Overview
Sisu Data made software to investigate why a business metric changed. Instead of asking an analyst to test each possible customer segment by hand, its engine searched the underlying data for groups associated with a change. Stanford computer scientist Peter Bailis founded the San Francisco company in 2018. It raised $128.7 million in venture funding and counted Samsung and Upwork among its customers. In October 2023, Sisu joined Snowflake and said it would discontinue its standalone product. The deal terms were not disclosed. [1][2][3][4]
Founding Story
Bailis had studied large-scale data systems at Berkeley and joined Stanford as an assistant professor. His work on the Stanford DAWN project explored tools that made data and machine learning useful to more people. Sisu grew from that research setting: its commercial problem was that warehouses could store vast amounts of data, while finding an explanation inside them still took repeated manual queries. Bailis left Stanford's tenure track to work on Sisu full time. [1][5][6]
Timeline
- 2018: Bailis founded Sisu in San Francisco. [1][6]
- 2019: Sisu launched its diagnostic analytics platform and announced $52.5 million in funding, including a Series B led by NEA. [2]
- 2021: A redesign brought monitoring and analysis together; a $62 million Series C brought total funding to $128.7 million. [3][7]
- 2023: Sisu announced that it was joining Snowflake and would discontinue its standalone product. An investor and later deal research identify the transaction as an acquisition. [4][8][9]
What They Built
A team could point Sisu at structured business data and define a measure such as revenue or customer retention. Sisu then explored combinations of attributes to find segments associated with a change. For example, a fall in sales might be concentrated in a particular region and device type. The result gave an analyst a lead to examine, rather than proving that the segment caused the fall. [1][7]
This was a technical search problem as well as a presentation problem. Bailis described a parallel, primarily memory-based engine designed to make many irregular group-by calculations quickly. Sisu also worked on the interface: a 2021 redesign brought monitoring and diagnosis closer together, so users could move from noticing a change to investigating it. [7][10]
Market Position
Sisu sat between cloud data warehouses and business intelligence tools. Warehouses stored and processed the data; dashboards showed what happened; Sisu focused on the next question, why a measure moved. An analyst interviewed at its launch saw a large opportunity but questioned whether the startup could expand beyond one diagnostic function. Sisu later moved into monitoring and other workflows. [1][10]
Business Model
Sisu sold enterprise analytics software. Its launch announcement named Samsung and Upwork as customers and said the funding would expand engineering, sales and other teams. The company did not disclose a public pricing schedule in the sources used here. [2]
Traction
In 2019, Samsung was already using the platform. By the 2021 Series C, Sisu named Mastercard, Autodesk, Wayfair, Equinox, Udacity and Gusto among customers. Bailis said revenue had more than tripled in the preceding year; the company did not disclose the revenue base in that account. These are company-reported signals of adoption, not independent measurements of retention or profitability. [1][7]
Outcome and evolution
In October 2023, Sisu said it had decided to join Snowflake and would discontinue the independent product. Snowflake already had a relationship with Sisu: the companies had been listed as partners before the deal. Investor a16z described Sisu as having joined Snowflake, and PitchBook's 2023 deal report listed an acquisition. Neither establishes the price or which specific Sisu capabilities survived inside Snowflake. The closure of the standalone product is clear; its customers' individual migration outcomes are not. [4][8][9][11]
Key Lessons
Sisu found a clear next step after a dashboard: find the groups that deserve investigation. Its experience also shows the challenge of building a durable standalone product around one part of a larger analytics workflow. The public record supports the product's adoption and the Snowflake deal; it does not establish why the company sold or whether it delivered a financial return. [1][4][7]
Sources
- TechTarget: Sisu's data analytics tool
- Sisu launch and funding announcement
- BigDATAwire: Series C and product architecture
- Sisu: joining Snowflake
- NEA: The Next Chapter in Data Analytics
- Peter Bailis curriculum vitae
- BigDATAwire: Sisu's $62 million round
- Andreessen Horowitz: Sisu
- PitchBook: Q4 2023 AI and machine learning report
- TechTarget: Sisu redesigns its analytics platform
- Peter Bailis: Snowflake partner announcement