The Revolution Analytics Blog
I recently received my monthly email from David Smith of Revolution Analytics highlighting the June articles from his wonderful blog on the latest developments with the R Project for Statistical Computing. RA's business charter is to commercialize open source R, making the platform suitable for serious business applications. I'd encourage anyone interested in statistics to take a look at David's site, even if you're not an R aficionado. Among this month's goodies are intriguing graphical analyses of competitive Nordic skiing, the recently-completed World Cup and the BP Gulf oil spill debacle.
My latest RA blog read focused on two posted interviews with new Revolution Analytics CEO Norman Nie; the first by Fox Business Network, the second by Internet Evolution Radio. I was wary of the Fox interview, as I am with anything from Fox – and it turns out for good reason. The 5-minute clip with an unprepared moderator tells us pretty much nothing about RA. Indeed, Nie had to correct one interviewer mistake on the current statistical market leader and veer off on his own to mention open source – the driver of the RA business model. Why am I not surprised?
The IER interview lands much closer to the mark. The moderator lets Nie drive from his experience with statistical analysis and enthusiasm for open source and the R language. Nie also draws on his past as a co-founder of SPSS, discussing the common challenge of building a productive GUI to shield users from a hard-to-learn language. The need to surmount R's limitations for business processing is front and center. Nie confided that Revolution Enterprise will embellish R to exploit multiple cpus and introduce file-based processing to counter current memory limitations. And the academic world will enjoy these enhancements free of charge.
The Fox fumble notwithstanding, the R/Revolution Analytics story is a lot richer – and more interesting – than two guys divining a disruptive new technology for a class project, then dropping out of school to nurture their creation, quickly becoming the rich darlings of the venture capital world. Indeed, my take on the background for the R/Revolution business “proposition” is rather complicated and includes the following points:
- The data deluge has created the mandate for companies to learn and profit from their information assets. Many successful firms like Amazon, Google and Harrah's have adopted a “science of business” rigor to running their operations, using the experimental method and statistical analysis as strategic guides. Business intelligence, which organizes and analyzes data supporting this science, is consequently one of the fastest growing IT fields.
- Time and again, analytics beats the experts in predictive accuracy. So increasingly, analytics are replacing experts as primary inputs for decision-making in many areas of commerce.
- Within BI, the sub-specialties of statistical analysis/predictive modeling/data mining are currently at the top of the hot list. Traditional BI looks backward at performance while predictive models are forward-facing.
- Commercial statistical software has been with us a long time. SAS is the market gorilla while #2 SPSS was recently acquired by IBM, the giant tech bellwether that's bet big on analytics with its Smarter Planet strategy. SPSS is 40 years old; SAS is 35. Both of these popular platforms revolve on old languages that unfortunately look much like they did in the 70s. What was attractive then is severely dated now.
- A modern statistical language/platform, S, was created in the 70s and 80s by the same innovators at Bell Labs that gave the world Unix and C. S was later commercialized as S+ and enjoyed a modicum of success among statistical cognoscenti.
- A little over 10 years ago, a couple of academics created a freely-available, open source dialect of S, naming it R in honor of the originators. The functionality, grammar and syntax of R are almost identical to S. R has taken the academic world by storm and is now lingua franca of statistical computing at top statistics, social science, mathematical science, financial science and biological science programs around the world. R has an open and extensible architecture that's readily exploited by world-wide developers who enthusiastically program the latest statistical/mathematical techniques and make them available for free to the R community. Like SAS students of 30 years ago, these vanguard R developers wish to continue with their platform of choice as they migrate to commerce from academia.
- About 3 years ago, a new company, Revolution Computing, was formed to commercialize open source R, hoping to make its adoption more palatable to a business world that's wary of non-commercially-supported software. Recently, RC changed its name to Revolution Analytics and hired veteran academic and statistical software pioneer Norman Nie as CEO.
- Nie is no novice to analytics. He was co-creator of SPSS as a grad student at Stanford in the 60s, and served as its CEO for years, ultimately helping to shepherd the sale of the company to IBM. With over 40 years invested in statistical software, Nie brings a wealth of experience to his role at RA. He certainly doesn't need the money for retirement!
- The challenge for Nie and RA is to enhance the free, open source R platform to make it attractive for paying customers, taking on a host of new and existing vendors, including the entrenched SAS and SPSS, along the way. One immediate hurdle for RA is the need to enhance R's capacity to meet the processing demands of ever-expanding data.
- RA must simultaneously appease and cultivate the R community of developers that does most of the platform work gratis. Unlike other commercial open source software companies that essentially “own” the major OS programmers, RA must herd the proverbial cats of R's free-thinking world-wide development community, many of whom are ambivalent or worse about commercial open source.
What, in sum, must Revolution Analytics do to be successful? 1) Evangelize, as David Smith does so well, on the science of business and the power of analytics/R 2) Convince the marketplace that R's the modern analytics answer to support the science of business, and 3) Introduce an enhanced R platform that adds discernible value over what's currently available for free – all the while playing nice with a spirited pro bono labor community. Piece of cake!