The state of the art

Putting KM technologies in context

Eighteen months ago, ICL founded KM Technology Watch, a community of practice with members across the company’s 20,000 staff, devoted to sharing experiences and developing a comprehensive understanding of the KM technology market. Arif Azar and Justin Souter outline their views on the current state of the art, describing the tools available and providing a broad perspective of how to implement a KM technology strategy.

Today’s organisational reality – the internet hangover

The dotcom bubble may have burst, but the internet and its associated technologies have forced a reassessment of working practices, attitudes and future strategy. The internet’s promise of faster, cheaper, better, while nowhere near being delivered in full, has nevertheless changed the business world for good. It seems that we are all being asked to do more with less: less money, less people and in less time. KM is seen as being able to help deliver on part of this agenda.

However, because of the challenges of measuring associated productivity gains, those running KM programmes need a roadmap to get them started, and some rules of thumb to help choose and maintain direction.

We are assuming that readers are familiar with the main themes of KM literature and broadly concur with the consensus that IT budgets are becoming tighter in anticipation of tougher market conditions. We also take for granted that Windows will continue to be the desktop platform (although most vendors provide server software for at least one of the different UNIX flavours, most commonly Solaris, as well as for Windows). Furthermore, it is assumed that more specialised tools are a luxury, with mainstream line-of-business tools such as portals and basic search tools being the technologies easiest to justify.

The KM technology marketplace

If we turn our attention to the types of technology in this area, one quickly comes to the conclusion there are too many products from too many vendors chasing too few clients. This means potential purchasers need to be confident in both their choice of technology and their evaluation of a vendor’s survival in this particular market. We see the market split three ways: Lotus, Microsoft and specialist/niche vendors (‘pure KM’).

With K-Station, Lotus has brought a mainstream KM product to market: Discovery Server, which aims to locate not only documents, but also expertise. We believe that while interesting, this product has yet to fully mature and needs time to prove its worth. At the time of writing, IBM has announced it is merging K-Station into its own branded WebSphere Portal Server – it seems that a wider consolidation of Lotus technologies into IBM’s product range is underway (representing yet another factor to consider).

Portals allow content to be viewed in a uniform way. Out go the multiplicity of web page designs and page layouts. Users become familiar with where everything should be, and so can concentrate on their objectives. Portals present information according to organisational, workgroup and individual preference, provide a platform for collaboration and access to numerous data repositories (databases, ERP and legacy systems) through proprietary interfaces; products combining these features are known as enterprise information portals (EIPs). Bear in mind, though, that this is a presentation layer – think about packaged or separate middleware that allows you to integrate data from different repositories with other applications.

Microsoft’s approach had focused on delivery and collaboration, using offerings such as Exchange. Microsoft now has a more coherent framework with the release of SharePoint Portal Server (SPS) and Content Management Server (CMS). These two have been slotted into the .Net Server branding, and can be used in conjunction with Commerce Server.

SharePoint, which has been promoted as a workgroup-level product, offers proprietary search and entry-level document management functionality, with a portal-style front end using a ‘digital dashboard’. Once integrated with the personalisation from Commerce Server and the web publishing framework of CMS, it is not hard to see them being packaged as an EIP. Watch this space.

Although SPS is a ‘1.0’ release, the SPS product team are working with integrators and clients to address current issues, and we are actively consulting with customers to make this product the basis for their knowledge-sharing infrastructure. As such, it is a cost-effective solution from a credible vendor, which should provide a wider platform for organisations to support their business. Looking at CMS, it already has over 150 clients from its days as NCompass Resolution, but there is some way to go before full integration between all three products. We would advise that you wait for the next releases of SPS and CMS if this is an issue for you.

We have talked about Lotus and Microsoft. We consider these to be infrastructure products. The third slice of the pie comprises numerous ‘point’ solutions that tackle specific issues. Instead of going for breadth, they concentrate on depth.

Drawing from our own experience, ICL has numerous outsourcing deals, a number of which are very large public sector mandates. Across these projects, standardising on products such as Convera RetrievalWare (www.convera.com) addresses the need to provide a search engine to cover the intranet and other internal data sources. RetrievalWare has much of the search ‘workings’ pre-configured, which can be extended to create a set of predefined queries and a Yahoo!-style taxonomy (where subjects are grouped according to relevancy and users can navigate through this topical hierarchy). When making technology purchase decisions, we recommend you take into account the extent to which the core tool can be extended using features such as these.

