I‘ve written a lot about Knowledge Management (KM), and recently especially about a revolutionary bottom-up approach called Personal Knowledge Management (PKM). The recommendations in my articles for instituting KM or PKM in your organization (processes, infrastructure etc.) assume, however, that you actually have an ongoing budget and full-time, permanent people to support your KM initiatives.
I’ve discovered, however, that many organizations have neither budget nor people specifically earmarked for KM activities. In fact, organizations whose IT people are principally project-focused, and which expect IT (or learning/HR) to look after KM matters, really have no way of adopting any of my KM recommendations because once the project is finished and the IT tool “launched”, the only ongoing operational people they can call on are in the help desk. And those people rarely know anything about KM.
So I thought it might be useful to develop a set of KM principles that can be considered at the design stage of any project, so the major KM landmines can be avoided. I’ve come up with these 16, which I think are the most critical. For each, I’m aware of at least one case where ignorance of the principle, or failure to consider it appropriately, has caused considerable grief to an organization:
The Objective of Knowledge Management (what drives these principles):
To enable users/customers to obtain critical and relevant context-rich information, and to connect and collaborate with experts and colleagues, easily and promptly, so that they can be more effective performing their jobs.
Information Content Principles:
- Content acquired and maintained should be driven by need, not availability: The information systems should obtain and accommodate information content that is needed – i.e. crucial to the performance of people’s jobs – rather than information content that happens to be available (notably newsletters, manuals and other content that was previously maintained or routed manually or by e-mail). There is a tendency to use information systems to house content that has been acquired that no one is quite sure where to put. As a consequence, most ‘knowledge repositories’ are cluttered with documents, messages and data that serve no critical need and simply get in the way of finding the information that is critical. Some organizations now spend almost as much effort in ‘content rationalization’ (getting junk out of repositories) as they do in content creation.
- Content needs should be those determined by users/customers, not suppliers/managers: As a corollary of the above, what is crucial should be determined by the users/customers of the information systems, not by the suppliers of information. This requires a program of continuous outreach to users/customers – a process of observation and ‘customer anthropology’ that enables the content architect to understand what information is needed by different constituencies of customer, and how it is used. In the absence of such continuous programs, at least a series of upfront interviews and surveys of different segments of users/customers should be undertaken to ascertain what their crucial information needs are. Otherwise, what gets stored in information system is the stuff that suppliers need a place to store – and that is rarely the content that has critical value.
- Content management systems should focus on personal content management and group self-management of content: Many studies of information systems have discovered that users/customers are unable to use knowledge and technology resources available at their own desktop effectively. Yet most content management systems focus on teaching users/customers how to use centralized repositories and tools, rather than how to organize and index the content on their hard drive, how to subscribe to and publish their content, and how to find it later – and, for collaborative groups, how to take responsibility and authority to self-manage their collective content and work-product in collaborative work-spaces.
- Information systems should enable and capture conversations: The reason face-to-face and telephone conversations are still the principal means of knowledge transfer in virtually all organizations is that they are effective – they allow quick, iterative, context-rich learning in a way that impersonal and asynchronous methods cannot match. The best information systems encourage conversations by making them easier to set up (see principle 12 below), and by capturing and sharing their critical content (by recording the conversation or by distilling its decisions, information, agreements and actions using a mindmap or other tool, and ‘publishing’ the result).
- Information management tools should be simple and intuitive to use: That often requires sacrificing power and functionality, but it has the advantages of obviating the need for training (that no one has time for anyway, leading to misuse or underutilization of many tools), making it easier to change as requirements change, and reducing development and maintenance cost and development time (many tools are commercially available and many of these are free). Often, if there is no simple, intuitive tool available to satisfy an information need, the best answer is not to introduce a tool at all, and use the existing, manual information process instead.
- Needs analyses should consider the ‘cost of not knowing’, not just the cost of knowledge: Although most of an organization’s information needs are best determined by interviewing and observing front-line users/customers, someone in the organization (perhaps someone in risk management) should also be assigned to look holistically at the cost of potential knowledge failures – what it would cost the organization (lost revenues, compensation for damages, damage to reputation etc.) if it failed to acquire, understand and communicate critical information (such as information on a competitor’s revolutionary product, or information about a critical weakness in the organization’s processes that could expose the company to major litigation or business interruption). Enron’s collapse, the extent of devastation from hurricanes and epidemics, and 9/11 were all consequences of ‘knowledge failures’ with astronomical ‘costs of not knowing’ that could have been prevented with an appropriate investment in knowledge.
