Problem to be solved - The Intellectual Production of Engineers p.155 “Group formation requires a change in thinking” is difficult to convey.
-
From my observations in workshops, it seems that the most confusing step in the KJ method for many people is the group formation. This is probably because it requires a major change in the way we think about information processing. In this section, I will explain how you should change your way of thinking. - (5.2.4) Group formation requires a change in thinking
- I have written a lot about this “change of mindset” and “major change of thinking” but not much has been conveyed.
- We thought it would be easier to understand if we introduced specific episodes.
Group formation is a very important part for the KJ method to have a beneficial effect. Often, however, people do it incorrectly and to no avail. What is the purpose of doing this? It is to discover structures that you do not yet have from the information you have gathered. To do this, we need to organize groups subjectively from the bottom up, rather than objectively from the top down. The direction of processing needs to be reversed.
- (Prodding comment: it may be difficult to understand that we are talking about direction and subjective/objective at the same time here. Might be better to remove subjective/objective).
Specific examples
.
First, let me give you a concrete example. I once helped a marriage counselor put together an idea. The director wanted to gather his thoughts on communication between men and women. So we asked him to prepare 100 sticky notes first, and then we decided to have a two-hour meeting.
On the day of the meeting, his first words were, “I have come up with a classification. The classification was as follows.
- Male/Female
- Passive-Active
- Feelings and Actions
This is a typical failure when doing the KJ method. If you determine the classification criteria first and then classify accordingly, the classification criteria will only be reproduced.
So I advised him to do this.
- Put aside for a moment the classification criteria you have considered.
- (1) Spread 100 pieces
- (2) View 100 pieces
- (3) Find one pair of two cards that seem to be related.
- (4) Move it to the work space.
He looked at the spread of sticky notes and chose one pair.
Next,
- (5) I asked, “Why do you think these two seem related?” I asked.
- (6) I then listened to what he said and asked him to make a new sticky note with key words that came up in his remarks.
This is how we got them to experience the group formation process through in a minimal amount of time. (*1)
- First, instead of categorizing, we look at the sticky notes and find one pair that looks related. two pairs are a minimal group. This is the “label collecting” phase of creating groups.
- Then explain why those stickies are in one group. Then, make a new sticky with a brief description of it. This is the “making a table nameplate” phase.
Thus, three sticky notes, a group consisting of two sheets and their tableplates, were placed on the work space. Since the number of sheets was still small, we did not bundle them at this stage.
My next piece of advice was, “Are there any of the remaining sticky notes that might be related to these three? If so, pick them up and place them nearby. This is advice to make the task smaller and the goal closer. The goal of “Let’s form a group” is vague and far away, but “Are there any ~ in the remaining stickies?” is much easier to do because it can be completed by looking through the remaining stickies. (*2)
The rest is a repeat of this process. Going back to (2), look at the stickies that are not yet grouped and find one pair that seems to be related.
After about three times of doing this, he seemed to get the hang of the “find something that looks related and put it nearby” process, and he was able to continue working on his own without my direction or questions.
After a while, 16 groups were organized, averaging 6 cards. The average number of sheets was 6. Some groups had only 3 sheets, while others had more than 10 sheets. (*3)
I asked each group, “What kind of group is this?” and listened to the explanation. If the explanation was too long, I shortened it by asking, “In a nutshell?” If the explanation was one word, I asked in-depth questions to lengthen it, thus assisting the process of creating a nameplate of appropriate length. For example, “What is this group?” “It’s the ‘indifference’ group.” “And what kind of indifference is this indifference?” What kind of “indifference” is this “indifference”? (*4) This resulted, for example, in the following nameplate
- Indifference to the other party
- act without thinking
- Focus on appearance
Compare this to what he had in mind before grouping. He initially thought of several “sections” that would divide the entire sticky, such as “male/female” and “passive/active”. This is a structure we already have and we push it onto sticky notes top-down classification. On the other hand, his experience in this grouping was to attach one sticky note at a time, creating 16 groups with an average of 6 stickies. This is Bottom-up grouping.
He then performed another grouping for the 16 tableplates, which finally resulted in four groups. At this stage, each group is about the same size as two divisions multiplied by two in the top-down classification. So, would the nameplate for this group have been “passive x male” or something similar, which is a multiplication of the two divisions that we had initially thought of? That was not the case at all. For example, the human behavior pattern of “not wanting to leave one’s reach” was extracted from the sticky notes of 24 specific examples.
Thus, top-down classification and bottom-up grouping are going in opposite directions; it is the bottom-up grouping that we do in the KJ method. This is because the KJ method is for discovering unknown patterns, and classifying stickies according to known patterns hinders the discovery of new patterns.
*1: (6.3.1) Minimum feasible product *2: (1.2.2.2) Tutorial brings the goal closer *3: This large group of 10 or more pieces was put on the front cover later, and the process of putting the front cover on the smaller, more immediate group was done first. In the process, the feeling that “this group should be broken down” was expressed. *4: (6.2.4.2) Clean Language and Symbolic Modelling
This page is auto-translated from /nishio/「グループ編成には発想の転換が必要」加筆案 using DeepL. If you looks something interesting but the auto-translated English is not good enough to understand it, feel free to let me know at @nishio_en. I’m very happy to spread my thought to non-Japanese readers.