Bad data eats IT budgets

Effective data integration is vital for business success

New Asset  

Effective data integration is vital for business success





In most businesses, IT departments take little notice of data quality or ongoing data management. Senior management also tends to under-estimate the need to focus on these challenges.

According to a global data management study by PWC, 75% of companies surveyed admitted that defective data impacted them financially. And 33% were forced to delay or scrap new systems. One company reported that data problems had caused it an £4.4m loss in one year alone.

Gartner Group says data quality is a serious problem that is imperilling business initiatives, future growth and long-term survival. It has found that more than a quarter of critical data in Fortune 1,000 businesses will continue to be inaccurate or incomplete until 2007.

But why is keeping data accurate and the management of that process such a big problem for so many companies?

Most firms purchase multiple data lists for numerous business functions. The result is a number of disparate databases with much duplication of data. While many companies are tackling this issue by embarking on data integration projects, they give little thought to the data itself.

Gartner says throwing technology at data quality issues does not solve the problem. Clearly data integration is more than integrating the framework or infrastructure that the data resides within. Users need to optimise the data within each system. While this may seem an easy task, it seems to stump many companies.

Data optimisation involves cross-checking, converting and combining information into a consistent whole. This is difficult because the data is peppered with errors. There are also anomalies in the way data is entered, which makes deduplication difficult.

And using out-of-date information affects business productivity and profitability. Imagine chasing a sale, only to find out later that the company was actually bankrupt. The knee-jerk reaction to such a problem would be to "buy a list" or a conduct a data cleansing exercise. Don't be fooled, data management is a long-term, methodical process that requires commitment and focus.

IT departments which recognise that data integration projects should involve someone with a background in information will go a long way to saving their projects from failure. Information experts who bring a background in editorial content and research will be able to introduce content optimisation processes and appropriate data taxonomy.

Users cannot afford to ignore data quality. By investing now businesses will reap rewards in improved operations and decision making.

Paul Brown is vice-president of OneSource Information Services UK

Read more on Privacy and data protection