Microsoft Enters Web Traffic Analysis Fray

LAS VEGAS -- Microsoft Corp. launched an initiative at this year’s Comdex to push the digital footprints that Web site visitors leave on Web logs into structured data that can be queried and combined with other warehoused data for business analysis.

Calling the initiative Business Internet Analytics (BIA), Microsoft is basing the offering on its flagship relational database management system, SQL Server 7.0.

Barry Goffe, SQL Server product manager at Microsoft, says the company’s approach addresses three problems businesses have when assessing how visitors are using their Web sites. Many Web traffic reporting tools run directly off the text logs, meaning users can only find answers to preformatted questions answered by canned reports; usable data can take days or weeks to get back; and these products have trouble handling large quantities of data.

"We’re trying to help dot-com companies make better decisions more quickly," Goffe says.

By moving Web traffic data into SQL Server 7.0, users can use SQL Server’s OLAP Services to ask any question of the data the next day, or they can use the relational aspect of the RDBMS to compare the data against other business data kept in relational stores elsewhere in the enterprise, Goffe says.

There is little doubt that many users are frustrated by the parameters of Web traffic reporting solutions, so Microsoft is entering a potentially large market. For Microsoft, associating SQL Server with such e-commerce reporting could result in additional SQL Server license sales and wed more customers to its technology.

The initial version of BIA will be a toolkit based on SQL Server 7.0 and a set of methodologies for Microsoft business partners to incorporate and sell as services. Microsoft is looking to fold BIA into a packaged technology that can be sold directly to customers. An intermediate step will be to integrate BIA features into upcoming releases of SQL Server and Commerce Server.

Microsoft ran a pilot implementation on its own internal Web site. With 500 Web servers and more than 35 million unique visitors per month, Microsoft generates 200 GB of clickstream data per day.

The company used the BIA methodology to process and load that clickstream data, starting each night at midnight and making the data available in SQL Server 7.0 OLAP Services for analysis by about 500 users within Microsoft by 10 a.m. the next day.

Interested parties within Microsoft included content developers, who assessed how many people accessed their content and where they linked in from; business managers who monitored Web ad traffic, partnership click-throughs, and other contractual obligations; operations managers, who assessed whether server farms were optimally dispersed from a geographic standpoint; and executives, who needed thumbnail overviews.

The technology fits into an emerging solutions group that allows companies to combine clickstream data with other data in a data warehouse. In the area of one logical step -- using the data to create, share, and sell consumer profiles -- one analyst speculates that the technology may be outracing business ethics.

Charles King, an analyst at Zona Research Inc. (, noted in a recent commentary on an IBM Corp. Internet marketing partnership, "While it is easy to understand why retailers and marketing professionals would be wild about these new toys, the way consumers will respond is less easy to predict. In the fierce debate raging about who owns consumer data, and how it can be sold and shared, retailers and marketing concerns are acting as if the dust has already settled. To our minds, this assumption is premature."

Must Read Articles