Sisu Nabs $62M to Grow Data Analytics Biz

Sisu Data, a spin-off of a Stanford DAWN project that’s bringing AI to bear on the “combinatorial explosion” of deciding which data variables to track, today announced that’s raised $62 million in a Series C round of investment. The company also is rolling out a new dashboard and new data exploration capabilities in its offering.

As Sisu Data CEO and founder Peter Bailis explained to Datanami last year, Sisu is based on a research program that he led as an assistant professor at Stanford DAWN. (Bailis, who was the founder of the DAWN project, left Stanford in 2020.)

“What we found was, in many cases, the challenge went from not just understanding what’s going on – any BI tool or environment can do that,” Bailis told us. “When you have this super wide data [with] all of these different features and columns and so on, essentially understanding why the metrics are changing” is the challenging part.

If you knew exactly what SQL query to run (not to mention perfect data), then it would be easy to figure out the root cause of why a given metric is going up or down, according to Bailis. But without precognitive abilities, which are in short supply in this world, analysts are left to “brute force” this type of analytic activity by generating lots of hypothesis, turning them into questions, and fine-tuning their SQL queries until they get an answer.

That insight forms the basis for Sisu’s offering. According to Bailis, a Datanami 2021 Person to Watch, Sisu essentially turns the problem around. Instead of waiting for a metric in a dashboard to change and then tasking an analyst to figure out why it’s changing, Sisu uses machine learning to continually monitor the entire data space and automatically bubble up the interesting bits.

“You declare the metric. You tell us the attributes. We’ll go and do the slicing,” Bailis said in the 2020 interview. “It’s essentially geared at running very large hypothesis tests, statistical tests, to understand what variables are interesting and important, and then [what] are ways in which you can transform those variables.”

Sisu aims to answer “why”

Think of an OLAP cube with the ability to slice and dice data across multiple variables, but without the need to actually materialize the data (because that would be too slow). “Because we’re doing a lot of these irregular group-bys, we basically have a columnar MPP primarily memory-based parallel dataflow engine, which basically does smart data encoding and a bunch of …parallelization and so on, which makes us from a hardware perspective, also fast,” Bailis told us last year.

In the past year, Sisu has nearly doubled its headcount and more than tripled its revenue. It has customers with names like MasterCard, Autodesk, Samsung, Upwork, Wayfair, Equinox, Udacity, and Gusto.

Now looking to scale up its go-to-market with the $62 million Series C funding, which was led by Green Bay Ventures and with participation from a16z, NEA, and new investor Geodesic Ventures. The round brings the San Francisco, California-based company’s total venture haul to $128.7 million.

“Sisu is creating a new way for organizations to not only analyze their data, but actually use it to make the best decisions in improving the operations, profitability, and success of their business,” Green Bay Ventures Co-Managing Director Anthony Schiller said in a press release.

Sisu today also launched explorations and dashboards, two new capabilities designed to allow businesses to better visualize and understand their data.

“Explorations allow users to quickly and easily dig into, pivot, and visualize their metrics, without requiring code,” Bailis said in a blog post today. “Dashboards allow users to collect explorations, track changes to their metrics, and to share them with others.”

Sisu noticed that users were still using pivot tables to explore their data, and then move into Sisu to do more fine-grained analysis. That will change with dashboards, which puts ML-powered diagnostics at user’s fingertips, “so you can automatically reveal the key drivers behind changing metrics with a click,” writes Sisu’s Davide Russo in a separate blog post today.

Related Items:

Peering Into the Crystal Ball of Advanced Analytics

VC Ben Horowitz Dishes on Hadoop, AI, and Data Culture

Sisu Seeks to Answer Why

Editor’s note: This story was corrected. Peter Bailis was an assistant professor at Stanford, not an associate professor. Datanami regrets the error.