Overview
How FoodQuest Built a Centralised Cloud Lakehouse Across All Its Restaurant Brands
FoodQuest is a UAE-based multi-brand restaurant management organisation operating iconic franchise brands including Al Baik, Denny's, TATEL, and TOTO. As the business expanded rapidly across brands and locations, its data infrastructure had not kept pace, leaving operational decisions dependent on fragmented reports, manual processes, and disconnected systems across POS, finance, and workforce management.
TGH designed and delivered a cloud-native Medallion Lakehouse platform on Azure, ingesting data from all source systems via Azure Data Factory, processing it through Bronze, Silver, and Gold layers on ADLS Gen2 with Azure Databricks and Delta Lake, and governing access through Unity Catalog, giving FoodQuest a single, trusted, scalable analytics foundation across every brand and location.
TGH built us the centralised data platform our multi-brand expansion demanded. We now have reliable, automated ingestion from POS, finance, and workforce systems, stable Medallion pipelines, and executive analytics that actually reflect what's happening in our restaurants, in real time.
The Challenge
Five Data Infrastructure Gaps Blocking Operational & Executive Visibility
FoodQuest's rapid multi-brand expansion had outgrown its data capabilities. Each restaurant brand and system generated data independently, but none of it was connected, governed, or trusted enough to support the kind of real-time, store-level and executive-level analytics the business needed to manage operations and drive growth decisions.
Fragmented data across POS, finance, and workforce systems: POSIST POS API, NetSuite Finance (GL & cost data), and the Workforce System (shifts & attendance) all operated in silos with no unified ingestion layer, making cross-system reporting impossible without manual extraction and reconciliation.
Manual reporting with high latency and limited reliability: Store managers and finance teams relied on manually compiled reports with significant delays, meaning operational decisions were being made on stale data that didn't reflect current trading, staffing, or cost positions.
Inconsistent schemas from POS APIs impacting downstream reporting: The POSIST POS API returned inconsistent schemas across different restaurant brands and locations, causing downstream reporting failures whenever schema changes were pushed, with no resilience or schema drift handling in place.
No governed, scalable platform for store-level and executive analytics: Without a structured data platform, the business had no reliable foundation for store-level performance dashboards, food cost analysis, or executive KPI reporting, limiting the leadership team's ability to make data-driven decisions across the portfolio.
Limited visibility into sales performance, food cost, and margins: Key financial metrics, revenue by brand, food cost as a percentage of sales, and margin by location, were not available in near-real-time, leaving commercial teams flying blind on the levers that most directly impact profitability across the franchise network.
Our Solution
A Medallion Lakehouse Architecture Built on Azure for Multi-Brand Restaurant Analytics
TGH designed a cloud-native Medallion Lakehouse on Azure Data Lake Storage Gen2, orchestrated by Azure Data Factory for REST API ingestion and processed by Azure Databricks with PySpark and Delta Lake. The architecture follows the proven Medallion pattern, separating raw ingestion (Bronze), cleansed and merged data (Silver), and KPI-ready aggregations (Gold), with Unity Catalog providing fine-grained governance and access control across all data assets.
💡
Delta Lake's ACID transactions and schema enforcement were critical for FoodQuest, ensuring that schema changes from the POSIST POS API don't cascade failures into Silver and Gold layers, and that every data write is either fully committed or cleanly rolled back, keeping the Lakehouse always in a consistent state.
Implementation
A Five-Phase Engineering Delivery From Source to Insights
Phase 1
Discovery & Architecture Design
Phase 3
Silver & Gold Processing
Phase 4
Governance & Analytics
01
Source System Assessment & Medallion Architecture Design
TGH conducted a comprehensive assessment of FoodQuest's three source systems: POSIST POS API (sales & order data), NetSuite Finance (GL & cost data), and the Workforce System (shifts & attendance), mapping all data schemas, API structures, ingestion volumes, and business KPI requirements. The output was a complete Medallion Lakehouse blueprint defining the Bronze, Silver, and Gold layer schemas, partition strategies, and Delta Lake table design.
02
Azure Data Factory Pipelines: REST API Ingestion &
TGH built Azure Data Factory pipelines for REST API ingestion from all three source systems, using incremental and CDC-based loading strategies to minimise API load and ensure data freshness. All raw data was landed in the Bronze layer on ADLS Gen2 in Delta Lake format with full schema tracking, creating a reliable, complete historical record of every source system event without transformation or loss.
03
Azure Databricks : Cleansing, Merging & Schema Drift Handling
PySpark notebooks on Azure Databricks processed Bronze data into Silver, applying deduplication, null handling, data type standardisation, and cross-source merges. Schema drift handling was built specifically for POSIST POS API inconsistencies, allowing downstream Silver and Gold layers to remain stable even when source API schemas change unexpectedly, eliminating the pipeline failures the business had previously experienced.
04
: for Store-Level & Executive Analytics
The Gold layer was built as a set of business-ready, pre-aggregated Delta tables, including sales performance by brand and location, food cost as a percentage of revenue, labour cost ratios, margin analysis, and shift-level attendance metrics. These Gold tables serve as the direct source for Power BI dashboards and Databricks AI/BI Genie, ensuring all reporting is built on a single, consistent, governed data layer.
