Poor-quality data is biggest CIO headache
Published: 09 May 2006 10:05 BST
Poor-quality and duplicated data is leading to bad decision-making and undermining costly investment in business intelligence projects, according to industry experts.
Business intelligence software allows organisations to pull together information from various corporate systems and produce reports that aid management decision-making.
But industry players, analysts and customers warn that those decisions are often being made using poor quality, out-of-date or incorrect information because of a failure to undertake data-cleansing exercises first.
Speaking at business intelligence vendor Business Objects' customer conference in Cannes this week, Gartner analyst Andy Bitterer said most companies have about 200 data sources, and much of it is poor quality and inconsistent.
He said: "There's not one company that doesn't have a data quality problem."
Mike Pratt, data integrity manager at Business Link London (BLL), said many companies simply think business intelligence tools will solve all their problems without thinking about the quality of the data that the tools will draw upon.
He told ZDNet UK sister site Silicon.com: "Many companies don't do the groundwork. They need to ask what is it they are trying to measure; do they capture that data; and do they have the quality of data to make it meaningful."
Pratt said that by cleansing, purging and standardising its customer database as part of a business intelligence project, BLL has reduced duplication of addresses and contact details from 50 percent to 2.5 percent and saved £100,000 by not sending out irrelevant marketing to people who don't even exist.
Business Objects chairman and chief strategy officer Bernard Liautaud said data quality is one of the biggest headaches for CIOs and senior executives today because of the amount of information enterprises hold.
He said: "There's too much data and it's duplicated hundreds of times. The mistake companies make is that they start from the data they have. They need to ask what data do their users need and what are the questions they are asking. Understand the questions, how they can be answered and what kind of data is needed."
Did you find this article useful?
84 out of 158 people found this useful
Digg
Slashdot
Del.ici.ous
Stumble
Reddit
Full Talkback thread
1 comment
