What was CONA’s initial BI focus?
From its inception, CONA BI’s goal was to establish a self-service BI environment that enabled our multiple bottling partners to create and manage their reporting based on local needs and preferences. The CONA BI team achieves this by providing a cohesive system for data management, model development, and security while enabling our Bottlers to share the cost of supporting the ecosystem. We realized that our approach to data needed to evolve, and we would need a modern data management platform.
What were the problems that CONA faced?
One undeniable issue loomed over us as we grew and matured: query performance. Despite multiple attempts to tune the environment with many experts, it was an issue we were not able to overcome. To account for the poor performance many teams went as far as resorting to extracting and blending their data outside of the legacy data warehouses in other database systems. This problem pushed us to evaluate what we have versus where we would like to be in the future.
As we started to document the foundation for our future, we realized the focus we placed on performance masked even more issues. With the cost of our legacy platform being driven by storage, our environment was costly and growing. Objectives such as data archiving and needs to keep data for long periods of time to support Advanced Analytics only exasperated the storage cost problem. In addition to the physical storage cost, we found our implementation of our legacy platform caused high operational expenditures and administrative costs. Non-SAP data integration, sharing data amongst internal business partners, and managing security were all overly complex and costly. This led to the position where our customers did not have the freedom to fully utilize the environment according to local needs. In the end, the pain points we needed to address were Capability, Cost, Performance, and Flexibility.
What did CONA need to focus on to combat these issues?
We started with aligning on Guiding Principles for evaluating destination technology options. We wanted to ensure that we had the right balance between Cost and Capability. We also wanted a destination that we could grow with, accommodating our needs today and in the future. We utilize many SaaS systems, so blending, connecting, and moving data from other data sources was a key focus area. Because our Bottlers work so closely with multiple strategic partners, we were looking for a platform that allowed us to provide information to other organizations with ease. Lastly, the hurdle that started our journey, Performance was still a primary focus. We wanted a solution which met very specific SLA expectations with regards to Extract, Transform, Load(ETL) query, and reporting performance. Finding a platform that was able to meet our guiding principles became the strategic priority to achieve our next steps in our growth.
What direction did CONA choose to advance toward?
After careful consideration of our needs, we explored multiple destination options. We evaluated each platform based on consistent success criteria, which we aligned with our Guiding Principles. In November of 2019, after the evaluation of multiple options, a team of leaders from CONA and each of our Bottlers officially decided Snowflake was the right option as our destination data warehousing service provider. We finally had our direction, and we started our 2 years journey to migrate to Snowflake.
How has Snowflake Inc. impacted the success of CONA?
Although having the destination was a substantial milestone, the pressure was on to realize the value creation opportunities which led us there, and execution became critical. Since choosing Snowflake Inc., we’ve been working through our migration, focused on ensuring that we integrate the lessons learned from the past six years. Since we were freed from the constant threat of poor performance, we were able shift our focus on value added opportunities like enhancing our new BI models to also better accommodate local customizations, enabling our Bottler BI teams to more effectively meet their BI and business analysis needs moving forward.
Now that we are 90% of our way through our migration, we are starting to experience many of the benefits we expected to achieve including:
- Near Realtime Data Ingestion – This is a major hurdle that we’ve been working through for the past two years. We now have made a decision to move forward with HVR to help move our data efficiently from SAP to Snowflake.
- Native Snowflake capabilities – Modern cloud-native architecture, that decouples compute from storage, allows us to truly enable data capabilities in one platform, which is critical to connect the dots between datasets and generate Business insights that were not possible before.
- Excel interaction with Snowflake – An Excel add in that was available for our legacy platform has been a staple of our organization and our customers. There was not a relevant Excel based option for interacting with Snowflake, so we built one. We call this SnowSquall, and we plan to sell it in the open market.
- Code Management via DevOps.
- Performance – With Snowflake, we can select our performance level, meaning we’re now able to meet individual requirements that help balance cost and speed.
- All data in one place. As we continue to evolve, focused on ‘Best of Breed’ technologies, we’re now able to bring our Blue Yonder, Salesforce, and PACE data together into the same data warehouse that has our SAP data.
- Local Freedom – Our Bottlers now have the freedom to build and manage their local Snowflake instance as they see fit. This allows us to be connected for common solutions and independence for local solutions.
- Sharing information – Snowflake’s Data Share capability is unlike any we’ve experienced, making it easy to share data with any organization that has Snowflake. Sharing data from Data producers to data consumers, without having to moving it and copying it, is impactful.
We are looking forward to close out the final pieces of our migration, expecting to deliver our full project commitments On Time, In Full, and Within Cost. Even though we aren’t completely done yet, the early results are clear that the forecast for BI at CONA is excellent with Snowflake.