NEW TOOL ANALYZES CUSTOMER DATA FASTER
A start-up company is claiming it has new technology that can quickly analyze customer behavior from a variety of sources, from clicks on a Web site to purchases at the store down the block.
So far, Digital Archaeology (http://www.digarch.com) has raised a modest $8.5 million in venture capital while creating a quick and flexible data mart-building capability. With its announcement recently of c-Discovery, it has placed a customer analysis application on top of its Discovery Suite data mart technology for slicing and dicing customer behavior.
Many firms already address this area as suppliers of data warehouse, data mart and data mining tools. What's different about Digital Archaeology, said Mitchell Kramer, database analyst at the Patricia Seybold Group, is the speed with which it can get adata mart an information store house up and running.
In addition, it is bringing a proprietary, high-level query and analysis language to business users instead of database administrators to explore the data mart, said Mark Smith, program director for application delivery strategies at the Meta Group.
Based on its new techniques, for which it has applied for patents, Digital Archaeology's product suite "can result in virtually instant implementation of data marts that integrate query, charting and analytic capabilities," Kramer said in an analysis.
Lying underneath c-Discovery is the Discovery Suite 2.0 set of tools for building customer analysis data marts. Unlike most data mart tools, Discovery Suite imports all the data of a Web server log file instead of selected data. Or, in the case of a bank, it might import all data from automatic teller machines, rather than scrubbing or cleansing the data first.
Discovery Suite 2.0 is available immediately on Windows NT for $85,000 or Sun Microsystems' Solaris for $120,000. C-Discovery is available immediately on Windows NT for $130,000 or Sun Solaris for $170,000.
David Frankland, Digital Archaeology's president and chief executive, said Discovery Suite deals with data in bulk, rather than selecting it to fit the preconceptions of the data mart builder. When a specific query needs to be answered, the underlying database server builds "an extended data set" out of the pool of information in response to the query.
While existing Online Analytical Processing engines do much the same thing, the Discovery Suite database engine does not need to structure the data or build indexes to access it, as relational based engines need to.
Having executed one query, Discovery Suite can turn around quickly and execute another by generating a new extended data set or reusing the set just generated.
"You can put the data in a more limited format for the analyst to look at, or you can go back to the raw data," Frankland said.