Trends: SAS, SPSS Push into BI Stronger than Ever

Two mainframe number-crunching mainstays make their bids for broad BI dominance

If you do any data mining on your mainframe hardware, you’re probably familiar with SAS Institute Inc. and SPSS Inc. Both companies have enviable pedigrees in the data-mining and statistical-analysis spaces, and both launched their bread-and-butter businesses on the mainframe. To a large extent, both continue to enjoy thriving mainframe software businesses, too.

But both companies long ago shifted their sights beyond the mainframe, successively conquering the client-server and, lately, the Web-client worlds.

More recently, both companies have successfully pushed out of their core market segments and into the broader business intelligence (BI) space—with strong results. According to International Data Corp.’s 2005 BI market share study, SAS is the number two overall BI vendor (trailing only Business Objects)—with growth that outpaces the market as a whole—while SPSS last year cracked the Top 10 and posted a double-digit growth rate.

Both companies expect untrammeled growth going forward, too. Thanks in part to the post-September 11 landscape and in part owing to the age of compliance, which has upped the ante for insight into one’s vital innards, SAS and SPSS officials believe they can leverage their demonstrable strengths in data mining and analysis to gain BI dominance.

That confidence probably isn’t misplaced. According to IDC analyst Dan Vesset, SAS’ and SPSS’ analytic chops have functioned as engines for broader BI growth, and will almost certainly do so for some time to come. “SAS [has found] success in specialty analytic applications that take advantage of its advanced analytics tools. Examples include applications for various types of forecasting, optimization, and descriptive and predictive analytics,” Vesset points out.

“SPSS is the second-largest advanced analytics vendor. Its focus on predictive analytics paid off in 2004 and 2005 after several years of lower-than-market growth rates. In many cases, SPSS has also been able to cross-sell its query, reporting, and analysis tools into its base of advanced analytics customers.”

More to the point, Vesset argues, neither the big BI pure plays nor the packaged applications vendors (e.g., SAP AG, Oracle Corp., or Microsoft Corp.) have yet delivered analytic technology that is the equal of, not to mention superior to, that of SAS and, to a lesser extent, SPSS. In the short term, in fact, Vesset doesn’t envision any compelling challenge to SAS’ dominance of the advanced analytics market, and he expects that company to continue to grow its query, reporting, and analysis practices at above-average rates.

One potential bulwark to further growth is usability. SAS and SPSS were long the domain of pointy-headed statisticians, and—even though both companies have struggled to recast their analytic tools as user-friendly plays—there’s the lingering perception that both toolsets present daunting adoption curves.

To some extent, says SPSS user King Douglas, a senior analyst with American Airlines, those perceptions were valid. In SPSS’ case, Douglas says, usability was a problem—at least when SPSS first transitioned from its mainframe-friendly plain-text output system to a new GUI-based (and client-server-oriented) implementation.

“I went from the mainframe, [and there were] no problems, [but] after a few iterations, [SPSS] went to a new output system with Version 7, and since it was a brand new output system, it was very, very awkward; [it] was causing all kinds of trouble. The old output system basically was plain text. So once they went to this new system—all graphics—it made things harder to do than before,” he notes. “This went on for a few iterations, then they snapped out of it. From version 13, 14, and 15 they’ve really gone back and made a star of SPSS.”

Douglas helps head up AA’s in-house market research arm, which taps tools from both SAS and SPSS to support its activities—although AA is most dependent on SPSS. For the most part, he says, going from the mainframe to client-server SPSS was a fairly painless transition. “Going from the mainframe to SPSS on the desktop, for me, was very easy, because you have an interface [in which] you just type in the commands. If you have a variable (say, income) and I want to see the distribution of this variable in this data set, I just type ‘fre’ or ‘freq’ [frequency] and it gives it to me,” he says.

“It was easy to move over, but then [in the GUI environment] they have this interface where you can point and click, drag and drop … and you can construct any analysis you want, click the paste button, then save it in a new window, and once you save that, you can continue to modify that and save it, or rerun it at a later date. So it gives you the ability to accumulate completely debugged programs that run perfectly.”

This is one reason Douglas believes that SPSS’ transition from number-crunching specialist to BI general practitioner—and would-be power player—could take place in a similarly effortless fashion.

Exploiting Their Expertise

For one thing, he says, SPSS has demonstrable number-crunching expertise, and in the company’s new SPSS 15—which debuted last week—this expertise is yoked to a broad array of data management tools. In a data warehousing space in which one of the prime problems is ensuring the quality and consistency of data, SPSS’ analytics seem tailor-made, he argues.

“SPSS is almost infinitely expandable, if you’ve got a computer powerful enough to handle it, and it’s got great algorithms for handling and managing large sizes of data,” he comments. “It’s a great data warehousing tool. It’s largely extremely good for managing large amounts of data, creating new variables, cleaning up the data, inspecting the data, getting it to cough up hidden information, and so on.”

SAS officials, too, take a similar tack. The company released its first-ever all-in-one BI suite more than two years ago, and SAS is currently prepping a major new version of that product. Since then, SAS has expanded more concretely into the performance management space, too.

The idea, officials say, is to encourage existing users of SAS’ data mining and statistical analysis tools, along with SAS’ market-leading ETL software, to take the plunge into SAS BI and PM. Or—to put it another way—if you’re already using SAS to do your number-crunching and data extracting in the back-end, why not use SAS on your client desktops, too?

That, in a nutshell, is the approach that both SAS and its less prominent competitor are taking. “It’s a natural extension of their investment in SAS,” said Mark Digman, director of industry and solution product marketing with SAS, in a May interview. “We want to shine attention on the fact that to do this, you’ve got to have the analytics piece … [that includes] optimization, forecasting, data mining, trend analysis, and not just query and reporting,” he added.

These are things SAS has been doing more or less since its inception, Digman contends. “When I look at performance management or performance improvement, we’ve been helping companies do that for 30 years now. What happened is there have been lots of different names around lots of different initiatives and market trends, and performance management is the latest one. But when you look at our tools, we’ve always enabled that, but it’s been for specific areas, specific needs,” he argues.

“We’ve been doing these [data mining and statistical analysis] for year now, and over the last few years, because there’s been a BI push in the market, we’ve had to make it clear that we do BI, too. Now we’re seeing more of a performance management [push], now we’re going to make a much bigger push from us.”