Unified data platform for a multi-site catering and hospitality business

Contract catering and hospitality services provider · Microsoft Fabric · Medallion architecture · Power BI

Engagement at a glance

Fragmented SaaS landscape unified into a single governed Lakehouse
Microsoft Fabric platform with bronze, silver, and gold medallion layers
Three reporting domains served: internal management, external clients, and purchasing intelligence
Client-facing dashboards delivered as a value-added commercial service
Sundew led platform architecture, data engineering, and Power BI semantic model design

The business context

A multi-site catering and hospitality business sits at the intersection of five operational disciplines: financial control, point-of-sale operations, workforce scheduling, purchasing, and sustainability. Each discipline runs on its own SaaS platform, and each platform holds a critical slice of the business. Without a unified data platform, the leadership team cannot get a coherent view of performance at the site, client, or organization level, and the commercial clients the business serves cannot get the reporting that increasingly defines a contract food service relationship in 2026. The brief was to unify the data, raise the bar on data quality, and turn reporting from an internal hygiene function into a commercial differentiator that the business could present to its own clients.

The challenge

Unifying SaaS data is conceptually simple and engineering-hard. Each system carries its own schema, its own update cadence, its own API quirks. Stitching them into one Lakehouse where every metric is reconciled, every dimension is canonicalized, and every report is governed is the kind of engineering that separates a working data platform from a fragile data swamp.

Beyond internal reporting, the business needed to deliver meaningful, trust-building dashboards to its own clients, which raises the bar on data quality, consistency, and presentation, because a client-facing report carries reputational weight that an internal dashboard does not.

The Sundew approach

Sundew designed and delivered a modern data platform on Microsoft Fabric, bringing together financial, point-of-sale, workforce, sustainability, and purchasing data from multiple SaaS providers into a single, governed Lakehouse environment.

Data ingestion and medallion architecture

Data is ingested in batches via REST APIs across all source systems, feeding into a medallion architecture. Raw data lands in the bronze layer, is progressively cleaned and standardized in the silver layer, and is modeled into business-ready datasets in the gold layer. This layered approach systematically enforces data quality before any metric reaches a report, so every dashboard the business publishes has the same provenance. Lineage is traceable from the report back to the source row.

Reporting across three domains

Management dashboards for internal leadership cover profit-and-loss reporting across sites and business units, operational efficiency tracking, and site performance benchmarking. The leadership team can view operations at the organizational, regional, or single-site level of granularity.

Client dashboards are delivered directly to the catering company's own clients. These reports turn operational data into a value-added commercial service, including sales performance by location, customer satisfaction metrics, CO2 emissions reporting, food waste tracking, and reduction trends. In an industry where ESG and sustainability reporting are increasingly contractual, this layer turns the reporting platform itself into a commercial asset.

Purchasing intelligence covers vendor pricing analysis, spend analytics across categories, and inventory management visibility, providing the purchasing function with the depth of data required to negotiate, consolidate, and optimize across the supplier base.

Power BI front end

Power BI serves as the unified reporting layer for both internal and external audiences, with role-based access ensuring each stakeholder, internal leadership, regional manager, purchasing analyst, or external client sees exactly what is relevant to them. The semantic model is governed centrally, so the metric an internal manager reads matches the metric the client sees, with appropriate aggregation and access rules applied.

The outcome

The platform transformed a fragmented collection of SaaS data into a single source of truth, enabling faster, more confident decisions for internal leadership while simultaneously delivering sustainability and performance reporting that strengthens client relationships.

The client-facing dashboards shifted the conversation with end clients from "did we hit the SLA?" to "what is the waste-reduction trend across your sites this quarter?"-the conversation that retains and expands food service contracts in the modern procurement environment.

What this means for your enterprise

If your enterprise runs on a stack of SaaS platforms and the data has never been unified, the question is not whether to consolidate; it is which architecture pattern to consolidate on. The medallion architecture on Microsoft Fabric is the right choice when batch ingestion is acceptable, and the reporting audience includes both internal and external stakeholders with different access requirements. The harder commercial insight from this engagement is that the client-facing reporting layer is no longer a nice-to-have. In service industries with embedded sustainability and ESG requirements, it is becoming a differentiator for winning and retaining contracts.

"When the data platform is engineered well, the reporting itself becomes a commercial product."

Talk to our Experts

Technologies and capabilities

Ready to start?

See how Sundew's Managed Engineering Services fit your industry roadmap.

Schedule a Call

Frequently Asked Questions

How does Sundew handle API changes from SaaS vendors without breaking the entire pipeline?

The medallion architecture provides a natural buffer. Raw data is preserved in the bronze layer, so when a vendor updates their API schema, the impact is isolated to the ingestion and silver-layer transformation logic rather than cascading into reports. This separation is what keeps a Lakehouse from becoming a fragile data swamp.
Can client-facing dashboards be configured so that external stakeholders only see their own data?

Yes. The Power BI semantic model is governed centrally with role-based access control, meaning each external client sees only the data relevant to them: site performance, waste trends, and CO2 emissions, without any visibility into other clients' operations or internal financials.
Why did Sundew choose Microsoft Fabric for this engagement?

Microsoft Fabric's multi-workspace architecture and Power BI's row-level security make it well-suited when reporting must serve both internal and external audiences from a single governed semantic model. Combined with native Lakehouse support and medallion-pattern compatibility, it was the right fit for batch ingestion at this scale.
How is data governance handled when internal and external stakeholders share the same underlying datasets?

The semantic model is governed centrally, which means metric definitions are consistent across all audiences. What changes is that the access layer, aggregation rules, row-level filters, and workspace permissions determine what each stakeholder sees, not separate, independently maintained datasets that drift apart over time.
What differentiates Sundew's approach to modern data platform delivery?

Sundew leads platform architecture, data engineering, and semantic model design within a single engagement rather than treating them as separate workstreams. The result is a platform where data quality, governance, and reporting are resolved together, turning consolidated data into a reporting layer that internal teams trust, external clients act on, and the business can take to market as a commercial asset.