05
Unity Catalog Governance, Power BI Dashboards & Go-Live
Unity Catalog was configured across all Lakehouse layers, implementing fine-grained access control by brand, location, and role, with full data lineage tracking and audit logging for enterprise compliance. Power BI dashboards were built on the Gold layer for store managers and the executive team, and Databricks AI/BI Genie was enabled for natural language analytics queries. Production go-live was executed with zero data loss and stable pipeline performance from day one.
Results & Business Outcomes
Five Measurable Outcomes Across Data Quality, Speed & Scalability
The FoodQuest Medallion Lakehouse delivered immediate and sustained improvements across every dimension of the client's data operations, from automated ingestion reliability to governance maturity and the scalable foundation needed to support future brand additions and location growth.
🔄AutomatedReliable ingestion across all heterogeneous source systems, zero manual extraction
✅QualityClear separation of concerns with strong data quality controls at every Medallion layer
⚡FresherReduced processing time and dramatically improved data freshness for operational decisions
🛡️ResilientStable pipelines resilient to schema drift and operational failures from POS APIs
🔒GovernedSecure, auditable data access aligned with enterprise standards via Unity Catalog
🚀ScalableA reusable Lakehouse foundation ready to onboard new brands and restaurant locations
Reliable, automated ingestion across heterogeneous source systems, POS, finance, and workforce data now flows on schedule without manual intervention or data engineering firefighting
Clear separation of concerns with strong data quality controls, the Medallion architecture ensures raw data is preserved, cleaned data is validated, and Gold metrics are always business-ready
Reduced processing time and improved data freshness: Incremental and CDC-based ingestion replaced full daily loads, cutting pipeline run times and delivering more current data to dashboards
Stable pipelines resilient to schema drift and operational failures, POS API schema changes no longer cascade into reporting failures, protecting the Silver and Gold layers from upstream inconsistencies
Secure, auditable data access aligned with enterprise standards, Unity Catalog provides fine-grained RBAC by brand and location, with full lineage and audit trail for compliance and data governance maturity
Data Pipeline ReliabilityNear-100%, Automated
Medallion architecture with Delta Lake ACID guarantees eliminated silent data failures
Schema Drift ResilienceFully Handled
PySpark schema drift handling prevents POS API changes from breaking downstream Silver & Gold layers
Data Freshness ImprovementSignificantly Improved
Incremental & CDC-based loading replaced heavy full daily loads with near-real-time ingestion
Analytics CoverageStore-Level + Executive
Gold layer KPI tables now power both granular store dashboards and executive brand-level reporting
Technology Stack
Platforms & Tools Used in This Engagement
TGH built FoodQuest's Lakehouse entirely on Microsoft Azure's native data platform stack, combining the industry-leading Databricks Lakehouse with Azure Data Factory orchestration, Delta Lake storage, Unity Catalog governance, and Power BI for business intelligence delivery.
Ingestion
Azure Data Factory (ADF)
Processing
Azure Databricks · PySpark · Delta Lake
Analytics
Power BI + Databricks AI/BI Genie
Azure Data Factory
Azure Databricks
PySpark
Delta Lake
ADLS Gen2
Unity Catalog
Power BI
Databricks AI/BI Genie
NetSuite Finance API
POSIST POS REST API
Business Benefits
Key Benefits Delivered to FoodQuest
The Medallion Lakehouse platform delivered five strategic data capabilities that directly support FoodQuest's multi-brand expansion strategy and operational efficiency goals.
📥
Centralized Data Ingestion
All POS, finance, and workforce data lands in a single governed Lakehouse, eliminating siloed systems and manual extraction
🏗️
Medallion Lakehouse Architecture
Bronze, Silver, and Gold layers provide clean separation of raw, cleansed, and business-ready data at every stage
🔁
Incremental & CDC-Based Loading
Efficient incremental pipelines and change data capture dramatically improve data freshness while reducing compute costs
📊
Data Quality, Reliability & Observability
Schema enforcement, data quality checks, and full lineage tracking ensure every metric in every dashboard is trustworthy
🔒
Governance & Secure Access
Unity Catalog delivers fine-grained RBAC by brand and location with full audit trails for enterprise compliance readiness
🚀
Scalable Foundation for Growth
New brands and restaurant locations can be onboarded to the Lakehouse rapidly using reusable pipeline templates
🥇Medallion layers enforce discipline Separating raw, cleansed, and aggregated data layers prevented the technical debt that accumulates when transformations are applied directly to raw ingestion
🛡️Schema drift handling is non-negotiable Building schema evolution logic into the Silver layer from day one saved FoodQuest from the POS API inconsistencies that had previously broken their reporting
🔁CDC > full loads Incremental and CDC-based ingestion via ADF reduced pipeline runtime, cut ADLS storage costs, and delivered significantly fresher data to analytics consumers
🔒Governance from day one Implementing Unity Catalog at platform build rather than retrofitting later meant access control, lineage, and audit capability were available from the first production table
