Idea Seeding Better Than Brainstorming


Kevin Cheng and Tom Chi, at OK/Cancel have written an article sharing the creative process they use for creating their awesome strips. Idea seeding is the process where they use time constraints and design/refine cycles to improve their ability to create quality “product.” They also wonder about extending this approach to other areas where brainstorming is normally used.


Even if you aren’t interested in idea seeding, you need to check out OK/Cancel. Tom and Kevin are incredibly talented designers who run a popular website creating a great community focused on (software) interfaces and the people who create them. Their articles, cartoons, and the discussion threads are top shelf.

A recent favorite strip of mine shows whatmight be titled Project Managers Behaving Badly.

Brainstorming Out

Kevin details their old design process for creating new strips, which involved a 2 to 4 hour brainstorming session. While they met their quality goals for the strip (their product), they felt that the process was too time-intensive. Tom had used a process called “switch sessions” as part of collaborating on music, and they adapted that process for comic-strip development.

Idea Seeding In

Kevin thinks “Idea Seeding” is a better name for the process, which they detail as follows:

  • Constrained Idea Generation. Each person works independently for 30 minutes to document the ideas for four comic strips.
  • Edit And Refine. The collaborators swap strip ideas, and for 30 minutes, refine or complete the other person’s ideas.
  • Collaborate And Polish. In a final 30 minute session, they collaborate to discard, improve, and select the comics they’ve created.

Kevin and Tom express that this is working better for them than brainstorming did. As a “user” of their product, I would have to agree, at least on the quality of the strips – they are just as good or better than before. And the process is saving them a ton of time.

One of the many commenters expressed that this is just a variation of “individual brainstorming” which has been proven to be more effective. The commenter cites this study as evidence of the relative efficiencies. The study tested brainstorming results using 36 college students. In our earlier article about the Wall Street Journal’s coverage of similar brainstorming studies, we quoted Dr. Sutton as a detractor of these studies:

To put it another way, if these were studies of sexual performance, it would be like drawing inferences about what happens with experienced couples on the basis of research done only with virgins during the first time they had sex.

Bob Sutton

Why It Works

The compelling element isn’t the lack of collaboration in the first session, it is the artificial constraint placed on time. Kathy Sierra has written about exactly this effect here and here and here and probably more places. She shares fantastic insights about cognition in many of her posts.

The second session is really interesting in that it has a marked divergence from brainstorming. In brainstorming, one idea builds on another. In this refinement session, one idea improves another. Brainstorming doesn’t allow for negative feedback or other valuations of ideas during the gathering stage – ideas are prioritized at the end of the session. And those ideas are all “first generation” ideas. With an edit cycle, the original ideas are improved – which implies that their weaknesses are identified.

Better Than Brainstorming?

Certainly Kevin and Tom are happier with this approach than using brainstorming. And Kevin wonders if it can be applied to interface design (creating wireframes, etc).

Idea Seeding For Process Re-Engineering

One idea immediately jumped out for how to use this as a business analyst – process re-engineering. Process re-engineering requires both insight and out of the box thinking. It is both a creative and an analytical exercise. We could use the “three session” approach to achieve these goals.

We would also have the benefit of pre-seeding the process by leveraging documentation of existing processes. Using two business analysts, the 30 minute sessions might look like this:

  1. Each business analyst identifies four (or three?) process changes that might be valuable if they worked, and fleshes out the ideas as much as possible.
  2. The analysts swap changes-proposals. Each business analyst then works to refine the process changes from the first exercise.
  3. The two business analysts meet to review the results and determine which changes to combine, pursue, or discard.

Kevin mentioned that their process worked well because they had a low-overhead template that made it easy to record thoughts and ideas in various stages of completeness. We would need something similar for our process change exercise.

[update: Kevin’s idea seeding article also covered at the business innovation insider]

Idea Seeding for Requirements Gathering

Requirements gathering is more about discovery than refinement. The collaboration inherent in a brainstorming session should be more effective at discovery than idea seeding.


Idea Seeding works better for Tom and Kevin than brainstorming did – same or higher quality in less time. We can apply the same technique to process re-engineering, but probably not to requirements elicitation.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.