Information Context & Organization Principles:
- Related content should be linked and stored together: Documents, messages and data that are related should be kept together (physically or at least virtually through virtual copies and links) to provide users/customers with as much context as possible about the information they contain. An overview of how the content is connected should also be prepared and maintained, using either a meta-abstract or an index or table of contents. Most document repositories, unfortunately, require ‘submitting’ isolated documents to centralized repositories, so the context for the document is quickly lost. If there is no one assigned to manage content to ensure related information stays connected, it may make more sense to leave the content on the hard drive of the author (where it is likely to be organized in folders with related information) and allow customers seeking it to canvass it (see principle 9 below).
- The indexing of content should be self-managed by its authors and users, using folksonomy and taskonomy, not imposed taxonomy: The use of folksonomies (where authors and users choose their own tags for content, and over time converge on agreed-upon, but ever-evolving tags) and taskonomies (where content is indexed/tagged by how it is most likely to be used by others, rather than by subject matter) often seems anarchic and inefficient to managers and lovers of taxonomy. However, taxonomies are rarely ideal for all or even most users and are hard to change, and they must be built and maintained, sometimes at great cost. Although there is a (self-)learning curve to folksonomies and taskonomies, once that curve is overcome these means of indexing are more effective and (because they are self-managed) easier to change and cheaper to maintain.
- Content should be ‘pulled’ from where it naturally resides, just in time, rather than ‘pushed’ to central repositories just in case: Most people are too busy to contribute, and insufficiently rewarded for contributing, information to centralized repositories, especially when this requires indexing and abstracting. There is also a natural risk-aversion to contributing specialized information that might be misused by others lacking the skills or context to use it appropriately. By leaving it on the site of the author or group that ‘owns’ it, and putting in place canvassing systems that can locate and retrieve this information, in context, on a just in time basis, all these obstacles are overcome. The use of weblogs or other simple personal workspaces where each user can collect and manage their personal shareable content can also help in this process. Where security or risk of misuse is an issue, a permissioning system can be used to restrict access to certain documents, messages and data, or to allow users to learn of the existence of relevant information but only access it with the explicit real-time permission of the author(s). This can be as simple as putting a process in place on each user’s PC that asks the users/customers, whenever they store or send a document, message or data, which people or groups it may be made available to, and then harvesting the information and the related permissions on a regular basis to create a centralized copy, or canvassing all users’ hard drives on a just in time basis using standard search templates, subject to that permissioning, to respond to searches for information in real time.
- Information should be provided in the context of existing business processes, not as something apart from them: Much of the information that is sent to and made available to people (notably that contained in e-mails) is outside the context of what people do in their jobs and disconnected from the systems and processes they use to do them. That means that they need to ‘set aside’ their regular work to process this information. To the extent that information can be embedded within the systems and processes people use regularly, that information becomes an enabler of work effectiveness instead of a distraction from it.
- Information processes should encourage users to add value and meaning to ‘raw’ information: Much of what users/customers do with raw information is digest, synthesize, distil and interpret it, but the value they add by doing so is often not captured and shared with others. Weblogs and other simple personal workspaces can encourage and enable users to record and share their interpretations and insights. Librarians and other information professionals can be trained to do this as well, using visualizations, systems thinking, ‘pyramid principle’ analysis, story-telling and other techniques to make information easier to digest and grasp and more meaningful, rather than just ‘ripping and shipping’ raw data to others in the organization.
Connectivity & Collaboration Principles:
- Information systems should enable users/customers to find needed experts and connect with them expediently: The most significant KM challenge in most organizations is not finding content, but finding expertise – who in the organization, or even outside the organization, knows more than anyone else about a particular subject. The need to find experts is often urgent, resulting in lots of time wasted hurriedly following chains of suggestions on “who might know about that” until the right person is identified, and then telephone tag connecting with them. Information systems that simplify the identification of experts, help users discover the most effective way to contact these experts (including using new technologies like presence-detection), connecting with them through various simple, intuitive, real-time virtual presence tools, and scheduling appointments with them, can offer an enormous ROI.
- Subscription and publishing processes should be primarily self-managed and peer-to-peer, not top-down managed: RSS and other new standard technologies make it much easier to establish mechanisms for sharing information simply, peer-to-peer, both within and between organizations. Centrally-managed “pub-and-sub” systems require more maintenance and training than straightforward tools that enable individual users to publish their information content to those they think would find it valuable (and they probably know best) and to subscribe to others’ information that they find valuable. A large amount of organizational e-mail is currently used for ad hoc peer-to-peer publishing, because it’s simple and meets a critical need, but this is not an effective use of e-mail. The best mechanisms make publishing as easy and intuitive as putting a document in an out-basket, and subscribing as easy as adding your name to a sign-up sheet – the manual processes that “pub-and-sub” supplant.
