In recent work at the Creative Systems Laboratory at Carnegie Mellon University, we’ve been investigating new methods for online creative idea generation. One of the great benefits of leveraging the crowd to enhance human creativity is the diversity of ideation inputs from people all over the world.
Let’s reflect for a moment on how brainstorming activity is traditionally performed. Usually a group of people, perhaps designers working on a new product or service, or academics interested in generating new research ideas, gets together to ideate about a given project of interest. The brainstorm centers on an ideation prompt: a question that’s been developed as part of the project and that, if answered, would advance the project. The team might begin by framing a “How might we… ?” question. For example, they may ask “How might we build more efficient ways to crowdsource innovation?” And then the team will come up with as many creative solutions and responses to this question as possible, and vote on their favorites to determine which ones to move forward. The process generally looks something like this:
In one recent project, that’s exactly what we did. Based on some initial research into ideation techniques, and an interest in extending these techniques online, we brainstormed about online ideation and came up with dozens of new techniques. Two of these concepts are described below.
The first concept is what we call “chainstorming,” inspired by the idea of an online “broken telephone” (or “Chinese whispers”) game. When chainstorming each participant is asked to build on the work of a previous participant. The first person in the chain generates the prompt question and one or two ideas that respond to the question before sending it off to a network of friends. Each subsequent person sees the prompt question, along with a subset of the ideas from the previous participant, and uses these ideas to build on them and generate new ideas. So the chainstorming process would look something like this:
Chainstorming is conceptually interesting because it transforms brainstorming into a communication medium. Rather than requiring groups of people to build on the ideas of their collaborators in a shared local context, it enables ideation to be distributed as smaller tasks across space and time. Eventually, all of the ideas would circle back to the person or persons that originally posted the prompt, who would select their favorites.
The second key concept is what we call “cheatstorming,” which builds on chainstorming to determine which ideas are best. The basic premise of this paradigm is as follows: imagine ideation has been performed on a given prompt, resulting in 50 ideas. Participants vote on their favorite ideas, and some of them are selected for implementation. Now more ideation is performed on a different topic, resulting in 50 more ideas and additional voting. In time, many hundreds of prompt questions are asked, and thousands of ideas are generated and saved. Some have been implemented, and others have not. At this point, a wealth of valuable ideation has already occurred. Cheatstorming proposes that no new ideas are necessary for further ideation to occur. Given a new prompt question and a set of 50 random previous ideas to draw from, cheatstorming simply bypasses the concept generation phase altogether and jumps directly to voting on which ideas to advance:
We call this cheatstorming because no new ideas are actually generated. Instead, we’ve simply gathered ideas from elsewhere and chosen our favorites.
To test this concept we performed a simple pilot experiment. First, each member of our team generated 3-5 “totally random” brainstorm questions on Post-It notes, not in response to any particular question or stated need (e.g. “What is the easiest way to make the most people happy cheaply?”). Next, a set of 60+ solution concepts was generated equally at random, like this:
Finally, one of the previously generated brainstorm questions was selected at random and paired with 10 of the concept Post-Its at random. From these 10 ideas, the four concepts that most closely resonated as solutions to the given question were selected as “winners.” We repeated this process four times with four different questions.
We were both surprised and delighted by the results of this method. Not only did we have little difficulty identifying those ideas that best resonated with the questions being asked, the resulting set of ideas was remarkably unexpected and fresh. Most exciting, the process was fast, fun, and required low effort, and the solutions revealed unexpected combinatory patterns and juxtapositions. Consider the result from the following cheatstorm:
The question asks “How could we illuminate large cities for less money to reduce nocturnal crime?” Surprisingly, three of the selected solution concepts are screen-based ideas that all emit light. Not only was this an unanticipated means of illumination, it was also one that could provide other forms of safety from nocturnal crime—via an interactive “call for help” kiosk or informative map, for example. Furthermore, the fourth idea in this set, “airbag for walking,” suggests that perhaps solutions for reducing nocturnal crime could be built directly into a user’s clothing. Combined with the other cheatstormed ideas, this in turn sparks a train of thought that perhaps clothing should be illuminated, or—alternatively—that the city’s streets should be padded.
Cheatstorming demonstrates that ideas need not be created by the team for ideation to occur—they simply need to be interpreted as possibilities resulting from a collision of shared meanings. The only requirement for a successful ideation outcome is that the ideas introduced in the sharing stage are unconventional to the ideating individual, team, or culture (i.e. “strange”), and that they be interpreted as relevant (or not) to the ideation prompt. Furthermore, this method intuitively incorporates conventional brainstorming rules: it results in lots of wild ideas because they’re arriving from completely different contexts; there’s no shortage of quantity, presuming lots of previous ideation has previously occurred elsewhere; it implicitly builds on the ideas of others; and users have no choice but to suspend their judgment and work with the ideas they’re dealt.
In our subsequent studies, we found that it is the unique ability of cheatstorming to “dial in strangeness,” that makes it such a compelling paradigm for ideation online. When determining which ideas should be part of the cheatstorming input, ideas can be drawn on from related or totally unrelated fields. Thus cheatstorming provides far more nuance than existing methods of random input such as inspiration card workshops and, because it enables easy changes to the ideation methodology and content directly, it can therefore facilitate targeted and highly contextual “leaps” of imagination from an original set of ideas to a much wider framing of the problem domain.
For more, see our full paper: Brainstorm, Chainstorm, Cheatstorm, Tweetstorm: New Ideation Strategies for Distributed HCI Design. Haakon Faste, Nir Rachmel, Russell Essary, Evan Sheehan, Proc. CHI, 2013.
And our video preview, here.
Suggested Further Reading:
- de Bono, E. (1970). Lateral Thinking: Creativity Step By Step, Harper Perennial.
- Faste, R. (1995). A Visual Essay on Invention and Innovation,” Design Management Journal, 6(2).
- Isaksen, S. G. (1998). A Review of Brainstorming Research: Six Critical Issues for Inquiry, Technical report, Creative Problem Solving Group, Buffalo, NY.
- Osborn, A. F. (1963). Applied Imagination: Principles and procedures of creative thinking (3rd edition), Scribner.
- Yu, L., & Nickerson, J. (2011). Cooks or Cobblers? Crowd Creativity through Combination, Proc. CHI, 1393-1402.