Description of the Structured Dialogic Design process for collaborative problem solving

sdd.md

Structured Dialogic Design

The following is a rough overview of the collaborative problem solving technique called Structured Dialogic Design.

Steps

1. Determine stakeholder groups, and knowledgable and trusted people in those groups. Find diverse backgrounds, world experiences, skill sets, specializations, etc. Max ~200, min ~20.
2. Construct a good "Triggering Question" that will be the thing to answer during the design sessions.
3. Gather folx together in solid time blocks to do the subsequent steps. Usually this is in person, but it can be remote, too. Two full days tends to be time involved, but it can go faster, spread out, etc. as needed.
4. Come to a common understanding of the problem and its causes:
   1. Each person writes down a small number (~2-3) of main "key concepts" or "relevant factors" -- a short name or phrase for the concept, and a longer 1-paragraph-max description summary.
   2. In turn, each participant reads out the name of one of their most important concept that they wrote, until at least two concepts from each person have been written out onto a shared whiteboard.
   3. When all names are read, we go through in the same order and read the descriptions, putting them onto the whiteboard space too, with the concept name
   4. We then give each person in turn the option to ask a clarifying question about any of the concepts, and the originator of the concept can try to respond. Each question is hopefully clarified, but the conversation should be limited to ~30 seconds. All the convos should be recorded with the concepts
   5. The concept names are then clustered into related groups by rough "belong together" show of hands voting:
      1. Start with one group, with the first concept.
      2. For each remaining concept, for each group, ask if it "belongs with this group, or not". Put it with the highest "belongs here" group, or else put it in a new group of its own.
      3. Repeat this till there's nothing left to categorize.
      4. Ask if anyone has any last minute adjustments they think should be made, pull out the relevant concepts, and repeat the process, if there's any strong disagreements, and stop when the categories are "not too bad".
   6. Give names to each group of concepts. There should probably be only a dozen or two at most.
   7. Do pairwise comparisons asking a question like "In the context of the triggering question, would solving A help solve B? How about vice versa?" or "Given the triggering question, is A a bigger contributor to the problem than B is? How about the other way around?". Tally up the votes in each direction.
   8. Analyze the resulting graph of influences, to determine which concepts and problems are most important, wrt the question asked. This gives you a shared consensus understanding of what needs to be addressed, what's most relevant, etc. All of this should be recorded and shared.
5. Develop some action plan options:
   1. Repeat steps 1-6 from the concepts phase, but this time ask about "ideas for action". Record and share.
   2. Do a pair-wise analysis with ideas for action, paired against each concept, asking questions like "In the context of the triggering question, would this idea for action help us address this concept/problem?", similarly recording how many people say yes. Record and share.
   3. Split into groups of 3-5, and use the information about action ideas to develop a possible action plan, with steps to take, brief little descriptions of which problems the steps address, and why the sequence is that way.
   4. Re-assemble everyone, and share ideas together as a group. Discuss pros and cons. Optionally re-form groups and revise action plans, then repeat presentations.
   5. For each final action plan, ask the group "Does this have a good chance at addressing the triggering question?", and tally up all the yes'es. These are not exclusionary votes, you can say "yes" to as many as you want.