Recently, I presented at the Ark Group’s KM conference on using KM to advance the AFA game for law firms. Greg has previously posted on his reaction to KM based on the conference, and here I share mine. For my presentation I put together a case study from an actual exercise in building a budget from past billing knowledge, to use in setting a fixed fee. Per my case study, I note that “Common KM wisdom holds that analyzing past billings is a constructive effort for building new budgets to use in establishing fixed fees and other alternative fee pricing.”
This idea makes perfect sense until you actually attempt it. In the case study exercise, I went through a progression of effort, trying to isolate influences to establish some pattern or trend in past billing information. The first pass was analyzing like matters. The second was analyzing like matters from the same client. The third attempt, the primary focus of the case study, was analyzing the same case (from different jurisdictions), for the same client from the same time-frame (a product liability case). None of these efforts resulted in a consensus budget number. Instead wide variations in fee amounts and even timing of fees was discovered.
The KM lesson is twofold:
- First, even the best existing search and retrieve KM will not address this problem since the issue is not finding matters (that is a separate issue), but instead it is understanding the data related to them.
- Second, gaining useful knowledge about fees will require the analysis of volumes of poorly structured billing information. And this is where KM comes in.
But it can’t be the same old KM.
It can’t be the passive search and collaborate KM. It needs to be a new style of KM, whereby our technology is about understanding our knowledge. A recent example that demonstrates this thinking is ‘predictive coding’ from the e-discovery space. This new approach moves past search and into analysis. Instead of just trying to find relevant discovery information, predictive coding analyzes the data and proactively codes the information. This task has been almost exclusively in the human realm until this development. The Managing Partner of Squire Sanders presented on this concept at the same KM conference. He had even conducted an analysis, comparing human coding with predictive coding side-by-side to test the effectiveness. Predictive coding matched or beat human coding on every level.
This is Analysis KM - a new breed of KM that brings machine-enabled analysis to bear on our growing mass of information. The volume and complexity of our knowledge is so large, expecting humans to be able to understand and analyze it is crazy. So the answer to KM’s future lies in this new analysis direction.
PS - The picture on this post has very little to do with topic, other than thinking BIG. I just miss Moab is all.