Taxing your brain on management taxonomies

Some of those who think and write about law departments create taxonomies to classify what they study. A taxonomy creates superordinate categories that are similar on a number of different underlying dimensions. “Structure of legal departments,” for instance, allows a taxonomy that could include centralized, decentralized, hybrid, global, multi-level, etc. Or, “Staff” embraces lawyers, paralegals, executive assistants, clerks, administrators, clerks, contractors, interns, and more. The Association of corporate counsel categorized law departments along a number of dimensions (see my post of June 5, 2011: “traditional” vs “proactive” firms.). Maturity models also try to create taxonomic description (See my post of Aug. 10, 2011: model proposed for outside counsel management.).

Good taxonomies clarify; poorly done, they retard progress. As critiqued in the Acad. Mgt. Rev., Jan. 2012 at 83, some taxonomies have limited usefulness. They might create too few categories, notably just two, which ends up with sweeping alternatives that are too broad to allow precise understanding and powerful prediction. If someone sorted all software used by in-house legal teams into “productivity” or “information flow,” that taxonomy would run afoul of this weakness.

Second, some categories are either/or, dichotomous, when that is really not a clean distinction. For example, “progressive law firm” versus “conventional law firm” grossly characterizes and lumps a huge array of firms.

Third, if the taxonomy creates additional dimensions such as a 2 x 2 system let alone a 2 x 2 x 2 system, it is difficult to meaningfully label and find conceptual examples for each cell. The famous BCG taxonomy of portfolio companies had only four exemplars (cash cows, stars, etc.). If each taxonomic category needs its own name, this poses problems. Think of alternative fee arrangements along the three dimensions of quality enhancing, cost reducing, and transactional ease – very hard to label each of the eight permutations!

Next, the article says that even with a well developed taxonomic structure you are still left with a nominal measurement system. "Measurement scales are ordered in terms of refinement, from lowest to highest, as (1) categorical, to (2) rank orders, to (3) equal intervals, to (4) ratio scales. Taxonomic systems are expressed in nominal terms, necessarily implying that all people, things or activities within the category are equal in the characteristics that define the type. For legal departments, that would be quite reductionist.

Finally, even if you pick an ideal type for a category, the other things you assess there probably have a normal distribution around it. As a simple example, if one type of department in a taxonomy were “large,” inevitably other departments in that category would have some attorneys more or some attorneys less.