Autonomy (www.autonomy.com) arguably has the highest profile in this area. It takes a different approach to that of tools such as RetrievalWare or Verity (www.verity.com), and indeed is not marketed as a ‘pure’ search engine. Autonomy technology undoubtedly has a great deal of potential – the ability to action content in different media, such as voice and video, could be a major plus point. However, our own experience with the intricacies of the system suggests that the task of planning and system ‘training’ (as opposed to RetrievalWare’s pre-configuration) that is necessary prior to roll-out could be a headache. Our take is that Autonomy has focused on getting the search algorithms in place, and now needs to spend more time on enhancing usability for system administrators. A recent tie-up with IBM suggests that Autonomy is developing a network of partners to bring its product to a wider audience.

Much of this will be familiar, as search is an essential component of any intranet or knowledge repository. After all, what’s the point of capturing and storing something if you can’t find it? Looking ahead, we see the next frontier for retrieval being centred on visualisation tools. The emphasis here is on making the black box return more of its secrets in a meaningful way – instead of confounding the user with Boolean logic, these applications express the results in context, thereby making it much easier to make sense of what’s returned. These are tools that create clusters of meaning, whether using words or pictures. We have already mentioned taxonomies (see also Semio [www.semio.com] and WordMap [www.wordmap.com], for example), but we have a personal preference for the topographical maps of concepts generated by tools such as Aureka Themescape (www.aurigin.com). The system analyses a complex data source and groups topics according to how often they appear – the most popular being the peak of the ‘mountains’. One can even click through to the topic itself through this interface.

Having split the market by vendor type, it is also apparent that KM has two dimensions: discovery and collaboration. Broadly speaking, discovery concerns itself with retrieval and presentation of information and individuals, whereas collaboration is about working with other people to productive ends. Collaboration helps turn tacit knowledge into something more overt, once the people and objects have been identified.

Having tackled discovery and retrieval tools of different guises, the next aspect to consider is collaboration. Our thinking is that while access is the foundation, the results generated by search facilities must be combined to generate further insight and hence real business value.

E-mail is probably one of the most common electronic ways to pass around information, but instead of clogging the mail servers with files, be sure to use groupware and public folder functionality from, respectively, Notes and Exchange, to maximise return on investment.

Once you have established the principle of working collaboratively, embed it in the organisation through softer initiatives like culture change programmes, training and directed leadership. This will hopefully lay the groundwork for establishing cross-functional communities of practice. With yesterday’s hierarchies being replaced by flat management structures, CoPs provide a way to challenge the silo mentality and allow knowledge sharing between far-flung people working in similar topical areas.

Kick-start such initiatives by enlisting the services of a player like www.participate.com, who will provide online communities via a managed service/ASP model. This has the advantage of concentrating expertise in one place. Also, consider tools like Sametime and Quickplace from Lotus.

Pulling all this together could be the job of a portal. Having considered offerings from the ‘big two’, organisations wishing to cast their net more widely should seek out products from independents like Plumtree, which bundles Autonomy retrieval software, Semio taxonomies and eRoom collaborative workspaces, to provide a complete presentation, retrieval and collaboration environment. Other vendors in this space include Hummingbird, Computer Associates and many more. Expect all but the best to leave the market, go out of business or be bought out – conduct a thorough due diligence process if you plan to make a portal the cornerstone of your intranet and knowledge sharing efforts.

Looking ahead, we believe a wider market consolidation is highly likely, given the condition of US and European economies. On the other hand, we see at least two different technology areas having the potential to solve issues not addressed at present. These are:

Peer-to-peer;
Mobile information.

Within the collaborative context, we see peer-to-peer as allowing more flexible and supportive working patterns, so that end-users can organise themselves into whatever configuration and organisational structure they choose. We view Groove (www.groove.com) as a leader in this field, with other peer-to-peer technologies including Microsoft’s NetMeeting, and even the venerable telephone.

