informa
/
Feature

In-Depth Review: SAS Enterprise BI Server

Our expert explores the strengths and weaknesses of SAS' flagship business intelligence entry, and outlines how it stacks up within an increasingly complicated BI marketplace.
Basic BI Analytics

Many users associate report-writing and cube analysis with basic BI analytics. This is true for basic scorecarding, "where-are-we-now" and "one-version-of-the-truth" compliance and reporting data. SAS Enterprise BI Server still concedes some features and functional advantages to Hyperion and Microsoft in OLAP services, or to Cognos and Business Objects in reporting, and to MicroStrategy and Cognos in real-time notification and business activity monitoring. But SAS is closing those gaps.

Moreover, basic analytics goes well beyond reports, charts and cubes. Being able to carry out problem analysis, basic testing and modeling is critical to many organizations. Base SAS, which sits at the core of SAS Enterprise BI Server, provides a wealth of financial, forecasting and statistical analysis tools critical for problem-solving and competitive agility. Knowing where you are is important. But being able to forecast where you and the market are heading takes BI to the next level. SAS is clearly positioned best among the BI vendors to deliver that capability.

The extra-cost SAS Enterprise Guide and SAS AppDev Studio allow technical specialists to construct SAS procedures that can be used as wizards and templates for gathering data for analysis, making forecasts with improved accuracy, or creating test models to assess different options. These SAS processes and procedures are then available for use in the SAS OLAP Web Viewer or the SAS Web Report Studio, among other SAS data display tools.

In effect, SAS is using its procedural language as the overall macro language for all its services. To make SAS procedures more accessible, Enterprise Guide provides a general workflow and GUI design system for creating wizards and macros while AppDev Studio provides specific processes and tools for creating SAS Management Console extensions, portlets for the SAS Information Delivery Portal, and more sophisticated SAS procedures and templates.

These tools have the complete SAS solution set at their command. So clients who have purchased extra services such as SAS Enterprise Miner or SAS/ETS for Forecasting know they can be incorporated into the BI processing. However, one drawback is that SAS, like almost all BI vendors, has a mixed record at supporting other vendors' BI tools. Getting SAS Enterprise Miner input is easy; but that's not necessarily so for IBM Data Miner or SPSS Clementine. However, in both basic BI statistical analytics and broader predictive and optimizing capabilities, SAS has a leadership position in the BI market.

Portal Problems

One would think that the overall BI industry would set in place good mechanisms for data-interchange among their various tools. On the data input side, vendors like SAS do a pretty good job. In data mining and OLAP, efforts by major players like IBM, Microsoft, SAS and others have had limited success. But the emphasis should be on "limited," because BI vendors are distinctly non-promiscuous in sharing their analytical results with other vendors. Nowhere is this more evident than in portals.

Portals should provide the long-promised "information at your fingertips." This is the proverbial "last mile" to users' minds, attention, and systems. Drill-downs should be able to switch from one BI source to another. Notifications and event response should be vendor-agnostic. Unfortunately, these services are hardly straightforward.

Take, for example, the problem of combining BI results from different vendors. First, the results must be output into a neutral container -- say, HTML, Excel tables, PDFs, or XML files. Then the data may have to be mapped onto common naming and formatting conventions. Finally, the combined data will have to be redisplayed with common features and display mechanisms. The portal challenge is BI's dirty little secret.

SAS, with its Information Delivery Portal, is no exception. First, information portlets run on only one client -- Microsoft desktop and the IE browser. Here we lose desktop platform independence, one of the key benefits of Web delivery. SAS information portlets and its AppDev Studio do a good job of bringing together raw data and combining just about any SAS System results. But after that, SAS falls down. It's true that SAS Web OLAP Viewer can display any OLE DB for OLAP or MDX-compliant source of data. But that still leaves most vendors out of reach. Likewise, report-writing and other BI outputs are subject to the "BI two-step": 1) Stop and translate into a compatible format and 2) Hope all convert properly so analysis can continue.

If BI vendors can agree to standards, then SAS's AppDev Studio and its XML-mapping facilities are in a good position to implement those standards fairly quickly. But as SAS Portal stands now, its dependence on Microsoft IE and me-too support for other BI vendors' results makes it "me-too" as well.

Conclusions

SAS Enterprise BI Server is very good but not necessarily best-of-breed in OLAP or reporting. SAS, like all BI vendors, has work to do in sharing and integrating output among analytical processes. Thus, SAS has portal problems. But in the arena of basic analytical processing and making its own data and results available to its BI users, SAS Enterprise BI Server has industry-leading capabilities. The key advantage is SAS Data Access and SAS Procedure language which allows developers and end users to readily customize access to the complete inventory of SAS applications. Thus, SAS has system-wide BI integration capabilities and reach that most other vendors can't match. If you are into BI for the long haul, SAS has to be on your short list.

Jacques Surveyer is a wide-ranging technology writer, see some of his other coverage at theOpenSourcery.com.