Analytics 'R' Us
By heading up Revolution Analytics, Norman Nie is putting himself in competition with SPSS, a company he himself started decades ago. Revolution Analytics -- previously known as "Revolution Computing" -- aims to take the open source R language, which specializes in statistical computing, into the business world and make it an analytics powerhouse.
06/11/10 5:00 AM PT
A growing recognition of the business benefits predictive analysis provides is positioning newcomer Revolution Analytics into a key role to help adopters of the R programing language migrate from legacy offerings. Until a recent funding infusion and a refocus of marketing goals, the startup did business under the name "Revolution Computing."
Revolution Analytics, led by predictive-analytics expert Norman Nie, recently released an alpha version of its commercial product based on the open source language known as "R." The company markets enhanced versions called "Revolution R" and "Revolution R Enterprise."
The new company, which launched on May 4, is headed by CEO Norman Nie. In fact, some 40 years ago, Nie founded SPSS (Statistical Package for the Social Sciences). That company's mission was to spearhead the widespread use of data in decision-making. Nie was one of three people who developed a software system based on R and the idea of using statistics to turn raw data into information essential to the decision-making process.
In his new corporate role, Nie is aiming to disrupt the market he helped create and take on his old company head-to-head using the technical and cost-advantages of open source R.
When funding partners approached him about taking over the new company and further developing R, Nie's first reaction was, "No way. I'm not doing this again," he told Linuxinsider. Then he gave the plan more thought.
"I started thinking about what it would take. I began to understand the opportunities and the limits of the GPL (General Public License). All of a sudden the picture occurred to me of an unbelievable software analytics company that could handle the biggest data files and would have the speed to overtake the legacy companies," he said.
The R Factor
R is a language and environment for statistical computing and graphics. The open source GNU project software project runs on a wide variety of Unix/Linux platforms, Windows and the MacOS.
R provides a wide variety of statistical modeling and graphical techniques. It is highly extensible. A key strength is its ability to produce well-designed, publication-quality plots, including mathematical symbols and formulas.
R is the choice of statisticians, Nie explained.
"There are no statistical expressions that can not be written in R," he said.
Revolution Analytics is working to be the rallying point for the commercial development of R. Nie sees a good chance for the company to become for R in the predictive analysis space what RedHat is to the Linux community.
To that end, the company is spending money to focus the open source community around its website, he said. The commercial product will be free to academics under a dual license.
"We want to take advantage of the use of R and put our IP (intellectual property) on it for commercial adoption," he said.
In essence, Revolution Analytics is putting its money on an academic infrastructure of R users to earn a return from the commercial users. The R language has its genesis among university researchers and graduate students.
That once-academic fervor for statistical analysis is moving into the business world. A new generation of college-trained statisticians will bring their programming skills in R with them. Why? Because it is so popular in Academia with researchers.
"So that has the potential to spur its adoption. Its use in academia can be an appealing factor," David White, senior researcher analyst for the Aberdeen Group, told LinuxInsider.
R Gets A+
The company has made some significant improvements in the community-developed version, White said. For example, it is now much easier to use analytics. The improvements address a key area where users lack key skills. They don't need a PhD to use it.
That sort of improvement comes on top of the existing benefits of using the R language. That is the cutting edge advantage the company brings to market.
"Nothing specifically like R has ever been developed for research analysis. It is the only language ever developed by statisticians," Jeff Erhardt, COO of Revolution Analytics, told Linux Insider.
Other programming languages such as C and Python do not have the same level of productivity. Over the last five years, use of R has literally taken over the academic world, he said.
Known as "Revolution Computing" when it opened for business in 2007, the company received $9 million in Series B funding from North Bridge Venture Partners and Intel Capital in October 2009. That's when Nie took over as CEO.
Despite the company's unique position of being in the forefront of the R evolution, the investors wanted a change in direction to better confront the marketplace. The predictive analysis space was dominated by only a few companies, noted Nie.
The launch of the alpha version of the enterprise version last month came with the announcement of a corporate name change.
Prior to the new round of financing and Norman Nie joining, Revolution was focused on commercializing R, but not specifically for predictive analytics use. Their primary efforts as Revolution Computing were focused on bringing parallel computing technology to R, as opposed to a focus on the broad analytics market, COO Erhardt explained.
"A change is going on in the industry. Previously, proprietary firms such as SAS and SPSS dominated the field. Revolution Computing is taking a new tack with its open source software development," Dave Stodder, principal analyst for Perceptive Information Strategies, told LinuxInsider about the company prior to its name change.
The recent corporate name upgrade may mirror a maturing of the concept behind predictive analytics. The software field for predictive analytics is wide open as it is relatively new, noted the Aberdeen Group's White.
What the company is doing is driving down the cost of using analytical tools. This allows software developers to include analytics in their applications, added White.
"This new field of predictive analytics is really just a variant of what was statistical modeling. R just sits on that wave. With the kind of corporate backing we're providing, I think it is an incredible opportunity. I see great disruptive innovations," Nie said of his company's viability for success.
Nie faces a major sticking point in pushing the commercial version of R against legacy options. For many old-school businesses, open source software is banned.
"A challenge to the adoption of R is its open source roots. A lot of enterprises shy away from open source software because of its potential for not being supported," said White.
For example, a representative from a large insurance firm looking at R told him that basically hell would freeze over before open source ever was used in that company.
Then and Now
Another obstacle that Nie faces is the different marketing strategy he faces. When he first marketed the R concept, no incumbents existed. Today he must contend with SAP and his old company, SPSS, which is now a part of IBM.
"So part of our focus is educating the market about the Revolutionary process. Our pitch is to show the scalability and usability compared to these other solutions. All of the innovation in predictive analysis is coming from open source," he said.
Core Goals for Development
Navigating the newly renamed company through a growing predictive analytics space is much like facing a perfect storm, suggested Nie.
You have all these graduates trained in using R, plus a growing demand in business for statistical analysis, coupled with legacy solutions that are very costly.
The startup may already have a head start. To push forward, Nie is following a three-part plan.
One, he has to scale R so it can be used in enterprise applications. Two, he has to build in greater usability. Three, he needs to develop migration tools to enable people to move all their legacy products into R.