Big data is a big deal. Did you know Walmart handles one million customer transactions per hour that is imported into a database that is 167 times more than the information contained in all the books in the US Library of Congress? So staggering is the amount of data we produce, that new software to handle the explosion continues to be built so the world isn't drowning in a big nonsensical pile of it.
Fortunately, some of these software solutions have been in the works for more than a decade. That's where Ayasdi comes in. Initially formed within Stanford Topology Lab and funded by the Defense Advanced Research Projects Agency (Darpa) for about $750,000 and IARPA (intelligence research arm of the DoD) for $500,000, Ayasdi has emerged as a big data contender for the pharameutical, oil and gas, and financial services industries, as well as the government.
Palo Alto, Calif-based Ayasdi, which was incorporated in 2008, announced Wednesday morning that it's raised $10.25 million in a Series A round of funding from Khosla Ventures and Floodgate to solve complex problems and discover new insights from scores of data. Floodgate had invested in the seed round of $2 million in 2010.
"Our mission is to help people extract information from data automatically," said Gurjeet Singh, CEO and co-founder of Ayasdi, in an interview. "If you think about how people work with data today, they do that by asking questions. The only way we really learn from data is accidentally hitting on the right question," he explained. Twenty years ago, that was OK, when teams of analysts were brought on to come up with questions, Singh explained. "But the scale of data we're dealing with today is so large that asking questions is flawed."
So rather than be top-down thinkers, people need to start being open-minded. "Start your process without any bias," said Jeff Yoshimura, VP of marketing at Ayasdi. "Then let the software flag anomalies and things you didn't know, and build a model around the answers," he said.
The approach to discovering these insights seems quite unique.
"We're using machine-learning methods and anthropology to discover insights from data," said Singh, adding that "The key insight is that data has shape and topology helps us to discover those insights [by seeing the different shapes]."
While the topology provides a visual aid, in many ways the mission to extract insight from raw data is similar to that of other software companies, such as Palantir, Tableu andTibco's Spotfire, and to startups that have emerged in the big data space. But unlike other recently-funded companies, Ayasdi has set its sights on multiple industries, rather than just one vertical.
Guavus recently announced a $30 million round from Investor Growth Capital (IGC), QuestMark Partners, Artiman Ventures, Sofinnova Ventures and Intel Capital to be a big data platform for the communications industry. Health Catalyst, another big data company that recently raised $33 million focuses on servicing hospitals and the health care industry. Platfora is yet another company that recently raised a round of $20 million for its business analytics solution.
Ayasdi's customers range from Merck, UCSF, Second Genome, Darpa and the FDA. But while the client base runs the gamut, a lot of customers are pharmaceutical companies, said Singh.
"If you think about today, the pharma industry is primarily around data," he said. "The last 10 years, the industry has transformed into data." The analysis of the data is where drug discoveries and new diseases come from. Unfortunately, the data can be analyzed for years without any new discoveries.
In one case study, resarchers had been sitting on data they'd collected for 15 years, according to an Ayasdi study that the company says has been scientifically validated and published. While scientists had been studying the data for many years, the Ayasdi software discovered a certain group of breast cancer patients that had survived without having to undergo certain treatment. Ayasdi said that it discovered this cancer insight in five minutes with one data scientist.
One data scientist vs hundreds of thousands can make a big difference in expense.
"Bringing one drug to market takes about 10 years," said Singh. "It costs about $1 billion and takes about 500 people working on it for a decade." Then after you have sunk in this many resources, your chance of success is about 10% as only one in 10 drugs make it to market, he said.
"The reason pharma companies like us is because we can dramatically reduce time and resources," said Singh, estimating that Ayasdi can reduce the time to market down to two to three years.
It's ambitious. But if any team can do it, this team of Stanford University mathematicians seem well-equipped to make it happen.
Singh founded Ayasdi with co-founders Gunnar Carlson, a professor of Mathematics at Stanford University and Harlan Sexton, who has a PhD in Mathematics at Stanford University.
(Image source: masterfile.com)