Close

January 31, 2011

Learn to build issue trees by watching

Matthew Juniper, at Cambridge, demonstrates how to build issue trees. His approach is very similar to ours and he has an excellent video to describe it so you might want to watch it; here are the highlights I like best.

Work on one specific question. Matthew starts his issue tree based on one question, not several. I think he is absolutely right: first you have to identify the specific question you need to answer.

It is not about being mutually exclusive. We talked about the importance of thinking in a mutually exclusive and collectively exhaustive way in most of the posts on this site. What we’ve called MECE so far, Matthew calls independent and complete. And he is right: when you break down a problem into its potential root causes or solutions, it isn’t about mutual exclusiveness but, rather, independency.

In the strict sense, branches in issue trees only have to be independent, they don’t have to be mutually exclusive.

In the strict sense, two events are mutually exclusive if the occurrence of one precludes the occurrence of the other. So, to be mutually exclusive, the two events must be independent, but two independent events aren’t necessarily ME.

When you are diagnosing a problem or looking for solutions, you want the various branches of your trees to be independent, but you don’t necessarily need them to be ME: the reason why we are late in our cable negotiations might because we prevent a timely close or because our suppliers prevent it. But one doesn’t preclude the other: in fact both we and our suppliers might be partly responsible. So these branches are not ME but, rather, independent and organizing our tree with this construction is completely acceptable. On the other hand, when we want to go from NYC to London and we separate the means of transportation between flying or traveling by sea, we identify options that are both independent and mutually exclusive: traveling by sea precludes me from traveling by air and vice-versa.

So the point is that branches in your issue trees should always be independent. Sometimes, depending on the nature of your problem they will also be mutually exclusive.

Start with thinking creatively, then organize your ideas.

To be an effective problem solver you need to be good at two types of thinking: creative and critical. Creative thinking allows you to think about various ways to answer a specific question; it’s the collectively exhaustive/complete part of the equation. Critical thinking is about organizing the material in an efficient way so that you don’t duplicate efforts; it’s the mutually exclusive/independent part.

Think creatively (coming up with potential answers) before thinking critically (organizing them).

 

Before you organize stuff, you need to come up with that stuff. Matthew illustrates this very well, showing how you first brainstorm to identify all the possible solutions to his question. Then he moves on to organize them.

Right around 5:05 in the video, he realizes that selling clothes, the idea he had originally identified, isn’t quite complete and he substitutes “clothes” for “belongings”. Just a few seconds later he realizes that, that too, isn’t quite complete and breaks down the branch into “belongings”, “services”, and “body parts” which, he notes, is indeed a one-off way to make money (you have to love British humor).

So the organization process is iterative but it starts with the creative thinking part.

First be exhaustive, then decide what you want to do.

As Matthew explains around 9:20, first come up with all the possible things that you can do and then decide which ones you are not going to pursue. That’s another simple-but-powerful approach: first write down every possible answer and then filter them.

It’s like riding your bike. To do a good job, you need to pedal and to break but not both at the same time, and you need to pedal to build speed before you break. By forcing you to think about all the crazy possibilities out there (pedal), you force yourself outside of your comfort zone, thereby avoiding to satisfice/settle for the obvious solutions.

If you go straight to eliminating ideas (brake), you will blackout some of your thinking. Also it won’t help having others collaborating on your trees (they’ll ask you why you didn’t include these ideas). So think about everything, include all of it in your issue tree and when you decide not to pursue some possibilities, transport these from your tree to the out-of-scope section of your problem identification card.

Insightfulness comes from thinking about your problem. Your first cut at organizing your tree might not be the best one, but it will help trigger your thoughts to identify others (~ 10:30). So try to think about different ways to organize your tree, compare them and select the one that works best for your specific situation.

Enlist others. Matthew also talks about the importance of having several people working on the same problem as a way to improve the solution process (~ 13:00). We’ve also talked about this: I think he is absolutely right; effective collaboration is probably one of our best tool.

Use similar elements. He calls that similarity (~ 14:30), we call it parallelism but we’re talking about the same concept: for your tree to be effective, you need to break it into elements that are alike.

Be prepared to spend some time on the process. Along his presentation, Matthew makes a point of explaining how the process can take a few hours or even a day. I think he is right and, in fact, for larger problems you should be prepared to invest more time still in developing your issue tree. Since your tree will guide your analysis, it really is the roadmap that you’ll follow for the rest of your research so for big, complex problems, be prepared to spend several days—or even a week—developing your issue tree.

One Comment on “Learn to build issue trees by watching

Matthew Juniper
January 5, 2012 at 05:49

Hello,

I have moved the issue tree movie to this site:
http://dl.dropbox.com/u/10418369/MPJ_website/issue_hyp_trees/flash/issue_tree/issue_tree.html

You can find the other issue and hypothesis trees from the links on my website:
http://www2.eng.cam.ac.uk/~mpj1001/MJ_teaching_pg.html

Best wishes,

Matthew Juniper

Comments are closed.