Most SAS Institute customers historically have worn lab coats and pocket protectors. Now the company wants to attract more customers who dress in business suits.
For almost three decades, privately held SAS has been a top provider of software for extracting and analyzing large amounts of information using advanced statistical-modeling techniques. In the pharmaceutical industry, especially, its software is a de facto standard, used by Merck, Bristol-Myers Squibb, Eli Lilly, Schering-Plough and others for analyzing drug-testing data and performing other analysis. In fact, in 1999 the U.S. Food and Drug Administration specified a SAS data format as its preferred method for drug-trial results that must be included in regulatory submissions.
“There’s probably not a biomedical or pharmaceutical company in the country that doesn’t use SAS,” says Barbara Tardiff, vice president of informatics at Regeneron Pharmaceuticals.
But to grow, SAS is moving beyond its traditional set of scientific computing users and trying to attract business managers who could use its tools to crunch financial or sales data. To win over those business customers, however, the company must bulk up its brand name while massaging a reputation for mind-warpingly complex products.
One measure of how complicated SAS’ software can be: The documentation associated with Base SAS 9—the latest version of the company’s original statistical-analysis software, which includes its own programming language for working with data at a very fine-grained level—comprises an encyclopedic 13,250 pages. “You have to be fairly fluent in statistics to use it,” Tardiff says.
The notion that SAS’ software is usable only by number-crunching eggheads has turned off more than a few prospects. Bill Sanderson, information-technology manager in the state of Utah’s finance division, picked Cognos to provide data reporting for 1,000 employees because “you didn’t need to understand the underlying data. SAS was still more oriented toward being a statistical tool.”
James Goodnight, president and CEO of SAS, derides software from competitors like Cognos and Business Objects as simplistic packages that merely relay queries to databases and pretty-up the results in colorful charts. “It’s the biggest farce in the world that they call themselves business-intelligence tools,” he grumbles. He notes that SAS provides analytic capabilities that less-sophisticated reporting software lacks, such as nonparametric analysis, which is used when the parameters of the variables being analyzed are unknown.
But since its products are seen as specialized tools for statistical experts, SAS doesn’t always show up on the radar of business executives. “Because you don’t have large numbers of business people using SAS, it isn’t something that hits them in the face every day,” says John Hagerty, an AMR Research analyst.
To raise its stature in the business world, SAS in the last two years has focused on industry-specific applications built on its data-analysis technologies, such as software to help banks detect credit-card fraud. The company has reorganized its sales group to line up with 12 vertical industries, including telecommunications and retail, instead of geographies. “If you’re talking to a bank, you’ve got to speak their language and understand their business,” Goodnight says. SAS also says it is increasingly catering to non-technical users with new products like Web Report Studio, a Web-based query and reporting tool designed to present data in business terms.
Still, the company may find itself needing stronger marketing medicine to expand beyond its installed base. About 80% of SAS’ $445 million in revenue for 2004 through mid-May is from existing customers renewing their annual software contracts. “The engine that really drives SAS is renewals,” Goodnight says.
Indeed, many traditional customers rave about SAS. Besides analytical muscle, the appeal for many SAS customers is that its software provides a wide range of capabilities, from data extraction and analysis to reporting and graphing, all in one package. “For everything SAS does, there’s probably another product on the market that does it better, but then you’d have 16 different formats you would need to translate,” says Derek Morgan, a database administrator with the Washington University School of Medicine who has been using SAS software for 17 years. Customers also praise the company’s technical support; Morgan says it is “the absolute best of any software company I’ve ever dealt with.”
Goodnight and company are working hard to win similar kudos from business users. How well they will succeed is uncertain—and even SAS’ powerful analytical software would have trouble turning up a reliable answer.
Headquarters: 100 SAS Campus Drive, Cary, NC 27513
Phone: (919) 677-8000
Ticker: Privately held
URL: www.sas.com
Employees: 9,251
Founded: 1976
Business: Data-analysis tools and applications
Executives: James Goodnight, president, CEO and co-founder; John Sall, executive vice president and co-founder; Keith Collins, senior vice president and chief technology officer.
Products: Base SAS provides data access, manipulation and reporting. Other products include business-intelligence and data-warehousing software, and more than 50 business applications, such as marketing-campaign development.
Market Share: 13% of $8.5 billion business-intelligence and analytics software market in 2003 (AMR Research).
Competitors: Business Objects; Cognos; Hyperion; Informatica; IBM; MicroStrategy; Oracle; SPSS; Teradata.
The Technology
SAS sells more than 100 software packages, and one problem with such diversity has been that different applications weren’t necessarily singing from the same songbook. For example, a report run using Tool A may not have included sales taxes while a report generated by Tool B did, because SAS products maintained separate definitions of key metrics and the data that supported them.
With the latest major upgrade of the architecture that underpins its products, version 9, SAS has tied together all those software components with a common metadata layer, which provides information about the data that resides in various places. That has eliminated much of the redundancy among SAS’ tools and improved consistency of the data, says Forrester Research analyst Keith Gile.
Another technical improvement in SAS 9 is that 20% of SAS’ procedures—groups of instructions that perform specific tasks such as sorting and data summarization—are multithreaded. That means they can tackle multiple calculations at the same time, so an overall job can be processed faster. Just how quickly depends on the type of data and the application processing it, according to the company. SAS says over time, more of its code will become multithreaded, but that certain procedures that don’t involve complex data manipulation, such as printing, would not benefit from the technology.
Reference Checks
Lloyds TSB
Shahram Sharifi
Credit Risk Director
Project: U.K. bank with 15 million retail customers uses SAS statistical applications to analyze credit risk analysis as part of approving 100,000 loans each month.
Toys “R” Us
Rick Muldowney
Dir., Guest Relationship Management
(973) 617-5878
Project: The 1,600-store toy, game and children’s clothing retailer uses SAS tools to figure out which customers to target for direct-marketing campaigns.
Progress Rail Services
David Klementz
CFO
(256) 593-1260
Project: Alabama-based railroad-services company uses SAS software for budgeting and financial forecasting, and to analyze operational metrics from 70 facilities.
Quaker Chemical
Irving Tyler
CIO
Project: Specialty-chemical maker analyzes multiple areas of its business, including customer segments and pricing of raw materials, with SAS software.
Regeneron Pharmaceuticals
Barbara Tardiff
VP, Informatics
(914) 345-7400
Project: Drug developer uses SAS statistical-analysis tools and SAS Drug Development, which allows the company to manage clinical data in an FDA-compliant format.
Washington University School of Medicine
Derek Morgan
Database Administrator
Project: St. Louis medical school uses SAS software as its primary tool to analyze, report and manage data for multiple biostatistical research studies.
