On the web: Getting to grips with growing content

In keeping with the goals of the Office of the e-Envoy, which sets out that all government services should be available by 2005, the DWP has recently launched its intranet future-development programme. Bernie Bowker outlines the aims of the programme and the key challenges that he and his team had to overcome to ensure successful implementation.

The Department for Work and Pensions (DWP) was created after the last British election with the principal aim of implementing the Government’s Welfare to Work strategy. The department is a large organisation with around 125,000 staff in over 2,000 locations across the country, responsible for delivering support and advice through a network of services to people of working age, employers, pensioners, families and children, and disabled people.

Like many other government departments, the DWP is under pressure to meet the objectives of the Office of the e-Envoy, which is responsible for ensuring that all government services are available electronically by 2005, with key services achieving high levels of use.

To improve the delivery of public services and achieve long-term cost savings by connecting online government services around the needs of customers, the DWP recently launched its intranet future-development programme. In essence, the programme is a business-change initiative aimed at modernising the way the DWP conducts business. The present DWP has in excess of 500,000 files (even after a massive data-cleanse exercise) and is growing in line with other comparable intranets at a rate of 100 per cent a year. The intranet is, therefore, a real driver for improvement, sitting at the heart of the DWP’s modernisation programme.

The most common project in today’s enterprise that has the potential to become a major headache is the corporate intranet. As such, it’s imperative to agree the objectives of the project well before implementation. For us, the primary focus was to establish some control over intranet content.

With new content generated so quickly every year, the need for an automated content-management system (CMS) is imperative. Without one, it will be impossible to keep information up to date. Keeping track of people and teams with permission to publish information on the intranet, and knowing who authorises it, will also become more complex and difficult. We opted for both a centralised and decentralised approach to managing our content, allowing people to publish and control their own content where possible. There is a central IT resource responsible for building and maintaining sites, but we also have a few business areas, especially ones that go back to the old agency days, that have their own staff who are skilled in Dreamweaver and maintain some sites locally.

As well as the need to manage content, finding information will also become more difficult as information becomes increasingly complex and the intranet boasts greater functionality. Search engines, like the one we currently have in place, can be effective, but often retrieve too much unwanted information. So, on the search side of things, there is quite a lot of user dissatisfaction at the moment. Our present engine is not configured optimally. We need to ensure that we have a strategic way of searching within an information architecture giving users the same experience no matter where they are.

Improving accessibility is also a key ingredient of the project. Finding information and the functionality required to do one’s job is only part of the problem. As more business information is made available online, users could be faced with having to navigate through several site areas in order to find relevant information. Also, as business usage increases, people will want information that is tailored to their needs, and not those of the intranet-design team. This will involve not just personalising content, but also changing how this information is presented.

As intranet usage becomes de facto for the DWP, and more information and functionality are available online, document management will become critical (eg, EDRM). Document repositories will need to be created, and archiving and retrieval solutions developed. Improved approaches to how documents are structured and presented online, coupled with greater re-purposing of content, are also priorities.

The culture of the DWP also needs to change. The intranet must become the medium for doing business, with people routinely using the intranet to collaborate and share knowledge. As an organisation, we must support the exchange of ideas by providing access to information resources across all levels of the department, and build a learning organisation by managing information more effectively and by promoting change actively. This is a massive challenge and one that must be met if we are to realise the full potential of our intranet.

Another goal of the intranet programme is to deliver web-enabled business solutions, ranging from electronic-forms processing to full workflow-based benefits processing linked to benefit guidelines. Much of the technology to deliver these solutions exists today, but the scale and cost of many of these solutions means it is important to have clear and well defined requirements before starting on their development. One difficulty is that without experience of enterprise-wide e-business solutions, it’s not always easy to establish these initial business requirements.

The development programme

To ensure the success of the programme, a series of inter-related activities to strengthen the underlying DWP intranet infrastructure was initiated. It is important to keep all DWP intranet programme stakeholders fully informed of our progress here, so we put in place a single, co-ordinated programme of work with progress reported on a monthly basis.

It was also agreed that without the successful implementation of this improvement programme, the intranet would not be able to support fully and maintain the increasing demands being made by business units.

The core areas of the programme include:

Data cleansing – over a 12-month period (which ended about three months ago), we removed duplicated, out of date and redundant content by as much as 40 per cent. More than 100 per cent of the original starting figure of intranet files has been removed;
Classification – a taxonomy based on government category lists was agreed, allowing for focused search capabilities;
Content management – following an evaluation of prospective vendors by EDS, the DWP’s lead IT service provider, we chose a leading CM vendor to provide our CMS.

Data cleansing

The data-cleansing process was a vital first step. It is the essential, logical forerunner of other programme activities and will provide a robust, accurate information base upon which taxonomy, content management and personalisation can be developed.

Independent reports by leading industry consultants in this area, such as Forrester and Ovum, have consistently referred to the unmanaged content and ad hoc site development that is part of the intranet evolutionary cycle. It is essential to remove duplicate, outdated and redundant information from the intranet before moving forward with other applications and components. In fact, inadequacies in our content or inconsistencies in metadata will be highlighted by other applications like a portal. If we do not tackle the underlying issues in intranet content, we will have limited success in realising anticipated business benefits. In the same way, if new content is provided but not in a controlled manner, benefits will not accrue.

