Cover Story – The road to content management

They soon realise that visitor numbers are likely to improve, and people are more likely to return to the site if information on it is “refreshed” regularly. So they give the web master a couple of assistants, and internal departments speed up the generation of content for their potential target audiences.

An element of personalisation is introduced as companies try to shape the view various different audiences get of the site in order to make it more relevant. All too quickly, the task of translating text documents into HTML without a proper content management (CM) system, becomes impossible.

The web master gets more assistants but people start to lose their way among the welter of pages.

At this stage, in all likelihood, consultancy help is called for and the company finds itself steered towards one or another of the major CM systems.This is where the organisation is likely to get its first major shock.As Mark Berner, head of knowledge and content management at e-consultancy Rubus points out, CM systems do not come cheap.

“If we are talking about a sophisticated electronic retailing site with a rich variety of content, say a travel site with streaming media clips and a lot of travel content, you are probably looking at a spend of up to several million pounds,” he says. Less information rich sites will need to spend less, but a CM spend will always be six figures or more.

In other words, this is a mission critical piece of infrastructure, at least equivalent to a major ERP implementation.

Not surprisingly, given the scale of the investment, picking an appropriate CM system is a skilled job. As Berner explains, the various applications on the market tend to specialise in different things. Take enterprise digital asset management, for example, a necessity where there are large masses of content in document and/or video format. This is largely the preserve of the media companies. The issues here are how to recognise the value of an asset, how to re-purpose it and how to work it to generate an additional revenue stream. Some systems are strong at managing the flow of content through an editorial process. Some are strong at multi channel delivery.

“The CM space contains a number of different vendors, including the likes of Interwoven, Vignette with its Storyserver, Mediasurface, Broadvision and others. They all have different approaches and strengths in different markets,” Berner says. Vignette is good at delivering content to a production environment and then managing the delivery from there to a live audience environment. It has caching technologies which bring efficiencies to content delivery for certain kinds of applications.

“In this kind of environment what you are looking for is a highly extended version of what the IT industry thinks of as software configuration repositories.

You need version control, the ability to check content items in and out; the ability to pull content together in a site build, and to do pre-testing and previewing, and you want to manage the process of delivering all the elements to a preview area.”

Interwoven’s toolset, Berner says, is particularly good at managing the production process and providing the appropriate levels of quality control.This is “pure” content management, and has little to do with content creation or content management. He sees Mediasurface, on the other hand, as being particularly good at both content management and content delivery, with the delivery framework being an integral part of the management component.

Yet another set of players come into the frame when one moves to content personalisation, and the need arises for personalisation engines. Here, vendors such as ATG (Arts and Technologies Group) and BEA with Web Logic have strengths. As Berner says, there are now plenty of proven tools, and it is getting much easier for e-consultancies to build out large, dynamic content systems with personalisation. But companies have to realise the substantial costs associated with serious CM implementations.

“If you look at the hardware element alone, a normal spend for a specialist application server in this area is anything up to $50,000 per CPU. So if you are building a large system with, say, 15 CPUs, to handle substantial numbers of visitors to a site, the cost is very significant,” says Berner.

This being the case, organisations need to pause before they throw themselves into the CM game. Dynamic web content, and proper content management may be mission critical to a media company, but many e-consultants are now warning their clients against getting sucked in to becoming web publishers, simply because they want to run B2C or B2B sites.

Tac Leung, information architect at e-consultancy Scient, argues that many companies drift into CM by default. “Their real requirement is to enable a community of users with common interests to exchange documents, and there is no real need for any editorial supervision of those documents.

If you do not need editorial supervision, and a formal approval and clearance process for copy, then almost certainly you do not need a full blown web content management system,” he says. Peer to peer file sharing can be accomplished much more easily through an Intranet-based publicly available document repository, such as Lotus Notes, he says.

Berner agrees. Companies need to think very carefully about the scale of what is involved with web publishing and the integration of dynamic content with personalisation services, before they get into such an expensive game, he warns. “There are plenty of instances where good site design and a good first line of content is sufficient for a company’s goals.

Personalising promotions adds to a site’s attractiveness, but it can be overdone. In our view, employing good creative design and working out how a proposition is going to appeal to an entire audience is frequently better than going down a personalisation route. Thomas Cook’s site is a case in point. Good general design can fulfil a company’s requirements in many instances, and where this is the case, it is vastly simpler than getting into dynamic content management,” he concludes.

The moral of the story for organisations seems to be that CM is a pool you shouldn’t paddle in, unless you have a compelling, strategic reason for your dip. And if you do decide to take the plunge, be prepared to reach deeply into the corporate purse to fund your swim.

Tac Leung, information architect, Scient

“Today everyone wants their website to be personalised to the n’th degree, and they all want maximum ‘stickiness’, which is interpreted as requiring dynamic content. What companies need to come to terms with is that there is a very steep cost associated with being able to refresh data at high frequency rates in a controlled manner.

An organisation such as the has the resources necessary to run a dynamic news operation. The average enterprise does not, and when it tries, it finds that it is terribly under-resourced for the job. Moreover, what we tend to find when we go in and look at these sites is that there was never any necessity for so grand a project.

At Scient we now conduct substantial research into the degree of personalisation that companies really require on their websites in order to meet their corporate objectives. When companies have a sense of direction and an understanding of their markets, they can build a realistic system that can use feedback from user visits to the site to refine the vision, but there is no need to make the whole site switch direction at every customer mouse click.

The value of CM systems comes from the fine control they give a company over what is published. One of the real strategic goals of these systems is to provide a consistent corporate tone of voice and a consistent corporate look and feel. In many instances, and for many audiences, this will not be a desperately important criterion. Where you are looking to enable knowledge sharing, for example, within or between organisations, for a specific community of users, often peer to peer networking will do the job perfectly well. The important thing here is to get the bureaucracy and the control mechanisms out the way and let the ‘worker bees’ talk together at the level of detail they find appropriate. The Internet, after all, is partially about enabling communication.

