What Can You Capture? September 25, 2009
Posted by Chuck Musciano in Random Musings.Tags: Best Of 2009, Information, Knowledge
trackback
There is a lot of discussion these days on the topic of “knowledge capture.” Recognizing that there is a lot of intellectual capital locked in the heads of employees, companies seek tools that will make it easier to find, categorize, and ultimately share this information. This is particularly important in industries where an aging workforce is retiring and taking a lifetime of knowledge with them.
As wonderful as it sounds, I don’t think knowledge capture is possible. Certainly not in the sense that it can be published and managed like a vast database, to be perused by those not yet enlightened. We are trying to treat knowledge like data, and that’s just not possible.
We have no problem capturing data. Data is nothing more than facts, collected and indexed. We have data streaming at us all the time, and computers are great at grabbing and storing all that data. We also have great tools for searching all this data. Google is the pre-eminent example; given a vast database of facts regarding which web page contains certain terms, Google will quickly tell you which pages match your query.
Stepping up from data, we have information. Information is data, correlated. Thus, a collection of temperature readings across the country is simply data. Those readings, correlated to produce areas of equal temperature, become information that will expose patterns that the individual data elements cannot disclose.
We have many tools to turn data into information, loosely grouped into that class of applications known as “business intelligence.” Depending on the usefulness of the tools, the correlated information may prove valuable or useless. Google is notoriously hit-or-miss on this; searches for certain phrases can yield bizarre search results, because Google cannot extend a clean semantic model onto the queries it services. For example, this blog gets hit on a daily basis by searches on “Scarlett O’Hara” because I once wrote one post about her. I am pretty sure that those seeking Scarlett are not looking for IT management advice, yet Google keeps delivering that result.
Google keeps making that error because it lacks knowledge. Knowledge is the result of placing information into the hands of someone with relevant skills and experience. With knowledge, information becomes useful. In our temperature example, the information represented by the temperature gradients allows a trained meteorologist to figure out when it might rain. In the hands of a lay person, it is just an interesting picture.
You can argue that some knowledge can be captured; that’s why we have books. I’d agree, but define what was captured as “shallow knowledge:” the part of knowledge that is close to the information, that involves repeatable activities, easily described. I can read a book and learn calculus, but knowing how to apply the calculus to a problem is far deeper and more intuitive. That’s “deep knowledge:” the complex ideas that knowledge capture seeks to document and categorize.
Regrettably, deep knowledge is almost impossible to capture. Consider the knowledge required to drive a car. Could you write down everything you know about driving? Could you even say it all? Absolutely not. The core of driving, what makes you good at it, was learned by practice and observation, absorbing lessons that are unspoken.
So it is with all deep knowledge. Imagine trying to write down how to negotiate with a vendor: reading body language, verbal cues, etc. You can’t; you learn it by watching a skilled negotiator. Even the skilled negotiator may not be able to express what they do; they just “know.”
The ease of data capture and the improvements in deriving information imply that the next level, knowledge capture, is possible and even easy. I contend it is not; while some rote knowledge may be captured, really important knowledge is only learned over time through continuous exposure to a master, much like a traditional apprenticeship. Our fascination with tools misleads us into thinking that there is a tool for everything; I don’t know that there is a tool for true, deep knowledge capture. I think that the real answer is one I’ve promoted before: know who knows, and learn from them before they are gone.
[tweetmeme source=”EffectiveCIO” alias=”http://bit.ly/cio109″ only_single=false]
The Effective CIO 
The issue of “knowledge capture” has been a topic for at least 20 years.
Some 20 years ago it was focused on building “expert systems.” Some did have success in initial processing of insurance claims, etc. There is also the Cyc effort led by Doug Lenat which continues to this day.
The “reuse” phase which led, eventually, to O-O classes was a technical version of this desire to capture knowledge and make use of it without reinventing it.
And, at least a decade or so ago, I heard numerous talks about how the knowledge of a company “walks out the door at 5 o’clock.”
So it isn’t just recently that people have recognized and look at ways to address the vast knowledge that exists but is inaccessible except through the narrow path to each individual.
And perhaps longer than this explicit search for ways to capture the information, people have talked about “corporate knowledge” and “how we do things around here.” These are other examples of the recognition of the collective information which is only indirectly accessible and disseminated/transmitted often over (relatively) long periods of time.
I do not dispute most of what you say, though. I just wanted to note that it is not some recent idea.
Well put. I’ve read somewhere that the human brain is the most powerful “computer.” Even with advanced AI, the human brain’s functions can never be duplicated by machines. And, through our experience, we develop instincts and reflexes. These can never be “captured” and handed over to the next person. The “tools” we have can never extract these deeply rooted knowledge that can only be obtained by going through a specific event, at a specific time, at a specific place, within a specific situation. There are too many variables involved to be represented on a flow chart or to be handled by an algorithm. Besides extremely gifted individuals or proteges, even a human apprentice can potentially fail to learn all the knowledge the mentor has to offer.
