Stats king SAS surfs the analytics wave
But can number-crunching software grandad stay on the surfboard?
Blog One obvious beneficiary of the rush to embrace enterprise analytics and "Big Data" is the SAS Institute Inc (or just plain SAS – rhymes with "pass"), the granddaddy of statistical analysis software. But it can be argued that SAS doesn’t really need the analytics boom; they’ve been doing quite nicely up until now without it, mining their profitable niche for all it’s worth.
And it’s worth quite a bit. Founded in 1976, the company has shown unbroken revenue growth for 35 years since. They topped $2.43bn in 2010, with 46 per cent coming from the Americas, 42 per cent from EMEA, and 12 per cent from Asia/Pacific.
Since it is a private company, it releases only the financials that it wants to release, so there’s not much known about their profitability. I’d put its margins somewhere between “really good” and “extraordinarily high".
Over the years, SAS has become the standard statistical analysis software package for government and business users alike. Their biggest segment is financial services, which provides a whopping 42 per cent of revenue. Their next biggest market is government at 15 per cent, followed by other industry areas such as life science, retail, and manufacturing, all with single-digit shares.
The company has vastly expanded their offerings beyond their statistical analysis roots. They have a full slate of business intelligence, resource management, and data management tools – all designed to give organisations a better handle on their customers and their business.
SAS has custom pitches and, in some areas, customised products to fit various industries: 16 individual vertical markets in all. Financial services gets the most attention, as befits its "largest segment" status, with many tools designed for both financial analysis and customer analysis.
So SAS is sitting pretty right now, seemingly perfectly positioned for the enterprise analytics wave. But SAS has to ride the wave expertly, or they’re going to be pounded by it. All of the attention to the analytics trend – and the accompanying higher spending – has attracted much more competition to what used to be SAS’s domain.
There are new competitive technologies such as Hadoop, and associated analysis tools including Hive and Pig, that evolved to handle truly massive amounts of data that might not ever make it into a regular database due to the sheer size or untidiness of the data.
These tools don’t have nearly the bells and whistles of SAS and traditional relational databases, but they have a lot of developers adding functionality and sophistication. They’re also free of charge – or so inexpensive compared to the alternatives that they seem free.
SAS also faces renewed competition from established vendors like Oracle, HP, and IBM, all of which are investing heavily to build their Big Data wares. While SAS is the dominant ISV in this space, the systems vendors (and others) are busily buying up everyone else.
HP purchased analytics DB ISV Vertica this year and has already rolled out new software and integrated analytic system bundles. Oracle is pushing ahead on their Exadata and Exalogic appliances. IBM has spent the most, somewhere north of $12bn, buying Cognos, SPSS, and a host of other companies over the past several years. They’ve also put out a number of analytics-infused systems with servers that are optimised for these workloads and come in x86, Unix, or mainframe flavors.
While SAS has competed with the big vendors before, this time it’s different. The big vendors today all see enterprise analytics as the "next big thing" and will push hard to get their fair share of the market – more than their fair share, they hope.
At the same time, SAS faces new threats from the data punk rockers driving new tools and techniques that still fall short of the massive power of SAS, but are good enough for many users – and much, much less expensive. SAS has faced upstart competitors before, of course, but those upstarts have never received the mindshare that they’re getting now.
SAS faces wars on two fronts against very different types of competition. They’re going to have to choose which technologies and partners to embrace or confront very carefully, or they could find themselves at a serious disadvantage down the road. ®
Rhymes with "pass"?
Yes it does - if you speak Merkin, or were born north of Watford Gap.
Just a matter of time till SAS crashes and burns.
Penguin herder got it right.
There are FOSS alternatives that are greatly superior: R and Octave (matlab clone) to name two.
Those enterprises who keep paying for the expensive annual support contracts and getting inferior software for it, are cases of plain inertia not much different than paying millions per year for Oracle databases.
The up and comers; the younger companies like Google, Facebook, Twitter, etc. don't use this expensive, proprietary and restricting stuff. They started from scratch, with the much better free options, which is part of why they are more successful.
This 2.5B behemoth is near its peak. IMHO.
In a word, R
R is free and very powerful. Problem solved.