Four steps for sharpening up data management

Many organisations understand the importance of managing information in a consistent and structured way for effective corporate decision-making, but most are daunted by the scale of the task they face in imposing order.

The vast majority of enterprises still keep structured and unstructured data in functional silos to facilitate storage and management by the IT department.

But such an approach does not make sense to the business, which often cannot find what it is looking for, does not trust the quality or integrity of what it does find or is unaware of what information exists in the first place.

One of the problems is that often, users still have to employ manual processes to bring relevant information together, particularly as they frequently store multiple copies of data on random disk drives.

The situation is aggravated by increasing amounts of information emerging from an expanding number of sources. For example, corporate data is found in everything from spreadsheets and user-developed Microsoft Access databases to mobile phones used to take photographs during workshops.

Filing disciplines
Another issue, says independent consultant Graham Oakes, is that few organisations today use the traditional filing disciplines associated with the days of paper.

"We were all told that computer search mechanisms would make it easier to find information and that we didn't need to structure our data any more. But while search can do good things, it has now been overwhelmed by the volume of data," he says.

Although many companies paper over the problem by buying relatively cheap storage, the fact remains that if different versions of related information are scattered across the organisation, it becomes difficult to locate, access and exploit effectively. So what should businesses do?

Step 1: Define information strategy

The first step in any enterprise information-management initiative is to assess what information the organisation has. But this is not a job for IT managers alone.

Instead, it is crucial to involve all parts of the business to establish what kinds of data the enterprise generates and consumes, where it is stored, who uses it and to what end, as well as the challenges they face.

In this context, it may be useful to set up an official team or steering committee comprising representatives from each corporate department to formalise the process.

Once the state of play has been determined, it is then necessary to undertake a rough prioritisation exercise based on the perceived value of the different categories of information that have been defined. High-value classes of information generally include customer, product or service and financial data, although each organisation's focus will be slightly different.

Desired outcomes
The next stage of the process involves establishing what it is the business wishes to achieve and defining desired outcomes and possible obstacles. These findings can then be used to prioritise action and to form the basis of a strategic roadmap founded on a phased project delivery model.

This roadmap should also include a technical architecture, not least because infrastructure changes will need to be budgeted.

Given tight budgets, it may make sense to focus initially on lower priority, quick-win initiatives rather than the highest-priority leviathan undertakings.

Matt Leighton, a UK sales leader at IT services provider Logicalis, says the quicker you can deliver results, the more committed managers will be. "It is hard work trying to keep people enthused for six to 12 months, but if you can deliver something in six weeks, they'll be much happier and it'll be easier to get funding for the next phase of delivery," he says.

Step 2: Establish governance models

One of the challenges in moving from a world where data is owned by business functions to one in which information is a corporate asset is that organisations are forced to create common definitions for their data.

For example, the average company may have several definitions for 'revenues' or 'margin', depending on context, but the aim is to reduce these variations to a single, universally applied term.

This task is often politically sensitive, so the creation of a cross-functional team or steering committee will be vital. But precision is important because such definitions act as the foundation for everything else.

Another even more fraught activity is attempting to understand who in the business owns different pieces of information and who should be accountable for their quality.

Independent consultant Graham Oakes says IT is managing information on behalf of the business, so they have to know who owns it. "It's very important to ask the question but, unfortunately, it's a difficult one to answer and often just as hard to work out," he says.

Loss of control
One of the problems is that, because a range of different departments benefit from certain data, individual business units may not be keen on being the ones paying for it to be collected and kept up to date. Alternatively, they may unhappy about the prospect of apparently losing control.

One way of addressing this issue is to set up workshops that focus on a specific data set, such as customer information. The workshop should include staff who produce the data and those who consume it for their business decisions. Participants can discuss what information is available, how and why it is used, so that they can understand the overlaps and gaps in responsibility.

This conversation opens the way to explicit questions about the data owner and who is ultimately responsible for it. Such information can subsequently be fed back to the steering committee.

Data-steward role
It is also important to create a second layer of responsibility beneath that of data owners. The role of data steward is needed to ensure data definitions relating to different information types are applied consistently and quality levels remain high.

The job of data steward is likely to be part time unless fresh definitions are regularly introduced — relating to new product launches, for example. Procedures must be in place so problems can be escalated to the steering committee.

One useful means of ensuring data definitions are made readily available to the rest of the business is to introduce a master data-management repository that can be accessed via the corporate intranet.

Such tools also help data stewards as they generally include a workflow element that can be configured to flag new additions to the library. The goal is to ensure additions are consistently mapped to the correct information category and the correct hierarchy maintained.

Step 3: Streamline business processes

To optimise business processes, you have to understand the lifecycle of individual categories of corporate data. This lifecycle will vary depending on the information involved, but may include creation and version management, reuse and disposal.

Part of this activity will entail answering traditional records-management questions about how long data needs to be retained for compliance. The answers gleaned will form the basis of corporate information policies, which need to be actively enforced.

Independent consultant Graham Oakes says information lifecycles relate closely to business processes and provide a different view on the same thing. "So if you start to understand the lifecycle of your information, you'll be able to identify bottlenecks and where problems lie and that's where you'll want to introduce process change," he says.

Identify dependencies
Such change might involve tweaking the way people work or automating certain tasks, but business-process modelling tools can help identify dependencies between activities. They can simulate process flows and model the impact of any proposed changes.

But such tools can be expensive, complex to use and take time to master. Because they are IT tools, the danger is that the business will abdicate responsibility for their use to the IT department, rather than remain actively involved.

Any business-process change will also alter working practices, so the introduction of a change-management programme may become necessary.

A good idea in this context is to formulate a clear communications plan. The plan should include an outline of how and what information will be delivered to staff, how frequently it will be provided and by what channels.

Such communications are likely to incorporate information about the proposed changes, the rationale behind them and how staff are likely to be affected.

Formal structures
Feedback needs to be captured at regular points using formal and informal structures to understand what works. A formal structure might include a clinic, during which managers or key influencers answer staff questions, while a more informal approach might simply mean having an open-door policy or chatting to people over a coffee.

Such feedback can also prove useful in devising training schemes. But Nick Millman, UK lead at Accenture's information management service, warns that the focus should not be simply on training personnel to use a given tool, but also on helping them to understand the context of the information at their disposal.

"So if the tool is providing people with detailed product profitability data, train them on how to think about the insights and actions they might derive from it. It's about offering training that mixes the technology and business-value angle," he says.

Step 4: Tight implementation management

Tight project management is key to most successful initiatives. Enterprise information management programmes are no exception.

Project managers need to be flexible to accommodate necessary change, but strict enough to stick to their guns if scope-creep threatens delivery schedules and budgets.

Choosing the right technology for the job is important and that means not being swayed by suppliers' promises. Some large suppliers may say they offer a fully integrated suite of enterprise integration software, but the reality may be somewhat different.

Many vendors have simply bundled together a range of loosely integrated tools gained from multiple acquisitions, but many suites are still not complete and only tackle elements of the problem.

Integration headaches
Depending on their requirements, organisations may be better off opting for best-of-breed components, whether document management systems or email archiving tools, with all of the inevitable integration and management headaches that such an approach generates.

But the absence of a one-size-fits-all approach to enterprise information management also means organisations can end up being taken down a blind alley if they make technology decisions too early.

Mark Smith, chief executive of Ventana Research, says the business information requirements must mandate technology decisions, not the other way around. "Otherwise, you could end up boiling the ocean," he says.