Company lawyers, of course, get very worried about the liabilities associated with unmonitored “bulletin board” type communication and knowledge sharing.

CM systems are then seen as having the virtue of clawing back editorial control. However, what we are doing is building agents that can trawl through documents, looking for terms that make lawyers anxious. This provides an alternative, more free form approach to the problem.”

Nick Gregory, vice president, marketing, Mediasurface

“There is a real problem, we feel, that companies encounter when they use CM systems that do not separate content from the HTML page definitions.

This problem emerges at full strength when you want to carry out a radical overhaul of the site, either because you have made an acquisition that affects your offering, or because you want to rebrand the site, and give it a new look and feel.

When this happens, CM systems that do not separate content from the HTML page definitions, require a massive recoding exercise. The contents are stored in a database, the page definitions are assembled and the content is pulled in to pre-defined slots on the page. That works well if you keep the page definitions the same, but when everything changes, it becomes a major task. You have to go back to the toolset and recreate all the affected pages.

We take a different approach. With our product, if a link gets broken, because you’ve changed the page definition it doesn’t crash the page.

Businesses are now increasingly concerned about time to market, about the cost implications of the systems they are running and about return on investment. The Internet may be an opportunity but it is also being seen a huge money drain, What businesses want from web CM systems is to demonstrate an ability to control costs without becoming the cause of further cost.

We used NOP to carry out research for us in the summer of 2000. They found that nearly eight out of 10 business managers surveyed saw the content on their websites as either out of date or inaccurate. At the same time, around the same proportion agreed with the proposition that their websites were critical to their business. There is a huge mismatch here.

This is known as the dead web syndrome. Companies find that they cannot get control of their web content so, in effect, they walk away from the problem – despite the fact that it is a powerful way to reach their most important markets.

If you take a typical piece of news related content. The day it appears, you want it to be strapped as breaking news. Tomorrow you want to move it to a less prominent page, and then you want to move it off to the archives.

These are business rule related events triggered by time and the CM system should handle this easily. Currently, many organisations have a web master doing this by hand. The opportunities for errors associated with this approach are huge.

Our approach is to have a content repository and a workflow engine that sits on top of the content, containing the business rules. Authors create copy wherever they are. The business rules determine the approval process and manage the life cycle of the copy.

In our view, knowledge management is on a convergent path with content management. Most organisations will find that they benefit from bringing the two together and handling their KM needs inside the CM system.”

Costi Perricos, consultant, Arthur Andersen

“There is no doubt that companies face some steep challenges in refreshing and updating content on their websites. The fact is that anyone who wants a dynamic website falls into publishing, whether they do so with their eyes open or not.

Even if they do not originate their own content, they are still going to be faced with the task of aggregating it, editing it and publishing it to the website for their audiences. Actual media publishers face a larger challenge of course, since they have to find ways of extracting value from non-print channels.

CM systems such as Broadvision, Vignette and Interwoven are ideal for this at the top end. These are large, scalable solutions with implementation costs of upwards of £200,000. However, there are smaller solutions. If a company only anticipates around a thousand people visiting the site, a hand made solution, using a Macromedia tool, for example, could do the job perfectly well.

People need to realise though that the issue of getting the content in media independent format is not as easy as it sounds. Many creative studios use Quark, for example. Until recently, this was a very proprietary, medium unfriendly format and extracting information from Quark to put it into a content repository was no easy task.

A second point is that print is just a flat blob of text. There are no semantics in a print document, even when it is digitised. You need to extract meaning from the content. What we mean by this is that in print form an e-mail address is just writing. On the web it is a URL that you can click on and activate a process. Similarly, when you transform, say a shopping review story into a web story, you want to add things to it.

You want the user to be able to click to get a map locating the shop, you want to search for reviews of products in the shop. All these additional elements make dynamic content very powerful, but they all have to be resourced and managed.

One of the key things that we stress to our clients is that when you install a serious CM system, it may look like an IT project, but it has implications for all the elements in the organisation involved in any way in content generation, or with an interest in the website. You need to think through the training and cultural implications.

On the positive side, you don’t need everyone in the organisation to be a skilled web creative person. What you do need though, is the editorial capability to react to breaking news instantaneously, as it happens. You have to change your publishing cycle to an instantaneous cycle, where this makes sense for your organisation.

We remind the clients too, that the web is a dynamic world. They do not have to host everything of value on their own site. You can form affiliate agreements and swap visitors with other sites. On the web, content and commerce collide and feed off each other.”

Darren Lee, consultant, NextPage

“We are an enterprise software company primarily focused in the content and KM management arena. Our interest is not in dynamic web content, but in the issue of content generally across an enterprise, or between an enterprise and its customers and suppliers. We have focused on what we call peer to peer content networks.

The idea here is that instead of consolidating all information into a central content repository in a CM system, you leave the information on whatever server it happens to be. This way you can leave the original author of the document in control of amending and refreshing it and you do not have to re-index and update the document from some central site.

What you need to achieve this is software that holds a central index of where all the documents are. Users can then jump from that to the document, which may be anything from an invoice to a technology white paper. The management software ties all the data sources together and provides a single view of all the information across what may be a multi company set up. No one has to centralise the documents and there is no central hosting job.

We are seeing a huge uptake of this approach from the big accountancy firms, for example. They can share information with their customers. They can give customers secure, protected access via the web to the work being done on their accounts, to billing information and so on. The principle here is that if my information assets are tied to yours, we both get very good value, and the relationship gains ‘stickiness’.”

Related reading