There is a tool for true, deep knowledge capture: our brain. It’s the universal tool that let all of us learn how to drive or acquire “deep” knowledge. It does so by identifying and classifying patterns of recognition. In an organization, the challenge of capturing business/corporate Knowledge (i.e. patterns of success) resides in the fact that captured Information (interesting) are not systematically processed, consistently evaluated, and publicly recognized as instances of success (useful). Otherwise, the staff (or to be more specific, their brains) would “know” how to solidify these instances into deep Knowledge over time.
Hi Chuck, Soctt Duncan beat me to the punch sharing the ideas that automating knowledge management has a long history and in my experience, a history of failure.
Apprenticeship — which has a history of success — is a solution to this problem. It’s not automation and thus considered to be expensive and “old world” by the technology industry.
People in the technology industry believe automation is the answer to everything. Abraham Maslow said it well, “He that is good with a hammer tends to think everything is a nail.” Best regards, -Steve
I wouldn’t say the history has been totally one of failure, just very few and limited successes compared to the great hope for capturing knowledge and automating its use.
I don’t think it is the technology industry alone that finds the idea of apprenticeship “expensive and ‘old world’.” I’ve worked in publishing, government, agricultural products, financial services (as well as technology). None of them had deliberate apprenticeship programs, i.e., ones the company backed and employed to capture/spread knowledge. I think they sort of expected this was happening somewhat automatically in an “of course people are helping/telling one another” spirit.
However, doing apprenticeship deliberately does take time (which translates to expense). But the main issue, in my view, in any formal apprenticeship effort is getting agreement on what knowledge it is that will be passed along. The formalization of apprenticeship (and any automation approach) can only go so far, it seems, for that reason. But even informally, it takes time and some deliberateness on the part of people to consider what it is they feel worth passing along as key to the organization.
So while I like the apprenticeship model, I think resistance to it could be greater than some “project” to try to automate knowledge. Informal things can be marginalized and ignored more easily than formal ones. I think some level of apprenticeship formality is needed in terms of overt organizational support/coaching for it happening. I would not see a formal “program” that tries to mandate how everyone would do it being successful.
Great post, Chuck, and great comments as well. We live in a work culture that prefers engineered, technological solutions to social ones. We also live in a work culture that devalues manual work, which is where most apprenticeships are. And we have adopted the Taylorist tenet that people are, essentially, interchangeable parts. No one admits this, of course, but it is clearly the Theory in Use in most organizations.
But when you’re dealing with knowledge, you’re dealing with something that doesn’t simply uncouple from people. As a practical matter, part of knowledge is relationships. As a practical matter part of knowledge is judgment.
I’m with Scott. I remember some of those expert systems. They worked, but only up to a certain level of complexity. We could identify what to do if gauge A went to red. We could even offer guidance on what to do if that happened at the same time as gauge B went red. Much beyond that, the complexity began to stymie us.
We were stuck because of something every accident investigator has experienced. Most plane crashes, hazmat spills, or plant explosions don’t happen for a reason. They happen because two or more things happen at the same time and set off an event cascade that’s beyond the capacity of any system to predict.
Apprenticeship is a process to develop knowledge in its richness. But it’s a process that swims upstream against the quest for the quick, the inexpensive, the technological and the easy.
So far we have only one proven knowledge capture system. It’s inherently non-plannable. We call it culture.
[…] decisions. For a clear explanation of the difference between the two, the readers are urged to read What Can You Capture? by Chuck Musciano on his blog The Effective […]
Chuck,
I’m fine with data capture over knowledge capture. That way we’re safe from the computers realizing themselves and taking over the world. 🙂
While I am in total agreement with Chuck and the other estimed commenters that automated, data-driven methods cannot be used for knowledge capture, I am not sure that the apprentice-master model would work neither in today’s context, as pointed out by Scott and Wally.
What I am suggesting is that there is an inexpensive way to capture knowledge, similar to the way our brain capture information and transform it into knowledge (you can read more at the reference in comment #7 above).
I think Wally is much closer to my view when he said Culture is one proven yet un-plannable knowledge capture system. The difference, in my view, is that such culture must be conducive to learning, i.e. explicit in what information must be captured and preserved, and supportive of the knowledge acquisition process by all employees (based on their roles and responsibilities). That way, the problem of limited apprenticeship is eliminated and the corporate environment (reflective of the culture) is quite plannable.
Rather than “Knowledge Capture” I prefer to think in terms of “Knowledge Transfer”. After all, you can have the best capture system in the world, but unless you use what you’ve captured, you get a very expensive write-only database!
