To Mine or Not to Mine Your Own Data
In the world of data mining, nothing is easy. But in recent years some application server providers (ASPs) have developed a business model that promises to simplify the process and lower the cost of mining massive amounts of data.
Large-scale data analysis is difficult, costly, and time consuming. To ease the burden on enterprises, WhiteCross Systems Ltd. (www.whitecross.com) developed an ASP data mining service that lets large corporations outsource their data analysis needs.
Using its data exploration server -- WX/DES -- to power an analytical outsourcing service, WhiteCross says it can lower up-front costs and offer deeper analysis of large volume data. Recently, the UK-based company inked an agreement with British Telecommunications PLC (BT, www.bt.com), in which BT tapped the WhiteCross ASP service to help perform ad-hoc queries on its complex data. WhiteCross will help BT obtain a more complete view of its customers by analyzing product purchases, lifestyle information, and billing information.
Companies thinking about outsourcing data analysis should consider a few critical issues -- including possible compromised data security and the cost of buying data mining solutions versus renting them -- before deciding.
"In general, there will always be a fear of turning sensitive data over to a third party," says Mike Schiff, director of data warehousing strategies at Current Analysis Inc. (www.currentanalysis.com). "The fear won't go away, but steps can be taken to minimize the risk."
To lessen security concerns Schiff suggests using an established vendor and making sure there are appropriate safeguards in the contracts. He also says using an ASP that is not closely associated with or owned by a competitor will decrease security risks. Still, companies will have to learn to deal with a "loss of control" of their data before they can outsource their data mining needs.
While it will be ultimately up to the enterprise to weigh the pros and cons of outsourcing its data analysis needs, ASPs have a lot to offer, especially in the area of cost savings. ASPs offer companies the opportunity to "rent" services, software, and data analysts -- things that are expensive, hard to come by, and difficult to retain.
"The positives that an ASP can provide are lower cost of entry, both in terms of equipment and expertise," Schiff says. "Consequently, most companies will find it more cost effective to 'rent' the equipment than to buy it and this is one of the advantages that the ASP model offers."