At the DWP, we believe the intranet infrastructure must be in place before it’s worthwhile investing in a search engine, a CMS and, ultimately, a corporate portal.

The steps involved in the data-cleansing exercise are as follows:

Management of the implementation of metatags across the intranet;
Identification of areas of the intranet that can be cleansed;
Validation of all intranet content;
Management of the cleansing;
The actual physical job of cleansing and associated archive/retrievals;
The ongoing development of mitigating action to prevent problematic areas re-emerging;
Establishment of a CMS and a process to keep the data clean.

The cleansing project removed redundant content before implementation of the CMS. At the start of the exercise, 380,000 files existed, including duplicates, out-of-date and ownerless information, with the business risk of another inherited SERPS problem.

A small team of five people re-invented the concept of metadata. The logic was that all files needed to be examined to establish ownership and currency. Files that could not have metadata attributes applied were referred for controlled deletion by a data-cleanse team.

Over 200 web authors in the department were ‘tracked down’. The project rationale/expectations were explained to them via workshops, meetings and guidance. A data-cleansing scan verified the existence of correct metadata in a particular website. Any exceptions were highlighted for web authors to fix. Subsequently, excellent relationships were built up and maintained over this period. This process was run 3,500 times to ensure a clean interface for data migration to the CMS.

The site would now contain at least one million files had the exercise not been undertaken. New files were added to the intranet throughout the data-cleanse exercise. In addition to the starting point of 380,000 files, a further 554,984 were added, leading to a total figure of 934,984. In total, we deleted over 468,538 files, which was completed in November 2002. The current number of files under control on the intranet is 539,420.

My team worked very closely with the CMS service provider to develop the logic and non-standard functionality to apply metadata to non-HTML files and to convert the legacy metadata to the mandatory e-Government Metadata Standards (EGMS) set.

Overall, the team has made a considerable contribution to the DWP intranet objectives by ensuring content and information is reliable and up to date. The team has been acknowledged by the Office of the e-Envoy as leading the way in EGMS/metadata application.

Classification of DWP information – Taxonomy

A well-researched and constructed system for the classification of DWP information is essential for the success of any website (intranet or internet). It enables content to be organised efficiently, which allows for the speedy and easy retrieval of information, and for content to be put into context.

A taxonomy is required to complement the intranet search engine by enabling users to navigate through levels of increasing specificity to a list of titles and descriptions of documents held on the intranet.

Through discussion and input from the business, a complete and controlled categorisation list for all information published on the DWP intranet will be provided, using no more than five levels. Compliance to the Government Category List (GCL) will ensure that the classification for a document published on the DWP internet is compatible and can be easily linked into the GCL. Appropriate change control mechanisms for the taxonomy will be established. However, management of the approved and implemented classification will be controlled by a librarian function and, as such, will be governed by the communications directorate.

The high-level steps involved in the taxonomy project are:

Design higher (i.e. one and two) classification levels;
Seek PB/DB approval for top two levels;
Design lower (i.e. three to five) classification levels with business engagement;
Implement classification;
Classification builder as software element;
Map existing documents to appropriate categories;
Ensure new content is metatagged appropriately;
Educate and train publishers/intranet users;
Establish change control mechanisms;
Handover management of classification to live service.

Content management

A full business requirement for DWP intranet content management has now been produced and addresses the following areas:

The management of information through its full life cycle, including capture, creation, retention, archiving and destruction;
Tools and procedures for capturing content from contributors;
Processes for ensuring that submitted content is checked and approved before being published to a live system;
Classification of submitted content, including adding metadata, publishing and expiry dates, and version control;
Administration of the system, including creating users and roles, defining approval processes, and managing information on the intranet.

Tools are also required to manage the site, for example identifying and dynamically replacing broken links.

As mentioned earlier, the DWP intranet currently holds over 450,000 files, and approximately 125,000 staff will have access to it. This will increase significantly in content size and application functionality over the next few years. As the size and complexity grows, and the number of contributors increases, the costs of maintaining the intranet are likely to spiral unless an automated approach is available. There are also significant risks to the department if content is not controlled – giving the wrong advice as a result of out-of-date or inconsistent information can be extremely costly.

The next step

We’ve achieved our goal of establishing control over our intranet content. But we’ve still got a long way to go before deploying fully all the functionality that is available through the CMS product suite. Once we started evaluating the products that would enable us to establish the control we wanted, it became clear there was a lot more functionality with the product that we hadn’t thought about or were even aware of. If we had just implemented the control element, like we first wanted, there wouldn’t have been much in it for the business.

With the data-cleanse exercise completed, and a CMS and taxonomy in place, the next stage will be to migrate to a portal with enhanced search and personalisation capability. This will be very much a natural evolution for the intranet, a move that supports collaborative efforts and facilitates knowledge sharing. A pilot will start shortly, with a view to improving our business processes.

Bernie Bowker works for the Department for Work and Pensions. He can be contacted at bernie.bowker@dwp.gsi.gov.uk