In 1996, Microsoft entered the online analysis world by acquiring technology from the Israeli company Panorama Software. Under the name OLAP Services, this technology became part of SQL Server 7.0 in 1999. In SQL Server 2000, the product was expanded and renamed Analysis Services. The latest analysis product is a powerful and flexible online analytical processing (OLAP) provider, but it doesn't include a client tool for querying, reporting, or analyzing data. Customers who want to use Analysis Services without writing a custom client browser must therefore find a suitable commercially available client application. To help readers select a client application that meets their specific needs, we compared five such products. After at least two consultants evaluated each tool, we reconciled and summarized the findings. We made certain that at least one of the reviewers of each tool had worked with that tool in a client setting.
To find a current list of client tools, we looked at Microsoft's Data Warehousing Alliance Web site (http://www.microsoft.com/sql/partners/DWA), which contains a cross-referenced list of approximately 20 vendors that provide front-end tools. Rather than perform a trivial analysis of all 20 products, we chose to focus on five products we've used and evaluated at the request of multiple clients: Cognos PowerPlay, Crystal Analysis, Microsoft Office XP PivotTables, Panorama NovaView, and ProClarity Analytics Platform. In this article, we share our methodology and vendor evaluation files so that you can make an apples-to-apples comparison with other vendors' tools.
After you've read this review, one or two of the five tools will likely be on your short list. However, keep in mind that your best choice might not be any of them. The purpose of the article is not to declare one winner, but rather to help you understand the strengths and weaknesses of these products specifically and of client tools generally. You should acquire a good sense of which features are common to most products and which features are rare.
As our team analyzed the tools, we followed a two-stage process. First, we created a standardized Analysis Services database and developed a standardized template of tasks, then asked each evaluator to use the targeted tool to try out the complete list of tasks. In this stage, our goal was to be objective and consistent. Second, we asked each evaluator to prepare an overall assessment of the tool's usefulness for users in various roles, along with impressions of that tool's strengths and weaknesses. Our goal in the second stage was to be subjective and pragmatic. The subjective evaluation includes not only the results of the objective analysis, but also experience from working with the tool in a real-world setting. In this article, we present the high-level, pragmatic, somewhat subjective analysis on a product-by-product basis. Then, in the Web-exclusive sidebar "Getting Down to the Nitty-Gritty," at http://www.sqlmag.com, InstantDoc ID 26486, we recap the results of the detailed analysis, highlighting significant product strengths and weaknesses on a feature-by-feature basis. In all, we evaluated more than 100 features, which we divided into 18 subcategories, then grouped into four categories. The sidebar's Table A shows our ratings for these categories and subcategories.
Inevitably, we missed unique features of each product. Some product features are subtle, and you stumble onto them only after months of intense work. Some features are so specialized that our apples-to-apples comparison missed them. And we didn't cover some larger topics, such as scalability, seamless support for multiple OLAP server architectures, and tight integration between relational and OLAP reporting because they're beyond the scope of this article.
Vendor Overviews
The high-level overviews of each product follow alphabetical order. For each product, we summarize the benefits that the product offers to users in different roles, then highlight what might be make-or-break attributes. User benefits are important because features aren't the same as benefits. A feature that might provide considerable benefit to one user might provide no benefit at all—or might even be a detriment—to a different user. To effectively evaluate a tool, you must think in terms of the ways people will use it. We identified three archetypal roles a user might play and asked evaluators to assess how well the product meets the needs of users in each role. (The same user might perform different roles in different contexts.) The three roles correspond to the three general functions of a client tool:
- the power analyst—analysis
- the data gatherer—querying
- the report user—reporting
To perform free-form ad hoc analysis, power analysts require the full analytical power of Analysis Services. They're willing to learn the details of the source database design and of the query tool to obtain the necessary results, and they often create reports that others use.
Data gatherers have to be able to dynamically query a database without becoming experts in the source database or the query tool. They want a guided user experience that permits drill-down and pivoting, yet eliminates options that might create undesirable results (e.g., restricting the view to fiscal year 2001 while simultaneously displaying the months from calendar year 2002).
Report users require standard reports that might be brief or extended and which often include charts as well as tables. These users want to scan consistently structured reports without needing to drill or slice to find the desired values. Producing this kind of report typically involves creating static reports, either in printed form or as static HTML pages or other documents. Usually, headers and footers (which include such information as the date the report was generated) and titles that repeat on different pages are crucial.
As you identify which roles are most important in your organization or for your application, you can focus on the appropriate features of each product. During each vendor evaluation, we describe how well that vendor meets the needs of users in each role.
All the products reviewed support the basic OLAP operations of slice (filtering by a single member from a dimension), dice (displaying multiple members in a grid), and pivot (switch a dimension from one axis to another). The review focuses on how easy these operations are to perform and their flexibility. Most products come in both a full-client desktop version and a thin-client Web version. For products that come with a Web version, we highlight subtle distinctions between whether the client is truly zero footprint (the client doesn't download controls and leaves nothing behind) or whether it requires Java applets or even a full download of an ActiveX control.
Some products include make-or-break features. Identifying these make-or-break features is important because it can often help you simplify the decision-making process. Typically, a make-or-break feature either elevates the product above the others or removes the product from consideration, permitting you to bypass a more detailed analysis. But remember that features might change or another mechanism might meet the same need.
Prev. page
[1]
2
3
4
5
6
next page 
