Microsoft increased its foothold in the data science community last winter by acquiring Revolution Analytics, a major provider of software and services based on the open-source R project for computational statistics. The deal is expected to bring R capabilities to the Microsoft suite of products and facilitate the adoption of R-based solutions in the enterprise environment. Revolution Analytics will continue to support and develop non-Windows versions of its products and continue to contribute to the open-source community.
Founded in 2007 as a spin-off of Yale University, Revolution Analytics is today a significant pillar of the R ecosystem. The company has approximately 100 employees, supports numerous user groups across the world, and work with more than 200 commercial customers, including Jonhson & Jonhson, Merck, Pfizer, and Citigroup. It also has a strong presence in the R open-source community with many active projects such as ParallelR, RHadoop and its flagship distribution: Revolution R Open (RRO). According to a 2013 interview with CEO Dave Rich, Revolution Analytics's main competitors include SAS and SPSS (acquired by IBM), as well as smaller companies such as KXEN (acquired by SAP).
From a technical perspective, RRO is a downstream distribution of GNU R that provides two main enhancements: performance and reproducibility. The gain of performance, particularly noticeable on multi-core hardware, comes with the use of Intel Math Kernel Libraries (MKL) in lieu of the standard BLAS/LAPACK libraries. While GNU R can easily be recompiled and linked to MKL by a savvy user, RRO has the advantage of providing this feature out of the box. The second enhancement is, however, more unique. By maintaining its own mirror of the CRAN repository, Revolution Analytics can ensure that each release of RRO uses a fixed set of package versions and, therefore, provide consistent, reproducible results over time.
It is worth noting that Revolution Analytics neither owns nor created R. R is a free and open-source project administered by the non-profit R Foundation and initiated in 1993 by Ross Ihaka and Robert Gentleman, two statisticians from the University of Auckland. It is one of the two modern implementations of S, a statistical computing language designed by Bell Laboratories in the '70s. The other implementation, S-PLUS, is now maintained and sold by TIBCO under proprietary terms.
After more than two decades of existence, R is used today by about two million users, both in academia and industry. Google, Ford, Microsoft, Facebook, Bank of America, Uber, The New York Times, as well as several US federal agencies, are notable users of R. Nature reported in 2013 that nearly one percent of scholarly articles indexed by Elsevier's Scopus, one of the largest bibliographic database of peer-reviewed literature, cites R or one of its packages.
As pointed out by Helena Schwenk, IT industry analyst at MWD Advisors, many of the other software heavyweights already have R strategies, including Oracle, IBM, Teradata, Pivotal, SAP HANA, and SAS. “Microsoft is also working in a busy marketplace, [...] articulating the value of its approach and differentiating itself amongst these players must, therefore, remain a priority for Microsoft going forward,” she says.
Research firm IDC forecasts that the market for advanced and predictive analytics software will grow from $2.6 billion in 2015 to $3.5 billion by 2018. Revolution Analytics raised $38.7 million in venture capital and has revenues estimated at $12 million in 2013 by Wikibon.