Mobile information delivery seeks to distribute relevant information to mobile workers using wireless devices such as mobile phones, personal digital assistants, and various gizmos falling in between. Road warriors can be piped the latest headlines, cost structures and product news, while field engineers can be informed of their next appointment with directions about how to get there. Given that most digital devices are not presently subject to much co-ordinating discipline, the potential for creative synchronisation is substantial.

Implementation scenarios

This section attempts, at a high level, to outline a rough framework based on different types of organisations and situations that may be encountered. A good place to start is to examine what technology is already in place. At one professional services company we worked with, for instance, Lotus Notes and Mail was the collaborative and messaging infrastructure. Although we have found that Notes databases can present challenges for a search engine, using Lotus’s own portal tool and search engine is the logical way to go in this case.

An early architectural decision should be whether to build off-the-shelf technology, or else procure something that needs integrating with your existing infrastructure. Tools from Autonomy or Applied Psychology Research are based primarily around sophisticated proprietary technology, which need to be wrapped in user-friendly interfaces. Any cost model must include the price of engaging the consultancies that specialise in doing so.

Coming from a different perspective, the size of the organisation is also going to be important. A start-up or SME looking for discussion-group functionality could download the fully-featured evaluation version of Groove, while avoiding an expensive centralised architecture by using an ISP or Hotmail for external e-mail.

At the other end of the scale, companies the size of Ford and K-Mart (a US-based supermarket chain) have signed portal installations offering up to 200,000 seats. However, you may find it more cost effective to develop a digital dashboard, with components that can be downloaded from the web or created from scratch, which leverage common enterprise systems and business intelligence tools.

The key questions, therefore, are: does the chosen technology run on the hardware and software platforms already in place? Does it fit into the overall budget? Are there sufficient in-house resources to integrate and support it? How quickly will it bring a reasonable return on investment? We hope our deliberations will make your decisions in this area more straightforward.

Putting it all together

Almost without exception it will be the intranet that will provide your company with a knowledge-sharing foundation (even portals are viewed through a browser). A search tool of some sort is essential, even if it is only something like Microsoft’s Index Server. Most employees will be reluctant to download all they know into a system that could make them obsolete, so make them famous for knowing what they know by using a directory of expertise (for example, SigmaConnect [www.sigmaconnect.com] or Lotus’s Discovery Server).

Consult industry analysts for corroboration on portal vendors, then go with a product that has the right balance of price and performance. If you have sufficient budget available, you may also consider a taxonomy to provide structure for everything that isn’t in a database.

This list attempts to put in order of priority/affordability the types of tools we think should be considered for a rounded knowledge infrastructure, but ours is by no means the final word – ultimately, the decision will depend on how the jigsaw can be pieced to fit into the present financial and technical configuration within your organisation. Above all, remember that capture and storage are necessary, but not sufficient: you must be able to find that needle in the ever-growing haystack.

Summary

Readers should now have a better idea of what practical steps to take when implementing KM technology, and how to think about the issues given their own organisational context. Buying specialised solutions to solve particular business problems is appealing, but in reality most IT spend will be in areas that are easier to make a business case for, and going with recognised vendors helps soothe senior decision makers and should provide a safe long-term solution.

We believe readers should be thinking in terms of an overall framework or architecture to support a wider knowledge strategy, where the balance of effort is likely to be biased towards softer initiatives such as change management and training. In this context, technology is seen as an enabler, and not a panacea.

In order for vendors to survive the coming shakeout, we see a number of factors that need to be addressed: KM tools should be easily integrated with other products in order to provide clients with a rounded KM platform; set-up and configuration needs to be simplified; and, finally, vendors should consider providing the managed service model, to make their applications available to a wider audience (such as SMEs).

Readers should be thinking in terms of an intranet as their basic KM foundation, a search tool to find provide basic retrieval, and a directory of expertise to address barriers to knowledge sharing. With these in place, further investment will probably need to be targeted at a more sophisticated retrieval technology and a portal product to unify the navigation of corporate information. Good luck!

Justin Souter is a consultant within ICL’s knowledge management practice. He can be contacted at: justin.souter@icl.com

Arif Azar is a consultant within ICL’s knowledge management practice. He can be contacted at: arif.azar@icl.com

For more information on the issues addressed in this article, please contact Tom Knight at: tom.knight@icl.com