Here’s to using decision trees in the new year
- Subtitle: The Future Files
The concept is simple. The execution is not complicated. It just requires mental commitment, planning, and a form of projection. You lay out all strategies, decision options, and possible outcomes. For each option, you try to quantify the risk and uncertainty associated with it and, ultimately, calculate its expected value.
Do that and you will be able to choose and intelligently recommend to your client, based on the information you have at the time, the optimal strategy for a case or business negotiation.
A good definition of the concept is provided by Marc Lauritsen, president of Harvard, Mass.-based Capstone Practice Systems Inc., the developer of Choiceboxing software. As he stated in his article “A Decision Space for Legal Service Delivery” in the Harvard Journal of Law & Technology Occasional Paper Series: “‘Choiceboxing’ is a method . . . that helps people quantify and qualify options to reach more carefully considered decisions when selecting from groups of alternatives. It involves creating and sharing an online representation of a choice – one that functions like an intangible device.
Probably many of us occasionally use on our files one form or another of an elementary (let’s be honest — very elementary) mental decision tree. Possibly we run some of the elements at a subconscious level, rather than in any contemplative way. Hopefully most of you will agree that using decision trees in our practice is a good, prudent idea. It can be a flow chart on a piece of paper. And it may be far better if you use a more systematic method, such as a software program.
The problem is most of us (including me) don’t end up doing it, despite our best intentions. We want to do it. We agree we should do it. But we don’t do it — which explains why this is a New Year’s resolution article.
Some skilled mediators routinely use decision trees. They draw a flow chart on a board with input from counsel and the parties during the mediation. Typically, the decision tree will reveal hidden, yet powerful, information about the case, such as the statistical chances of being successful on a given issue (many of which end up invariably lower than what counsel and clients anticipated), and the cost of failing on any given issue (many of which are invariably higher than anticipated). I have found that tool to be quite effective to drive home to the parties and their counsel the hidden risks in a case.
For lawyers’ use in assessing risks and recommending a best course of action in cases or any other legal problems that involve other parties or events out of one’s control, specialized decision analysis software may help with ease of structure and formatting, input criteria, and with making sure all necessary elements are not only used, but used properly. I mentioned Choiceboxing earlier. Another vendor is TreeAge Software Inc. There are probably other vendors in the market.
To demonstrate the key basic elements of such software programs, I will use TreeAge’s basic legal analysis demo, which it calls Simple Legal Model. The model consists of a simple horizontal flow chart. Here is a screen shot of a sample:
|Click to enlarge|
The tree starts from the left with the party, goes on with a basic Shakespearean decision: to litigate or not to litigate (or rather, to litigate or to settle). Each branch develops from there. The litigation branch unfolds to proving the causes of action. Each of these develops into damages levels, and ends with a monetary value.
This simple chart must instinctively make sense to every lawyer. It seems to me a similar analysis can be done for commercial negotiations as well: for instance, whether to buy this business, merge with another, or do nothing. It would have branches to address the many business implications of those decisions, such as product ownership, market share, profit margins, growth plans, etc.
At its core and in more technical terms, a decision analysis tree is comprised of the following elements (using TreeAge’s model again for ease of reference):
(i) individual branches, which first deal with decisions (to litigate or settle), then with events that cannot be controlled (successfully providing a cause of action, providing damages, etc.), and ending with values;
(ii) numeric values, which are made up of (a) probabilities (what are the percentage changes of proving a cause of action or proving damages), and (b) values (what are the monetary values of each end result); and
(iii) internal model setting preferences, which control how a model is calculated and displayed.
Until I read Lauritsen’s above-referenced paper, I did not consciously realize a decision tree may also be used by multiple, even adversarial, parties together, towards a common goal. Indeed, mediators’ use of this tool, as noted earlier, is a case in point. As Lauritsen points out in his paper, the system (although he refers to his software) allows the following:
- “Provides a place for participants to record, update, manage, and share important information about a choice in progress;
- Prompts choosers for information about their goals and circumstances that are relevant to the choice;
- Suggests options, factors, ratings, resources, actions, and other things that choosers may find useful to weave into their choice in progress;
- Provides a single source of information that participants can be directed to, minimizing duplication of effort and inconsistencies in describing the choice;
- Lets participants see the whole developing picture, including input from others, minimizing reinvention and conflict.”
Whatever is your thing — paper and pencil or software program — this is a tool I think is worth your while.
Ben Hanuka may be reached at firstname.lastname@example.org. Ben has been practising commercial litigation with an emphasis on franchise disputes since 1998. He is the principal of Vaughan, Ont.-based Law Works P.C., a small boutique firm founded in 2012, whose files are in the cloud and feet firmly on the ground.
Column: The Future Files