TLDR: Food & Beverage (F&B) franchisees consistently face challenges in automating their financial control processes, consolidating financial data, and driving sales and profit growth across their stores. Ezora is a scalable platform built to address financial reconciliation, reporting, and analytics for these franchisees. Ezora needed a high-performance data warehouse to simplify data pipelines and deliver granular, sub-second drill-downs for its customer base. Firebolt’s capabilities in performance acceleration through indexing and storage optimizations helped deliver on these expectations.
The F&B Challenge
An F&B franchisee interacts with many disparate proprietary applications daily to run operations. Typical systems include point-of-sale data capture, financials, inventory and asset management, human resources, and workforce planning systems. These systems vary by franchise. A single franchisee with multiple franchises ends up interfacing with proprietary systems specific to each franchise. As a result, building a unified view of the business requires manual processes for bookkeeping, reconciliation, and financial reporting leading to inefficiencies and increased risk of errors. Complex, one-off data Integrations are necessary to generate store key performance indicators (KPIs) on sales, inventory, labor, fraud risk, customer experience, and operational costs.
Moreover, with the advent of providers like Uber Eats, DoorDash, and others, the system of record for a franchisee goes beyond the franchise. Analytics KPIs need to be tracked to understand delivery and financial performance and the demands these new services place on individual stores. These challenges and the need for highly skilled IT staff make the status quo hard to work with for F&B franchisees.
Ezora's platform eliminates the above challenges that franchisees face day in and day out. First and foremost, Ezora has developed intellectual property to simplify franchisees' data integration challenges. Ezora’s vast collection of custom connectors integrates across numerous proprietary systems, simplifying data processing, financial reconciliation, and reporting.
Ezora’s Financial Control Engine automates bookkeeping, reconciliation, financial reporting, and consolidation. Ezora’s Sales and Profit Growth engine empowers store managers, area managers, operations directors, and leaders with data-driven insights to make informed decisions and drive business growth. By leveraging industry-specific knowledge and expertise, Ezora ensures the accuracy of financial data. The platform offers store performance dashboards that display KPIs derived from sales, inventory, labor, fraud risk, customer experience, and operational cost data. Real-time, daily, weekly, monthly, and year-to-date perspectives allow users to monitor outcomes, trends, and comparisons against targets and benchmarks. Additionally, the platform facilitates the creation of forecasts, targets, and performance comparisons for enhanced planning and decision-making.
The Ezora platform was initially created as an on-prem service and migrated to AWS in 2016. Since moving to AWS, Ezora has delivered best-in-class availability, ease of integration, improved scalability, and faster innovation. Ezora’s logical architecture is shown below.
Data integration uses multiple technologies such as APIs, direct database connectivity, or secure file transfer. This flexibility helps address data integration across various F&B franchises with varying connectivity requirements, scale, and throughput. Ezora manages these integrations in the form of configurable data pipelines that are automated and managed by the SaaS platform itself. Data is acquired and consolidated into Amazon Aurora for transactional data, Ezora Data Lake for raw data collection, and Firebolt Cloud Data Warehouse for analytics, reporting, and drill-downs.
Firebolt's High-Speed Data Warehousing for Subsecond Analytics and Drilldowns
One critical challenge Ezora faced in delivering efficient and comprehensive analytics was the limitations of the Aurora Postgres database. These limitations hindered their ability to store and process large amounts of raw data while providing subsecond analytics and drill-down capabilities for end-user customers.
For example, fast ingestion and drill-downs using raw data were impossible with previous technologies. To compensate for these limitations, the Ezora team had to create and perform compute-intensive aggregations ahead of time while limiting the time of data analyzed to shorter time windows. Shorter time windows limited the business operator's ability to do detailed analysis over extended periods. To resolve these issues, Ezora’s technology innovation team leveraged Firebolt’s capabilities to build their next-generation platform.
Firebolt's architecture excels in rapid data ingestion and processing. It leverages advanced parallel processing techniques, columnar storage, and compression algorithms to optimize data loading and transformation tasks. This capability significantly reduced the time required to ingest and process new data, enabling Ezora to provide up-to-date and real-time analytics to their end users. Brian Mc Intyre, Chief Technology Officer at Ezora, says, “ We have seen greater than 10X improvement in data ingestion speeds and 30X improvement in query response times, allowing us to provide more timely analytics that keeps our customers well-informed about their business developments.”
Firebolt's data warehousing technology employs unique indexing mechanisms and query optimization techniques. These advancements allow for lightning-fast query execution, even on large datasets. By leveraging these capabilities, Ezora performs complex queries and drill-downs on granular data stored in Firebolt. All done at sub-second response times. This capability to access raw data has enabled end-users to explore data at various levels of detail and gain deep insights into their financial and operational performance. From a data engineering perspective, this also eliminates data and pipeline redundancy. “The advanced speed of Firebolt enables us to run all our analytics off the underlying transactional data sets without the need for any batched data or aggregation routines, allowing us to simplify our data model. We don’t need to denormalize or create secondary pipelines. This helps accelerate the timeline to turn around new data services for our existing clients and new client implementations. From the data engineering and client support team’s standpoint, we have reduced the time required to launch new data services by 40%, which translates to several weeks of work.”, according to Brian.
Ezora’s journey with Firebolt has not been without challenges. “When we started, Firebolt was an append-only data store, which meant we had to rule out use cases requiring updates, deletes, or find workarounds. Firebolt’s Engineering team worked with us through the launch of updates and deletes supporting new use cases. For example, employee time card corrections can be implemented directly on the analytics system with transactional granularity. We look forward to working with Firebolt Engineering as the platform evolves.”
The Business Impact of Sub-second Analytics
With Firebolt's subsecond analytics and drill-down capabilities, Ezora delivered an interactive and seamless user experience. End-users could effortlessly navigate through different levels of data hierarchy, exploring data points, comparing YTD metrics across multiple years, and uncovering trends and patterns. This interactivity fosters data-driven decision-making and empowers users to make timely and informed choices to improve store performance and drive business growth.
Alan O’Neill, Chief Executive Officer at Ezora, says,“Our customers can now dynamically analyze their data across brands, geographies, teams, stores, sales channels, and products without waiting for results or running out of bandwidth. Switching to Firebolt has enabled us to expand our enterprise-class clients and apply our solution to new market opportunities where the data volumes are much higher than our traditional customers. We can now provide sub-second analytics to retail clients with 10k + SKUs, proving to be a real growth market for our business.”