Celebrate the Success Stories of Knowledge Management - 2022 KMWorld Awards

How to be smart

We'd all like our businesses to be smart. But not smart in the bookish, horn-rimmed way in which you know everything except what counts. In fact, it turns out what we usually mean by being smart is knowing what matters. And knowing what matters is damn hard.

"What matters" is one of those loose but clear concepts that are indispensable and irreplaceable. It's slippery because what matters is so context-dependent. For example, what matters to a teen-ager is different from what matters to a baby. (Hint: The baby's demands are usually more rational.) And what matters to your business changes from day to day and workgroup to workgroup.

So, you can't tell just by looking at something that it matters. It matters only if you have some business interest that it addresses. And that's why it's hard for software to find what matters. For example, search engines can easily find you every document that contains the word "yak spleen," but they can't know that yak spleens matters to you--even though you don't know it--because it's been discovered that the coating of the yak's spleen has properties that would make it a great material for lining the wax lips that are at the heart of your global manufacturing empire.

That's why search engines are so bad at doing relevancy ranking, i.e., putting the 12,000 documents that match your query in order of importance to you. Instead, they rank them by how likely it is that the documents are about the terms you're searching for (and even that is a tough challenge). So the first 11,999 documents you're shown may be yak spleen recipes, and the document that matters most to you may be at the very end because it only mentions yak spleens once, in a footnote, as a low-friction material Yaklanders use to coat their figure skating blades.

Search technology alone can't solve that problem. Even if it solves its own problem of galactic magnitude (understanding not just what documents say but what they're about), that's only half the equation. To be smart, your knowledge management system has to also be able to ascertain your business interests.

The problem is that you don't know what your business interests are. Oh, you can copy some material from the mission statement and come up with some ideas, but you're interested in a much, much broader range of topics than that--so broad that you couldn't possibly write them all down (or build a comprehensive "topic tree"). For example, you may turn out to be interested in Malaysian politics if one of your suppliers is affected by new governmental regulations, or in transfractal numbers if a competitor has figured out how to use them to turn wax red.

So, how can software figure out what you're interested in? Well, it can watch what you write, what you search for, what you read and what you share. That may be enough over time to build a base that can be mined and used to make some guesses. Imagine the interior dialogue of a piece of software: "Ah, I see that Mary has written many internal reports on long-lasting, low-friction materials, and these reports have been widely read and shared. And here's a research paper on the Web that talks about yak spleens as a long-lasting, low-friction material. I bet this matters to Mary!"

Until the thankful day when software relieves us of all our burdens, we can take other steps to find what matters and thus become smart.

If building a KM system, don't focus only on retrieving information, sharing knowledge or making tacit knowledge explicit. Those are all either implementation "details" or actual distractions from the hard problem: finding out what matters.

You can't assume that what matters to your industry will be discovered first within your industry. So, encourage folks to explore widely. Let them roam the Web and count it as part of their job. Aimless browsing isn't the point, but healthy curiosity is. Enable communities of interest to form without restriction on topics (except where the laws covering obscenity and hate crimes apply, perhaps). The interests people have "outside" of work often are related in mysterious ways to their interests inside of work--because people generally aren't neatly segmented into two parts connected by a commute. These communities will develop knowledge that may turn out to be crucial as your company's interests change. In other words, what doesn't matter today may matter tomorrow, and your employees' interests aren't nearly as random as they may seem.

Scrub the talk of benefits as a way of explaining the value of your services and products. Think instead about why they matter to your customers. The difference is that benefits are put forward as Universal Goods and thus are often bland and vague (cut costs, increase quality, shorten time, yada yada yada), while explaining what matters requires thinking through your customers' real interests.

There's more to being smart than just knowing what matters. You also have to be able to learn as a normal part of business life, you have to have the courage to change (i.e., become what you're not) as rapidly as required, you have to appreciate the many styles and types of intelligence. But being able to find what matters would sure be a good start

KMWorld Covers
for qualified subscribers
Subscribe Now Current Issue Past Issues