Many industries, from medical internships to military training recognize this, and so do a form of “micro apprenticeship”. That is, they don’t just send someone to class and then let them loose on patients or the enemy. They follow a “Hear one, see one, do one” process, and the first “do one” is often under the close supervision of a more experienced peer or supervisor. The knowledge is not considered learned (ie transferred) until the sequence is complete.
When was the last time you came back from a class on a new IT technology or technique and got to observe or practice those new skills before being launched into a customer project?
JohnD – Yes…good change in terminology. Using “capture” can sound like trapping a bug, enemy, a cold, etc. or, at the least, permanently encapsulating something that has a life outside the “capture” and changes while what you’ve caught ages, perhaps dangerously. Using “transfer” sounds more realistic as well and suggests it can happen both ways and asynchronously between multiple parties.
Of course, formalizing “transfer” is like training, education, and mentoring. The latter is another way of saying “apprenticeship,” though I like “apprenticeship” more given “mentoring” programs I have seen that don’t go much further than showing you around the premises, giving you policy manuals, etc. But I think, for somewhat more “senior” people, “apprenticeship” sounds like reducing their status. It would not bother me, but I know the word suggests subordination of a junior person to a more capable one.
But “transfer,” however else we term it, is really better than “capture.”
Congratulations! This post was selected as one of the five best independent business blog posts of the week in my Three Star Leadership Midweek Review of the Business Blogs.
http://blog.threestarleadership.com/2009/09/30/93009-midweek-look-at-the-independent-business-blogs.aspx
Wally Bock
There are so many roadblocks to capturing or transferring knowledge in the business world. Management may not understand or support the initiative, employees may feel like they are giving perceived competition a leg up by sharing what they know.
Our PhaseWare Files blog has a nice series of entries about knowledge management, what is takes to get knowledge shared and how KM initiatives fail.
http://www.phaseware.com/PhaseWare-Files-blog/bid/20690/8-Reasons-Knowledge-Management-Initiatives-Fail-4th-in-an-Occasional-Series-on-KM-and-Social-Media
http://www.phaseware.com/PhaseWare-Files-blog/bid/15903/Yours-Mine-and-Ours-Knowledge-Sharing-Insights-No-2-in-an-occasional-series-on-KM-and-Social-Media
All excellent messages here. A critical dichotomy in this discussion is the fact that [based on Polyani’s differentiation of tacit vs explicit knowledge] only tacit knowledge [which cannot be directly cpatured] ever makes it to the workplace, to the point of application, while the only form of knowledge that can ever be exchanged is explicit knowledge, the artifact that is created when we try ot express/represent the tacit [idea] knowledge. All explicit knowledge is, of necessity a diminishment and an inaccuracy – the map is not the territory. But this elicitation of artifacts, and the highest possible level of fidelity available, is critical to the “transfer” of the original tacit knowledge. While the idea that I carry in my head can never be “transferred” to another person’s head [it will always remain in mine], with knowledge transfer, we attempt to replicate a facsimile of our ideation so faithful and complete that another can develop a similar ideation from the exchanged artifact. This doesn’t always work, but it is the most reliable vehicle we have thus far developed.
To converse more on this topic, please feel free to call me – I LOVE talking about this stuff. Or tweet me at galenmcp.
What a great conversation. Thanks to all for making this such a valuable exchange! I’ve found the key to blogging: suggest a topic and get out of the way of the experts.
I like the concept of “knowledge transfer.” “Capture,” I suppose, is like creating a backup tape that is never used to recover the data. “Transfer” better captures the complete activity, delivering knowledge to another person. (Thanks, JohnD & Scott)
The distinction between tacit and explicit knowledge is very helpful. I think this clarifies the real problem; the difficulty lies with the tacit, not the explicit. (Thanks, Galen)
I do believe that apprenticeships, whatever the size or shape, are the key. This is an overlooked area these days, and I’ll be writing a bit more on that in a few days. (Thanks, Wally, Scott, Steven, & Long).
I’ve learned a lot from all of you. Thanks for contributing!
What a great post and ensuing conversation. I think one of the reasons knowledge capture/transfer doesn’t deliver on it’s promise is because we have been unable to capture the thinking and conversation that transform the information into knowledge as it is applied over time. Apprenticeships are a great structure for that critical aspect of knowledge transfer. Yet I am beginning to believe that social media tools such as internal blogs and yammer can provide a way for more people, not just the few chosen “apprentices”, to listen to and learn from the masters in an enterprise. Wondering what your thoughts are about that.
Consortium for Service Innovation: http://www.serviceinnovation.org has some neat ideas regarding the capture of knowledge (within the help/service desk environment). Companies such as HP, MS, Marriot, Intel all participate in developing the KCS methodology with the Consortium.