Transforming a Digital Advertising Company’s Data Platform for the Future

When a leader in programmatic digital advertising saw their business accelerating, they knew their legacy data platform would need an upgrade. To stay ahead, they needed a faster, more flexible solution that could grow with them, without locking them into a single cloud provider. The company was ready to modernize their data architecture to improve performance, streamline costs, and build a scalable foundation for future innovation.

A technology leader in programmatic digital out-of-home (DOOH) advertising, operates one of the industry’s largest programmatic marketplaces. Their platform enables real-time bidding and automated ad placements across a diverse network of digital screens in malls, gyms, transit hubs, and urban spaces, offering advertisers unmatched targeting and reach.

The company’s success meant their legacy AWS Redshift-based data infrastructure was increasingly under pressure. Several critical issues emerged:

• Performance bottlenecks: Reports and ad hoc queries often timed out, degrading customer experience.
• Operational delays: Lack of real-time data visibility slowed down vendor payments and strained relationships with key partners like Foursquare.
• Rising costs: Their Redshift-based architecture was cost-inefficient, with compute resources that couldn’t easily scale to meet fluctuating needs.
• Growth constraints: The platform struggled to keep pace with growing volumes of event and analytics data.

After deciding to migrate from AWS Redshift to Snowflake, the company brought in SPR to design and implement a future-ready data architecture. While the platform choice had already been made, the company needed a partner who could bring deep technical expertise and guide the transition from strategy to reality.

SPR was tasked with building a modern, scalable data solution from the ground up. This included migrating both historical data and ongoing real-time streams from AWS S3, applying data engineering best practices, and creating a strong foundation for long-term analytics innovation.

“Our role was to help them build a complete, scalable data solution that would let them grow confidently,” said Mahadeva Bisappa, Data Architect at SPR.

The engagement focused on optimizing performance, reducing operational costs, and future-proofing the infrastructure to support Vistar’s expanding business and data needs.

SPR partnered closely with Vistar’s internal teams to deliver a comprehensive platform rebuild:

Key Technical Achievements:

• Full Schema Redesign: Reimagined the database structure for Snowflake, with dimensional data models optimized for reporting and analytics.
• Advanced Aggregations: Replaced manual roll-up tables with materialized views, cutting query times and simplifying data maintenance.
• Seamless ETL Processes: Engineered new ETL pipelines that extracted event data (stored in AWS S3 as Protocol Buffers) and transformed it for Snowflake using a combination of Scala scripts and Snowflake native scripting.
• Dynamic Scaling: Configured Snowflake’s compute clusters to scale up during heavy ETL processing and scale down during idle periods, drastically optimizing costs.
• Medallion Architecture Implementation: Applied a bronze-silver-gold (landing-cleaned-curated) model to structure the data lakehouse, enhancing reliability and auditability.

“We rebuilt the whole platform from the ground up,” said Bisappa. “Not just a migration, but an opportunity to modernize everything, from data models to ingestion processes."

SPR’s work ensured that the new Snowflake environment would seamlessly integrate with the company’s AWS-based operations, while offering the flexibility to evolve independently of cloud provider limitations.

When SPR wrapped up its engagement, the company had a fully functioning development environment on Snowflake, complete with historical data migrated from Redshift and real-time data pipelines flowing reliably from AWS S3. While the production rollout was scheduled for a later phase, the infrastructure SPR built laid the foundation for a smoother, faster, and more scalable future.

SPR’s hands-on work delivered immediate improvements in both performance and cost efficiency. Benchmarks showed Snowflake’s loading and query speeds were significantly faster than Redshift, and the platform’s ability to dynamically scale compute resources — and shut them down when not needed — led to smarter, leaner operations.

“We were able to demonstrate that with similar volumes, Snowflake loaded data faster, queries ran faster, and cost less. That’s a huge win,” said Bisappa.

Beyond performance, the SPR team’s implementation of a modular, medallion-style architecture and clean ETL processes meant Vistar’s internal teams could take full ownership of the solution, with minimal ramp-up time.

Outcomes included:

• Accelerated query and load performance, improving from sluggish Redshift execution times to near real-time reporting.
• Reduced operational costs through dynamic scaling and optimized compute usage in Snowflake.
• Improved maintainability and flexibility, with code that could be promoted easily across environments and adapted as business needs evolved.
• Smooth knowledge transfer, empowering the company’s in-house developers to maintain and extend the platform confidently.

For the company and SPR alike, this project was more than a technical migration, it was an opportunity to rethink what a modern data platform should look like. By decoupling from Redshift and embracing a cloud-agnostic, high-performance solution like Snowflake, The company positioned itself to be more nimble, cost-effective, and insight-driven.

One of the biggest advantages was how developer-friendly and accessible Snowflake proved to be, especially for testing, onboarding, and operational handoff. It allowed the internal team to iterate quickly and scale strategically, without needing deep cloud infrastructure expertise.

“The whole idea was to give them a real blueprint,” said Nerine Perera, SPR Data Engineer. “This is a scalable, future-proof data platform, not just a quick migration. And Snowflake made it easy to validate the approach, test, and grow from there.”

SPR also saw firsthand how important cloud abstraction can be for fast-moving businesses. Snowflake’s ability to run across multiple cloud platforms with a consistent user experience means that similar companies aren’t locked into the constraints of a single provider.

Lessons learned:

• Cloud-agnostic platforms like Snowflake offer powerful flexibility — organizations can scale and adapt without being tied to a specific cloud vendor.
• Dynamic compute provisioning leads to real cost savings, especially when compared to always-on models like Redshift.
• Simplicity matters — Snowflake’s intuitive interface and developer-friendly tools helped the company’s team take ownership quickly.
• Modern architectures enable rapid evolution — the medallion architecture approach gave structure and agility to the company’s growing data ecosystem.

These lessons reaffirmed SPR’s belief that modern data transformation isn’t just about tools; it’s about enablement, and building platforms that teams can trust, understand, and expand.