KM Essays

Just-In-Time Information Management

by Dave Hollander
Revised: Tuesday, January 04, 2000

These Essays are inspired by phrases and passages I highlighted while reading Thomas Stewart's book, Intellectual Capital  (ISBN: 0385482280). I consider this book a must read, not only for those in information/knowledge management but for anyone who finds value in good books discussing business trends.

These essays and quotes should not substitute for reading the book. There is a lot of expertise captured in the book. The focus of this work is to capture some of that expertise and find ways to put to work for HP not to explain it. If you must, look over this work and maybe it will convince you that the questions left unanswered here will make the book worth the read.

Just-In-Time Information Management

Currently, most HP employees use an information management technique that can be described as "Just-in-case" information management. Information is kept in personal spaces (email, hard drives, file shares, webs) just in case it will be needed later. Indeed, I am personally a spokesperson for this process: I have two websites with over 1000 documents, 75 Mbytes of stored and indexed email, hundreds of megabytes of files and a desk and two bookshelves full of paper. Why? Because, as a knowledge worker with limited long term recall skills I use parts of this data set daily.

Does my system meet my personal needs? Yes. Can the processes be used by others? Not easily. Is it cost effective for HP? Yes. Can HP increase the value of these information assets (my data and myself) with some supporting technology? Yes. This essay will explore some of the issues surrounding providing employees with the information management facilities that will help them make a change for the better.

One alternative is "just-in-time" information management (JIT-IM). Chuck Sieloff and others have begun to move us in this direction. (ex phonewin ...) As the name implies, the JIT-IM approach strives to make the right information available at the right time and the right place. One key to successful JIT-IM is for information users to foster zones of deliberate ignorance. [1] If I can ignore information that is not needed by me at this time, then I have more resources to put into the information that is relevant.

Just-in-time implies just when our own stream of consciousness is ready for the information--how does our brain do it? The human mind does an incredible job in processing information from its receptors. Large portions of the brain are dedicated to processing the real time information, analyzing the signal and recording the results. One of the primary responsibilities of these lower brain functions is to notify the higher processing centers when an event is detected. The higher level processing centers can work on what they want confident that they will be notified when something warrants their attention. How can we emulate the processing done by our lower brain centers in our information infrastructure? What would be the characteristics of this infrastructure?

  1. Events. Email or web push could deliver "information events" similar to the signals delivered by our  visual cortex to our higher brain functions.
  2. Filtering. A human will wake to a baby cry but sleep through a barking dog because our hearing processes are trained to detect what is important to us.  In a successful infrastructure, the incoming information stream must be filtered and significant events identified. The cortex must use a combination of business strategy and personal strategy to establish the context for identifying events.
  3. Memory. Just as we have long and short term memory centers which we rely on when processing events, the infrastructure must include reliable data storage. The stored data must be organized according to our business strategy and personal strategy and it must be available when I need it.

Far fetched? Not really. There are some good examples in common use today. For me, the W3C list server is a great memory aid. As a member of the XML Working Group, I would have had to save thousands of messages to be able to access to them at a later date. With the list I know where the messages are, I know I can always get them, they are organized and I can reference them.

I wonder if the balance between messaging, the web and caches can be identified by further thinking about the short and long term memory centers? More research needed...

[1]Stewart, p 134

Imagine A Factory

"Imagine a factory inside whose walls is everything necessary to make a product--machines, component parts, other raw materials, safety glasses and hard hats, testing equipment, forklifts, the works. But suppose this stuff is heaped and scattered about the building without rhyme or reason. Parts are never counted or sorted; every time a worker has to bolt the gearbox onto the product, he must leave his post and search for ten minutes before he finds the box of bolts and three minutes more before he finds the right one; bins overflow with random jumbles of spare parts and components; the testing equipment is a city block from where the finished product off the line-if "line" is the word, for the plant's arrangement seems more like a plate of linguine than a line. Half-finished examples of discontinued products lie all over the building, and navigating a forklift through the mess is like driving in Manhattan on the last shopping day before Christmas. Trash is never collected. Instead, every few months a bulldozer drives through the plant, pushing out the door whatever lies in its path."  Stewart, p109

Discussing this paragraph is a bit like gloating after a 49-0 football game--redundant and insulting. Regardless, my own personal information assets and those of most organizations I have worked with at HP are managed as described. Why? Why does managing intellectual assets elude us when we all can see the described insanity and would not tolerate it?

