CNET también está disponible en español.

Ir a español

Don't show this again

Tech Industry

DigiMine aims to take software big time

The statistical analysis start-up is set to release new data mining software that it hopes to use to woo Fortune 1000 clients and others in corporate America.

    DigiMine will release new flagship software Monday, hoping to attract Fortune 1000 clients and demonstrate to corporate America the potential for a complex statistical technique called data mining.

    The Bellevue, Wash.-based start-up will begin peddling its Enterprise Analytic Services to large companies struggling with how to manage overflowing amounts of data collected from customers on the Internet, through catalogs and at stores. New York-based clothing retailer J. Crew was the first customer to test the new product in a pilot program.

    But it is unclear how many other companies will be able to foot the bill for the service: The start-up fee ranges from $30,000 to $40,000, depending on how much data the company collects and how many reports it wants DigiMine to produce. Ongoing fees for data warehouse hosting, maintenance and daily reports are $30,000 to $40,000 per month.

    The timing of the expensive product is risky. As the economy turns sour, many companies are reining in spending on information technology. For example, Agilent Technologies announced last month a plan to slash discretionary information technology spending by 90 percent this year, and the company has ceased buying new personal computers for its work force.

    But the downturn does not worry 140-employee DigiMine, one of about a dozen companies jockeying for dominance in the emerging niche. Earlier this week, the company introduced a service to track the behavior of people who surf the Web from wireless devices.

    Nick Besbeas, DigiMine's executive vice president for sales and marketing, says he hopes plenty of companies will pay $30,000 to $40,000 per month for data warehousing, hosting, maintenance and daily statistical reports.

    "Your knee-jerk reaction is that's a lot of money," Besbeas said. "But if you're dealing with large amounts of data, the alternative is to engage a consulting firm for millions of dollars to do the same thing or spend millions of dollars to build a solution yourself."

    Data mining is the use of statistical analysis to uncover hidden patterns in otherwise random information. The science is expected to revolutionize the Internet during the next decade, resulting in a "personal Web"--both wired and wireless--tailored to an individual's preferences.

    Through data mining, marketers can target customers with personalized stock quotes, news updates, special promotions and other information they are most likely to use, dramatically reducing advertising budgets and boosting revenue. It is also entirely automated, reacting instantly to changes in a customer's behavior, unlike the vast majority of personalized services on the Web today that ask people to fill out questionnaires.

    The newest software is a high-end version of its core product, DigiMine Analytic Services, which costs between $10,000 and $20,000 to start and at least as much in monthly fees.

    At double the price, the fancier version adds features that large companies might demand. Data reports crunch numbers for discreet company divisions or products and also provide a broader analysis for senior executives looking for a bird's-eye view of the organization. The software can process hundreds of gigabytes per day of online and offline data.

    It also features a product called ReportBuilder, which allows statistics-savvy employees at the corporate client to create custom reports and tweak spreadsheets and algorithms as they wish.

    The move is aimed in part at correcting a myth that many data mining companies have overly dumbed down software so that marketing executives with scant statistical knowledge can use the products. Companies that employ statisticians can easily customize the new software to result in more complicated charts and analysis than the core product can deliver.

    "We're trying to focus on the details that most analysts want--the people who must go beyond summary views or custom reports," Besbeas said. "They want to get hands-on with data and manipulate it in real time. It's not for everyone."