Self-Service Business Intelligence On Hadoop


Dave Mariani, Founder and CEO, AtScale

AtScale announced on May 16, 2016, that it has raised $11 million in a Series B round, bringing its total funding to date to $20 million.  The round was led Comcast Ventures, with participation from UMC, AME Ventures, Storm Ventures and XSeed Capital.

AtScale was founded a few years ago to link the business intelligence users—and their favorite data visualization tools—to the Hadoop data platform used by their enterprises. “Hadoop was invented at Yahoo,” says founder and CEO Dave Mariani, “because [co-founders] Yang and Filo had this mantra of ‘we don’t throw data away because you never know when you might need it.’”

At Yahoo, Mariani was responsible for managing engineering for audience and advertising data pipelines and analytics ingesting 20 terabytes of data daily across multiple 4,000 node Hadoop clusters. What became the big data mantra of collect all the data, ask questions later, led to the creation of Hadoop, a new type of “data warehouse” where structure was added to the data, ”only when you had a question to ask,” says Mariani. With his background in business intelligence, it was “a revolutionary idea” for him.

That led to another revolutionary idea which was to provide a seamless link—“one data service”—between the business users and the terabytes of data on Hadoop, regardless of which BI tool they were using (at Yahoo, as is typically the case, there were at least 4 different  BI in tools in use at the time).

“The old definition of a BI platform which comes with its own built-in visualization interface is just that—an old definition,” says Mariani. “Our thesis is that consumers want choice and that they want to plug in the visualization tool that works best for the job at hand. So [at AtScale] we explicitly did not built our own visualization layer.”

When he raised Series A funds early in 2015, he got a lot of objections from VCs who thought AtScale should be “full-stack” like other BI tools. Instead, they work with data visualization providers (e.g., Tableau) one the one hand and the Hadoop distributors (e.g., Cloudera) on the other hand to provide the missing link. They do this without forcing the business user to learn new tricks. “The real Trojan horse for us is that we natively support Excel, we don’t require any driver to be installed,” says Mariani. “If you have Excel on your desktop, you are now an AtScale user.”

It turned out to be a convincing value proposition for companies that have tried to capitalize on the promise of big data bandwagon by investing in Hadoop, only to discover the challenges in getting their business intelligence user connecting to the newly collected terabytes of unstructured data. Within a few quarters, AtScale saw a five-fold revenue growth and new customers such as Aetna, Home Depot, Bloomberg, Groupon and Yellow Pages.

So Mariani and team found it easier to raise additional funds this year, in less-than-ideal market conditions, than it was to raise funds last year, when market conditions were much better. “Most big data vendors try to convince the market that because Hadoop is different you need a different platform to analyze it,” says Mariani. “We say that how people want to access the data hasn’t changed, only the data has changed.”

Originally published on