Once a year, the business intelligence [BI] industry holds its breath in anticipation of the publication of the most authoritative and comprehensive guide to BI products: The OLAP Survey, a sister document to ‘bible of BI', The OLAP Report.

The research is based on deep technical bench testing by BI analyst Nigel Pendse and his team, and a global survey of BI users' ratings of product features and functionality.

For many years at BI tools vendor MicroStrategy, the publication of the OLAP report was always greeted with dread. Pendse seemed to be hyper-critical of the product, highlighting technical shortcomings that typically put it behind rival offerings from Cognos, Business Objects, Hyperion and Brio. During the period from 2001 to 2003, he was keen to raise a red flag against MicroStrategy for some of its questionable financial management.

That was not the sensation at MicroStrategy in late 2004, when The OLAP Survey gave MicroStrategy the highest overall rating of any BI vendor. The company's eponymous product was regarded as the best available for analysing larger data sets – with a median user data volume of 313 gigabytes.

That was predictable, as handling and analysing large data volumes has always been the company's specialism. However, the product was also rated as the most likely to deliver project success and business benefit to customers, and was the product where customers typically invested in the largest number of user licences – a median of 1,100.

But confirmation that users were voting with their wallets too was not far behind. Revenues for the company's third quarter ending 30 September were up a stunning 42% to $60.6 million, with software licence sales growing even faster at 46%. And while Microstrategy added 90 new customers in the quarter, including office supplies company Office Depot and the FBI, its net profits, excluding a tax windfall, leapt to $18.8 million, giving it a 31% net margin.

What makes that picture even more impressive is that, only a few years ago, MicroStrategy was written off as a casualty of dot-com exuberance.

During that period, it over-extended itself into markets such as data broadcasting that were only tenuously related to its software. A revenue slump, huge losses, the build up of a vast debt and investigations into false accounting followed.

The rebuilding of the company has been a slow and humbling task for Mike Saylor, its co-founder and CEO. But it has been an ultimately successful one – to the point where MicroStrategy can legitimately claim to be the fastest growing BI vendor and take its place among the top three or four companies in that sector.

Over a two year period, it has repaid debts of $150 million, and is now generating $1 million of cash a week, according to Saylor.

On a technical level, much of the effort has been focused on broadening the scope of its product and on creating a single, integrated platform to underpin its range of BI tools.

It has fixed a weakness in reporting tools, for example, with the late 2003 launch of Report Services, which provide users with business, operational and fixed-format reporting.

It has introduced OLAP Services, which gave it online analytical processing (OLAP) capabilities for creating multidimensional views of data equal to that offered by PowerPlay, the popular OLAP tool from rival company Cognos.

It has changed its user interface so that its tools can be used by a broader set of employees and unified the underlying metadata of its tools to that they work as an integrated suite. And related to that, it has added the facility to generate reports directly from Microsoft Office applications, such as Excel spreadsheets.

Importantly, it has finally embraced Unix, after years of only offering its software on Windows servers. With the latest version of the product, MicroStrategy 7i, the company now supports IBM's AIX and Sun's Solaris flavours of Unix, and Saylor promises a Linux version in 2005.

The aim of all this was to eliminate elements or shortcomings that some customers find off-putting. "What was the traditional reason customers didn't buy MicroStrategy? One was that we didn't have the ease of use and look and feel of a product like Microsoft Access. Second, was that we did not do reporting like [the widely installed] Crystal reporting tool; and, lastly, we did not run on Unix. We've eliminated those gripes," explains Saylor.

The remaining thing to do, he says, is to provide native support for running queries against SAP's Business Warehouse BI services. That should be in place by the end of 2004, says Saylor.

There is still one major gripe that Pendse hears often, though, and which Saylor looks unlikely to address. "MicroStrategy is not cheap," Pendse observes. "But it has a strong following at the high end of the market where price matters less."

As the company's recent financial results might indicate, price has not been proved a deterrent so far. "Customers are making $10 million and above commitments for five years or more and rolling out production systems with 20,000 active users," says Saylor.