- Connectivity and collaboration (including e-learning) tools should be simple and intuitive to use: [This principle is directly analogous to principle 5 above.] This often requires sacrificing power and functionality, but it has the advantages of obviating the need for training (that no one has time for anyway, leading to misuse or underutilization of many tools), making it easier to change as requirements change, and reducing development and maintenance cost and development time (many collaboration tools, such as wikis and Basecamp are commercially available and many of these are free). Often, if there is no simple, intuitive tool available to satisfy an information need, the best answer is not to introduce a tool at all, and use the existing, face-to-face connectivity or collaboration process or another face-to-face process like Open Space instead.
- Information systems should draw on the collective intelligence of employees and customers: The majority of KM systems are internally focused on capturing data from the front lines for management, and communicating instructions from management back down to the front lines. Such systems fail to capitalize on one of the organization’s greatest assets – the collective wisdom of its employees and customers. Tapping this ‘wisdom of crowds’ requires mechanisms to communicate the organization’s most pressing business problems and management’s information needs out to the front lines and to customers, and then canvass or survey these employees and customers to capture, in a focused way, the information, ideas and awareness that they can bring to bear to address those problems and needs. This is a world apart from the collection of customer data and the customer satisfaction surveys that most organizations use – it requires candour, trust, an ability to ‘qualify’ respondents, and careful wording of the questions posed to the ‘crowd’. But the insights and knowledge that can be gleaned from such an investment – and the employee and customer satisfaction that comes from engaging these stakeholder groups in the organization’s decision-making processes – is immeasurable.
Knowledge Behaviour & Culture Principle:
- Information systems should adapt themselves to the organization’s culture and information behaviours, not try to change them: Many organizations have learned the hard way that social systems are complex systems and culture and behaviours change very slowly and only when they must, not just because it is a good idea. Human beings have a way of resisting changes, and also finding workarounds when the processes and tools they are prescribed don’t do the job. It’s essential to understand and accommodate those workarounds in the design of information systems, rather than trying to impose new systems on users/customers who have already discovered or invented something that works – such impositions will inevitably be resisted. Likewise, repositories, tools and spaces will only be adopted if they respond to a critical need that is not already being addressed by existing systems and processes – even if the boss mandates their use. Many organizations (especially large ones) also deal with dysfunctional information behaviours caused by inappropriate processes (especially reward processes), office politics, lack of awareness, and faulty sense-making skills. Ignore or try to change those behaviours at your peril. It’s essential to know which behaviours are at work in your organization, and what the organization’s ‘knowledge-sharing culture’ is, and why, and then adapt information systems, processes, and tools to accommodate the behaviours and culture, rather than trying to change the behaviours and culture through executive mandate, or coercive or reward systems.
I’ve spoken to a couple of people about this, and they say this list of principles is not only useful as a KM effectiveness guide for IT projects and learning programs, but can also be used as a ‘scorecard’ on the effectiveness of the organization’s KM programs, by scoring your organization on the degree to which existing processes adhere to each of these principles. The organizations I know best don’t score verywell, especially on principles 6, 9, 13 and 15.
How would you score your organization?
very nice … there’s a book there, the chapters are already stated, and you will be able to find much of the rest of the content (to be adapted, for sure) in a number of your other posts over the past several years.It’s a necessary book. I’ll help write it if you wish.
Thanks Jon. I suspect such a book, and the ideas in it, would be a hard sell. We would need to find ‘success stories’ for each principle, and some of these would be hard to come by.
Fabulous list. Thanks so much for creating it. Just a comment on #8 (folksonomies vs taxonomies). I really dislike saying that one of these methods is better than the other, as they both have sizable pros and cons, and your mileage is going to depend on your goals. For instance, I’d point out that in environments where completeness is important – i.e. “I need to find everything we’ve ever done with XYZ Org” – folksonomies are inherently problematic. I think the real utility is in systems that combine *both* methods – to allow people the meaningful vocabulary and explorator aspects of a folksonomy, mapped to a taxonomy via a controlled vocabulary to reduce some of the inherent messiness of a folksonomy. No question, though, that any taxonomy is going to take money and continuing support from knowledgeable people to work well.