A Cloud Journey Begins with a Business Problem, Not Technology
Today, cloud transformation has become the norm. Most organizations have either completed or are well into their migration to the cloud. Looking back on JANA’s journey, one lesson stands out above all others: successful cloud transformations do not begin with technology. They begin with a clear understanding of the business problem.
Eight years ago, I began advising JANA’s executive team on how to operationalize its pipeline risk models. JANA had already developed sophisticated mathematical models for predicting pipeline integrity and failure risk. The challenge was transforming those models into a production platform capable of ingesting, integrating, and analyzing large volumes of real-world client data.
Real-world data existed across numerous disconnected systems. It arrived in different formats, contained inconsistencies, and suffered from varying levels of quality. Before a single risk model could be executed, we needed to answer much more fundamental questions. How would we integrate these systems? How would we ensure the accuracy and completeness of the data? How would we establish confidence that the outputs produced by our models could be trusted?
These were not technology problems. They were business problems that required technology to solve.
At the time, however, JANA had virtually no enterprise data infrastructure. Much of the company’s analytical work was performed from engineers’ laptops, which was sufficient for research but not for delivering a scalable commercial platform. To establish a foundation, I designed and implemented an on-premises enterprise data platform. It significantly improved our capabilities, but it also highlighted the limitations of traditional infrastructure as the business continued to grow.
By 2019, cloud computing had matured enough to become a realistic option. Microsoft Azure offered services such as Azure Data Factory, Azure SQL Managed Instance, and Azure Data Lake Storage, but neither JANA’s Data team nor its Software Engineering team had meaningful cloud experience. Rather than attempting to learn everything through trial and error, I engaged Newcomp, a consulting firm that had already delivered several Azure implementations. Their experience accelerated our learning and helped us begin JANA’s cloud transformation with a sound architectural foundation.
In March 2020, we deployed our first Azure environment. The initial platform provided secure networking, identity and access management, data storage, orchestration, and database services. We migrated client data into the environment and began developing production data pipelines, not simply to migrate workloads, but to understand how the platform behaved under real operating conditions.
The proof-of-concept phase proved invaluable.
Within six months, we had identified several important architectural lessons. Azure Data Factory did not provide the level of stability we required for every workload. Our client onboarding process demanded more sophisticated managed file transfer capabilities than we initially anticipated. Most importantly, Azure SQL Managed Instance did not deliver the performance necessary for the geospatial processing that lay at the heart of JANA’s risk models.
Rather than forcing the business to adapt to technology, we returned to the original business requirements and re-evaluated the available cloud platforms. We compared Google BigQuery, Amazon Redshift, Azure Synapse Analytics, and Snowflake against the capabilities that mattered most to JANA.
BigQuery emerged as the strongest architectural fit. Its native geospatial processing, integrated machine learning capabilities, and fully managed analytical architecture aligned exceptionally well with the computational demands of pipeline risk modeling.
Instead of replacing Azure, I designed a federated cloud architecture that integrated Microsoft Azure with Google Cloud Platform. Azure continued to provide operational services, identity integration, and orchestration, while BigQuery became the analytical engine supporting JANA’s data platform. By leveraging the strengths of both cloud providers, we created a solution that was better than either platform could provide independently.
Before the platform entered production, we conducted penetration testing, remediated the identified vulnerabilities, and validated the final architecture with cybersecurity to ensure the environment met enterprise security standards.
Looking back, the experience reinforced an important lesson. Successful cloud transformation follows a disciplined methodology.
1.Begin with clearly defined business objectives.
2.Identify the capabilities required to achieve those objectives.
3.Evaluate technologies against business requirements, not vendor marketing.
4.Validate architectural assumptions through proof-of-concept implementations.
5.Select technologies based on measurable trade-offs involving functionality, security, performance, scalability, and operational complexity.
6.Design an integrated architecture that supports long-term business growth.
7.Secure executive sponsorship by demonstrating measurable business value.
8.Build the platform, validate its security, establish governance, and operationalize it with the right people and processes.
Technology was never the destination. It was simply the enabler.
The results validated that philosophy. In 2019, JANA had no enterprise cloud platform. Today, that platform has been deployed more than 80+ times across 35+ customer organizations, providing a secure, scalable foundation for pipeline risk modeling while becoming a significant contributor to the company’s revenue growth.
Perhaps the most important lesson is this: organizations rarely succeed because they choose the “best” cloud platform. They succeed because they understand their business well enough to choose the architecture that best supports it. Cloud services will continue to evolve, but aligning technology decisions with business strategy remains the enduring foundation of every successful cloud journey.