I think it has to do with the intangible nature of information. I can't see it pile up until a disk drive overflows or I have so many folders and files I can't find anything any more. At that point, we only react to the problem, find a way to free up some percentage of the disk space and go about our daily task. Even seasoned system administrators who proactively monitor disk space do not modify their or their users behavior to do meaningful trash collection.

Anyway, now that you have this image of our management of intellectual assets perhaps we can work together and redesign the factory.

Knowledge Management Goals

Traditionally, knowledge management focused on "capture and store": capturing expertise from humans and placing it into company owned structural capital wpeB.jpg (22876 bytes)repositories. The focus is on moving into a repository both:

It is easy to understand the motivation for this objective.  Structural capital is the intangible asset companies own outright--it's what managers can most easily control [p 163]. The belief is when they can convince a human to externalize their expertise, their tacit knowledge, it will turn that knowledge into structural capital. Then the knowledge and capital can be controlled by the company and will stay with the company when the employee leaves.

Unfortunately, this approach typically falls short of meeting the goal. While there are many reasons, four stand out:

  1. Humans and their organizational systems seldom provide the necessary motivation to perform the task. These systems seldom successfully manage the behavioral changes necessary.

    This is a classic problem for knowledge management projects; we all know people who only share their expertise when they can get "credit" for it and we all know of groups whose reward structure is a disincentive to sharing. Indeed, the movement toward performance based organizations pay a premium to those who share only when their performance will be increased.

    In the distant past, middle management played a key role in assuring that vital knowledge was applied where needed. They assured that the appropriate assets were applied to the task at hand. With the delayering of the organization structure, the movement that included the laying off or reassigning of significant numbers of middle managers, this knowledge asset management task has been subject to experimentation or neglect.

    What remains is an "adhocracy", a literal and figurative network that connects people to people and people to data. The fast and easy to use network enables virtual teams to form as needed, to develop an ad-hoc structure to address a task. To be successful, an Adhocracy needs to find a way to restore the perspective delayering might destroy [p. 187]. It is not enough to provide a facility that allows project managers (task owners) to "troll for brainpower", it must manage the resource to assure there is a fertile place to troll and reward those who are caught.

  2. Valuable information and intellectual capital is seldom static.   Someone needs  to keep it fresh and meaningful.

    Many KM tools provide rigid information hierarchies or rule sets for organizing and managing information (and hopefully the capital captured therein). These structures suffer from several limitations:

    One approach to solving this problem is to design programs to provide human resources  to keep information fresh. PSO's Structured Intellectual Capital (SIC) effort not only addresses responsibility for structural knowledge, but also linkages to and from human capital. The second focus should help to assure the structural capital will be maintained.

    Perhaps the the most crucial task that we could give our information stewards is to take out the trash. Amid information overload, the value added is the information subtracted; that is to say that an intellectual capital resource is more valuable when the expertise is captured in less information. [pps 170-174]  For example, of the 300,000 instances of the word server indexed by our intranet search services why should the pc-proxy home page created and 1994 and made obsolete several months later vie for our attention? Data reduction was another value that used to be delivered by the layers of management and we need to find an effective replacement. If we reward people for creating information, shouldn't we reward them for deleting information?

    Typically and unfortunately, the value that an information steward delivers is either to low or not clearly understood and ultimately the resource is reassigned. In the publishing world, the steward role is performed by the editor-in-chief or publisher who typically receive compensation proportional to the success of their publication. In industry, we need to find a way to reward our people assigned to information stewardship proportionally to  the secondary successes their effort enables.

  3. Another approach would be to use adaptive technology instead of static rule sets. I have not seen tools that do this, but I would love to. What I have seen is how search engines have adopted to this situation. Verity, for example, uses a rigid topic structure to make search results more relevant. When the topic is well defined and maintained the results can be excellent. Verity's competition created automated algorithms that dynamically generate structure and the results are seldom as relevant as Verity's with a good topic structure. But over time, the results are better because the topic structure goes out of date or is never completely delivered. As adaptive technologies emerge, compare them to static rule set based tools based on their long term value, not just a best case analysis.

  4. It is extremely difficult to capture enough of the context to make information useful.

    Without context, information can never be intellectual capital. Even patents, the most traditional intellectual asset, are useless with out context--could you deliver value to a customer by reading a patent on electrophoresis? Gregory Kenning, holder of patent number 5717602 and some members of our Chemical Analysis Group can but most of us lack sufficient contextual knowledge to put that information to use. And, no system short of a full University, is likely to be able to hold sufficient context to help us turn this structural capital into customer capital.

