Challenge

Surface needed to build an application could help them assess and track risk and returns for each company and portfolio over time using ongoing financial data, scores, and comments. A key requirement was to support sub-second latency from the moment data is entered or sourced into a Neon Database to the target analytics being displayed. This included running complex queries with several levels of nested joins.

Solution

Gyan Kapur, a managing partner Surface, considered two options for high performance analytics databases as a service - Clickhouse and Motherduck. While he believed both could work, he chose Motherduck for its combination of sub-second performance even with deeply nested joins, its simplicity, and its low cost.

For data integration, after looking at several options including Airbyte, Gyan chose Estuary for its ability to support streaming CDC from Neon to Motherduck, its ease of use at the lowest cost.

Benefits

• Fast implementation - The initial implementation from start to production took Gyan 2 days, including initial learning time. Changes now go much faster,
• Sub-second latency - The combined end-to-end latency of Estuary CDC and loading, along with Motherduck complex query performance, is less than 1 second. There is no waiting for results.
• 10x lower cost - Estuary and Motherduck combined ended up being 1/10th the cost of the alternatives.

“When it comes to analytics, I don’t like to wait. Estuary and Motherduck allowed me to integrate and create the right queries really fast so that I could focus on investing. Now, when I enter in new data, I can immediately see the impact of changes on investments and portfolios.”

• Gyan Kapor, Co-Managing Partner, Surface Ventures

Now that Estuary and Motherduck are in production, they just work, really fast. This gives Gyan extra time to add even more analytics, and more importantly, more time to work with startups.