Skip to content 
Patties Food Group
A Recipe for Scalable Data
Cooking up a modern data platform on Snowflake, dbt, and Fivetran to replace a fragmented legacy stack.
The client
Patties Food Group is a leading Australian supplier of savoury and dessert products for the food industry. A small data team has been maintaining hundreds of tables and serving thousands of reports for internal and external stakeholders. Their existing stack, built on Azure Data Factory, TimeXtender, Azure storage, and on-premise IBM Cognos and ThoughtSpot, had reached the limits of what it could support.
Key metrics and results
50% faster delivery of new data products
Cloud-native platform ready for AI and ML
Sales team empowered to prioritise targets
Data latency down from 6 hours to 15 minutes for the critical dashboard
The challenge
Patties had built up a low-code, business-user-friendly tech stack over the years, and it had served the company well in the early stages. As the business grew, though, the same stack began holding the data team back.
Business users had the freedom to create thousands of reports independently across BI tools, and this added a transformation layer with no version control behind it. The same metric could be calculated differently depending on who built the report, which would lead to discrepancies that were hard to trace.
Data pipelines drew from multiple sources on different schedules, several of them triggered manually, with no single orchestrator tying them together. On top of this, there was no way to assess the impact of a change before it went live, so bugs often reached reports before anyone caught them.
None of this was unusual. These are common growing pains for data teams on legacy, low-code stacks, whatever industry they sit in. Patties recognised the need for change and set out clear requirements for the migration: full data lineage from storage through to the BI layer, pipelines that are easy to maintain and debug, a stack that could support future AI development, and minimum disruption to the business along the way.
The Solution
Identifying migration objects
Before touching any code, the team needed to know what was actually being used. Using ThoughtSpot's API and Cognos's built-in features, we pulled real usage statistics across thousands of reports and tables. This narrowed the list of objects worth maintaining and optimising, and gave Patties' data team a clear, organised way to communicate the upcoming changes to business users.
Building the new platform
We moved the transformation layer from TimeXtender to dbt Cloud, storage from Azure SQL Hyperscale to Snowflake, and ingestion from Azure Data Factory and TimeXtender to Fivetran. On dbt Cloud, we set up a unified orchestrator along with a data catalogue, lineage tracking, and version control for the whole code base.
Development and production environments were separated, with tests required before anything reached deployment. Snowflake IP blocking and single sign-on across all platforms were added to meet Patties' security requirements.
Ensuring quality throughout
Tests were added to dbt models across the entire project, catching most issues at the development stage rather than after reports went live. To confirm the migration itself was sound, we audited the new models against the legacy tables, checking that every number matched before cutting over. The production pipeline has since run at a 99% service level agreement for six months.
Knowledge transfer
Handing over a new platform is only useful if the team behind it can run it. We held weekly training sessions with Patties' team, built around a hackathon format to keep engagement high and make the learning practical rather than theoretical. These sessions also gave our own team a closer look at Patties' legacy stack and the reasoning behind it. Regular check-ins kept everyone aligned on progress, surfaced blockers early, and made sure the team had access to the materials they needed as they picked up the new tools and ways of working.
The result
With a modern, version-controlled platform in place, Patties' data team no longer spends its time chasing pipeline fixes or reconciling mismatched metrics.
Data lineage now runs from storage through to the BI layer, so changes can be traced and debugged quickly rather than investigated from scratch.
Reports draw from a single, tested source of truth, and dashboards that once took hours to refresh now update in minutes, giving the sales team the timely information it needs to prioritise targets.
Moving to the cloud has also opened the door to the AI and ML initiatives Patties wanted to pursue. Overall, the team has shifted its focus from ad hoc pipeline fixes to applying cutting-edge skills to genuine business problems.
Let’s walk the walk together
Our data expertise translates into cutting-edge, modular solutions that empower people and help organisations thrive in the era of AI.
See related stories
We work with modern organisations to strategise and build cutting-edge solutions, help them adopt data & AI innovation, and nurture key competencies for scaling in the long run.