Ramp provides unparalleled revenue and market insights to customers. As it grew, the Ramp team spent more and more time focused on infrastructure rather than its unique data modeling approaches. That all changed when Ramp switched to a scalable, simple data cloud with Snowflake.
Ramp is a cohort-based customer and revenue prediction platform that helps consumer brands predict performance with startling accuracy. Since 2018, Ramp has helped finance teams at organizations to better manage resources, make more informed decisions, and automate revenue and user forecasting.
For Ramp’s customers, the true value of its service comes from its unique, powerful data models. Yet, as the prediction platform provider grew, its teams were spending less and less time on core business, and spending more time dealing with infrastructure growing pains. By running data warehouse and data engineering workloads on Snowflake’s Data Cloud Ramp improves performance and user experience, while delivering powerful insights to customers quickly.
How do you scale seamlessly, without worrying about keeping the lights on?
Although Ramp serves a roster of large clients, it has kept its team small. The challenge has been scaling to meet the needs of the largest venture capital firms and consumer organizations, while leaving room for its own business to grow.
But when you employ a few dozen people, you can’t devote entire resources to managing infrastructure. “Data modeling is where we provide value to our clients,” explained Francois du Toit, Co-Founder and CTO of Ramp. ”So we wanted to concentrate on that instead of our infrastructure.”
Unfortunately, du Toit and his team had no choice but to investigate issues as Ramp scaled up. Generating about 3 billion records per day, Ramp’s datasets grew quickly, and the team began seeing runtime problems with its Postgres system. Our developers were spending a lot of time trying to figure out where the issues were,” du Toit said. “Even a simple delete operation took up to ten minutes to complete.”
Scalability and new features with Snowflake’s Data Cloud
Ramp began assessing new tools to address these challenges. “A VC suggested that we should look at Snowflake as part of our assessment,” du Toit recalled. “Snowflake quickly beat the other vendors we were looking at in terms of price and features.”
After four months of testing, du Toit and his team had moved one of its databases to Snowflake for bulk data processing. They kept the other half in Postgres as part of their customer-facing application. Ramp fetches and delivers data into S3 buckets as well as uses dbt to transform data at each stage.
“At every point, Snowflake has been an absolute joy to work with,” du Toit enthused. “The documentation was clear, and that helped us get started faster than I thought we’d be able to. And for the few small issues we had with configuration or SQL differences, I was able to bounce ideas around with Snowflake’s startup team. They’ve been really helpful in supporting and shaping our platform.”
Ramp transforms performance and overcomes data processing challenges
With a new system that can scale with its growing datasets, du Toit and his team have been able to transform performance. “We saw a 92-97% performance increase compared to Postgres,” du Toit explained. “Batch processing operations that used to take hours now take just eight minutes.”
Beyond time savings, this performance increase has helped Ramp streamline operations and deliver market insights to customers at scale. “We used to run scenarios at night to avoid capacity issues,” du Toit remembered. “If they failed, you had to wait hours for results, during which time no one else could process anything. Looking back, it felt like a dark age, so I’m glad we were able to solve that challenge with Snowflake.”
Faster analysis expands customer insights and product offerings
The speed, performance, and scalability of Ramp’s new solution doesn’t only let its engineers work faster and more efficiently; it also helps customers get more out of the Ramp platform.
“Up until now, our analytics tools were for internal use. Then we’d share the insights with our customers,” explained du Toit. “But we wanted customers to see us as a platform and solution provider, not a consultancy. In the past, the lack of scale and performance meant we couldn’t offer our tools externally. But now we have Snowflake, we can let customers access insights directly.”
Now, customers can bulk process all kinds of what-if scenario queries across an extensive array of audience data. And those results come back in minutes. “Really, Snowflake has helped us bring the vision of Ramp to life: extensive insights for customers across lots of data—at scale and at speed.”
Technical teams can now make the most of their resources
As a smaller team working with several large customers, Ramp has found success making the most of its resources. Now, the team can do even more with less.
Since moving to the Snowflake platform, Ramp has cut the costs of its data platform by two thirds. “Many of the cloud data platforms seemed efficient, but only for organizations with constant high loads,” explained du Toit. “We tend to have high demand at uncertain points in time with varying peaks, and Snowflake was the most efficient platform for dealing with those workloads.”
Ramp delivers even more insights, data, and value to customers
Ramp already helps enterprises delve deeper into their customer insights and revenue trends to take their businesses further. Now, the company is looking to provide even more value with the latest Snowflake platform features.
“Customers want more access to their data in different formats,” explained du Toit. “So I’m keen to explore how Snowflake’s collaboration features, warehouse isolation, and row level access can help us build a scalable system where our customers can access Snowflake directly.”
In the meantime, Ramp is poised to continue improving its core offering with Snowflake’s Data Cloud. “Our data scientists have seen the success the platform team has had with Snowflake and are now using it themselves,” said du Toit. “Snowflake is the single source of truth they need to devise new types of insights and models based on reliable data.”
“Data is mission critical to our business. As we continue unlocking the full value of Snowflake, we’re becoming even more efficient in how we use data to deliver the best service to our customers.”