It is a cry for help that echoes through the corridors of organisations everywhere: “We know all the information is there – it is just in different places, in different formats and inaccessible.”
Just ask IT strategists at the UK’s National Health Service (NHS). As part of its efforts over the next four years to create an electronic patient record for every British citizen, it is working out how to corral information from scores of different sources (everything from X-rays and blood test results to GP notes and prescription records). Adding to that integration task, the NHS is having to develop sophisticated taxonomies that will govern access rights to different parts of the same record, so that accident and emergency teams can see vital medical history while administration staff only have access to a thin slice of the patient’s personal information.
Similar problems – if maybe not on always on such a large scale – exist within every enterprise, according to Richard Harris, research director at IT advisory group, Gartner. Businesses today collect vast amounts of information through transaction systems, email, scanned documents, demographic data and even voice. And making sense of that – turning it to the business’s advantage – is a huge challenge, he says.
As that illustrates, information management is now a major corporate issue. From the top to the bottom of the organisation, there is an understanding that information is critical to the business, and the management of that information is essential.
Defining and implementing an information management strategy presents, however, both a cultural and a technical challenge.
The cultural challenge relates to the way data has been gathered within organisations. Different groups – divisions, departments, subsidiaries, even individuals – feel they have ownership of data. They want to be able to gather it, define it and analyse it when and how they want. Those data stores are created by locally run applications or built up as local databases, resulting in numerous ‘silos’ in which access to data by ‘non-owners’ is limited or made impossible.
The critical challenge is to wrest control over the management of those ‘islands of information’ away from the controlling groups (while providing them with a sense that they still ‘own’ the data) – even as that data is opened up to other appropriate parts of the organisation.
The other key issue is technical. Providing access to data in an integrated and digestible form usually requires a major overhaul of existing infrastructure. Not only is that complex – it requires a major consolidation of data, standardisation of data, integration of data and data-cleansing effort.
One of the first barriers to overcome in developing an information management programme is the lack of understanding of its benefits, says Steve Gallagher, a senior director at Accenture’s information management services group. “The return on investment on spending £1 million on improving data quality, for example, as part of a wider information management strategy, is poorly understood,” he says. But knowing how many customers the business actually has or the true locations of those customers (in terms of accurate postcode) has immense value. The problem is that often such initiatives are competing for the same pot of money with a proposal for, say, a factory overhaul, he adds – a concept that most senior management can grasp more readily.
There are some outside influences, though, that have helped to focus management minds on information management concerns. Regulatory compliance tops the list. An inability to keep track of emails at drug maker, Merck, contributed to the company losing a high profile legal case involving its blockbuster arthritic painkiller Vioxx – at a cost of over $250 million.
There are plenty of other related business drivers supporting the adoption of more sophisticated information management programmes. Keith Funnell, desktop support manager at electricity company EDF Energy, highlights how his company saved on expensive storage capacity and streamlined access through the implementation of an email archiving system that filtered out duplicate email content. “The benefits of archiving digital information are seen at the operational level through efficiencies and at the business level through easy access to information required for compliance,” says Funnell.
As that suggests, those in charge of information management strategy need to have a wide knowledge of the issues involved: as well as the acumen necessary for getting senior management support for complex projects, they need to have a shrewd understanding of compliance issues as well as the diplomatic skills needed to take over control of a department’s jealously guarded data.
“Increasingly, enterprises find they must dedicate a core team to information management, [with skills] ranging across functional, business and compliance perspectives. A foundation of appropriate expertise is needed to support and sustain the enterprise-wide effort,” says Gartner’s Harris.
That is no mean feat. As Accenture’s Gallagher points out there are actually a shrinking number of IT professionals equipped with the skills to handle many of the complexities of information management – the building of database schemas, the categorisation of data, the definition of taxonomies and so on. “This goes beyond strategy. We are running out of people when it comes to this [implementation] phase of the work,” he says. The scarcity of professionals with skills necessary to drive an information management programme has encouraged many organisations to pass implementation of their information management strategies to third-party partners. That judgement call depends on the complexity of the existing information infrastructure.
But what is not in question is the need to make the existing information infrastructure more valuable to the business.
Data quality: the perennial problem
The vast majority of today’s businesses have sunk untold millions into their business applications. But getting reliable information from these systems is not always so straightforward.
Take consumer packaged goods giant Unilever, for example: it uses German business applications company SAP’s enterprise resource planning (ERP) software platform across its global business. The complexity of Unilever’s operations has meant that the SAP product needed to be implemented in 112 slightly different ways. Each of those instances makes the consolidation of company-wide data that little bit harder.
As companies standardise and consolidate, they frequently find they have a data quality problem, says Bill Swanton of analyst company AMR Research. “High data quality is necessary to make ERP work and the elimination of redundant data is necessary to capitalise upon the company’s global clout.”
But there are two stages to that elimination. First, organisations need to cleanse their existing data or data they are importing; then they need to start defining data from one central point so that its structure is consistent across all related databases.
The value of business information is undermined entirely if the data being analysed is inaccurate, incomplete or inconsistent. The extent of the problem is so severe that industry analyst group Gartner predicts that through 2007, more than half of all data warehouse projects will have limited acceptance, or be outright failures, as a direct result of a lack of attention to data quality issues
Stephen Brobst, chief technology officer at NCR’s Teradata data warehousing unit, says: “In my experience, no matter how bad people say their data is, it’s worse.”
UK-based bank Lloyds TSB calculated that poor data quality was costing the UK bank £200 million ($300 million). As a result, says Amanda Hughes, senior manager of wholesale data management at Lloyds TSB, she was granted the budget she needed to improve matters.
But that is the firefighting. The aim is to keep the data clean, and one key mechanism is master data management (MDM). By defining changes to data in one place and then rippling that through to relevant databases, MDM’s combination of business processes and technology ensures that key data (customers, suppliers, products, employees and so on) is current, consistent and accurate across the enterprise.