    There are several tools and techniques which  add wpe6.jpg (36145 bytes)context to increase the value of an information repository. However, as long as these tools are applied to the problem of capturing instead of augmenting human expertise they will satisfy only the most simple of the problems facing today's knowledge powered industries.

    What is called for is to consider what we all know and most companies recite as a mantra: our employees are our must important asset. Consider changing the design focus of a KM project to develop structural capital repositories to help humans find the knowledge they need now. Instead of trying to capture tacit contextual knowledge, provide tools that help you locate the human with that knowledge.

    To create such as system, the first thing to consider redesigning is the context structure. Typically, structured capital repositories use a fixed hierarchy to establish context for each information nugget. Typical are frameworks like department or project names, technology or product categories. These hierarchical structures were created from a capture and store perspective. While seldom used for organizations anymore, the classical hierarchical organization chart is still used as the model in some information and knowledge management projects! [p 123]

    A structure designed to deliver, not capture, intellectual capital would probably look more a like a network of interconnected ideas. Peter Drucker explained in his seminal lecture, Knowledge Work and the Knowledge Society, knowledge work, "requires that people learn and preferably early how to assimilate into their own work specialized knowledges from other areas and other disciplines." A network structure can help create the associations that will lead someone to a different discipline where as a hierarchy forces you to only classify within your own discipline. The proposed repository structure would be less about capturing information and more about helping people find people who have the specialized knowledge they need now.

    Central to this approach is to keeping intellectual capital with the human. Humans are very apt at maintaining their expertise, especially when motivated and given the necessary opportunities. 3M Labs has built organizational structures which focus on building up the expertise in their employees. Then, the leverage that expertise by giving  project managers from their   business various techniques to "troll for expertise" and "call for reinforcements". The organization structure provides many of the instruments for managing human capital that exist in our entrepreneurial system, including accounts for storing credit accumulated by working for business teams and venture capital boards to fund speculative ideas.

    In order to really allow humans to focus on building expertise, we need to create organizational structures that make it easier for humans to tap into the structural capital and facilitate socialization (human to human exchange) by easily locating expertise. Then, we need to significantly reduce the transaction cost of using this expertise. Today, outsourcing makes it easy to use an external resource, yet there is support for "renting" an internal resource.

  5. Finally, the analysis leaves the most important intellectual asset unaccounted for: customer capital.

    Customer capital is why any of these activities are important; wpeA.jpg (29757 bytes)generating it is the primary mission of any commercial enterprise and is strongly related to objectives 1 and 2 in HP's Corporate Objectives. In pursuit of profit, companies try to turn human capital into structural capital. Paradoxically, however,  structural capital is what customers care least about. From a customers perspective, those structures are best that obtrude least. (p 163)

    It seems intuitive to us in this era of "One to One Marketing" that customers are buying a relationship. Relationships are formed between humans, not with structural capital. Since the organization is responsible for the relationship, or at least the profitability of it (see next essay) it is imperative that the organization uses its resources to assure that its employees are representing the company. This can only be achieved by assuring that the employee is a willing and enthusiastic participant in the organization. KM efforts focused on making structural capital a tool for the human should help the company be perceived as a valued partner.

  6. I believe that efforts that focus on humans leveraging structural knowledge, not creating it, will have a greater impact on the corporate bottom line than KM efforts using traditional approaches.




Power in an Information Age Company

wpe9.jpg (13788 bytes)In the last chapter of the book, Stewart makes some interesting assertions about careers and power in an Information Age Company.

"For the Information Age manager, power doesn't flow from position or budget, but from subtler springs, as intangible as the assets that create value." He goes on to list:

  1. Expertise - in project management more than a technical field
  2. Reputation
  3. Honest salesmanship
  4. Entrepreneurship and negotiating ability

He concludes the chapter with, "A specialist should be willing to bet on the long-term value of his specialty, do whatever it takes to be among the very best, and take the risks and seek the rewards of a more entrepreneurial career."

As I read these passages, I drew the diagram on the right--I find it interesting and compelling to see a tri-lateral balance of power, just as in our US Federal government. In our industry, and at HP, I think that project power has been hidden.  Snakes, was a prime example of project power at work. At HP, I think project power has the slogan "we will just get it done" and is part of what has set HP apart.

If you believe balance is necessary for organizational success (if not, read the book), sustaining a balance gives organizational imperative to knowledge management. KM activities are necessary to meet the needs of both the project team and individual.

Project teams form to meet a need then dissolve. Supporting them requires:

Expertise is the primary asset of an individual, it is the goods that Me.Incorporated has to sell.