integration failures, improving data reliability and reducing manual intervention.
- 35% reduction in Azure Synapse compute costs by shifting from pay-as-you-go pricing to reservations.
- 90% reduction in time-to-insight, enabling business teams to act on information substantially faster.
- 30% reduction in platform TCO through migration to the Azure Databricks Lakehouse.
- Four consecutive quarters of year-over-year improvement in guest experience and operator performance scores as guest feedback was translated into actionable insights.
- Established a scalable data foundation for personalization, loyalty, advanced analytics, and AI/LLM use cases.
Leadership insight
Long-term transformation does not require making every technology decision at the beginning. A clear capability roadmap can provide direction while preserving the flexibility to make architectural decisions closer to the point of need, when business requirements and technology options are better understood.