BeyondCore raises $9M to automate data analysis

BeyondCore, a San Mateo, Calif.-based company whose software automatically analyzes data sets to identify the strongest correlations between variables, has raised a $9 million series A round of venture capital led by Menlo Ventures. The funding comes just more than a year after BeyondCore officially launched, and underscores a shift in analytics — in some cases — away from human intuition and toward computing logic.

“The reason we play 20 questions, or rather 20,000 questions, with our data is because computers can only answer very specific types of questions,” BeyondCore Founder and CEO Arijit Sengupta explained during a recent interview. So, he added, rather than humans asking whatever broad questions they want and then drilling down, they should just let computers ask lots and lots of specific questions for them.

The company is working on the 3.0 version of its product, which is delivered as a cloud service, but save for some new features the experience remains largely the same. Users upload data, select the variable they want to analyze, and BeyondCore analyzes everything to highlight the factors that affect it most. The company’s most notable feature might be an animated presentation mode that shows users the strongest statistical connections while an avatar explains the results.

A screenshot of the Analyst Overview.
A screenshot of the Analyst Overview.

Although they’re quite different technologically and probably in terms of ideal use case, BeyondCore is similar in theory to other startups such as (Structure Data Editor’s Choice award winner) Ayasdi and Emcien that also highlight connections across complex data sets without human interaction. The idea isn’t to eliminate humans from data analysis, but just to save them the trouble of having to think of the right questions to ask.

Sengupta explains the process as “hunting for signal.” Once the software identifies something potentially significant, analysts can dig in and clean the underlying data or manually explore it on their own.¬†Data scientists can even examine the statistics and export the code that BeyondCore generates into the R statistical analysis package.

BeyondCore has at least a handful of big-name users already, including McKinsey & Co., United Health Group and Sears.

Sengupta’s vision is that we’ll some day think it quaint that only highly trained analysts can run advanced analytics. “Nobody thinks about asking a potential business candidate ‘Do you know how to use Word or do you know how to use Excel?'” he said. “It’s a given that they do.” It could be a tough sell into certain organizations, but Sengupta thinks software like BeyondCore will make that the case for data analysis skills, as well.

You can hear more about BeyondCore in this presentation Sengupta gave at Structure Data 2013 as part of our Big Ideas program.

Feature image courtesy of Shutterstock user Sergey Nivens.