Two modes, two different jobs
Imagine a team trying to make a complicated sign-up process easier. One person suggests shorter instructions. Another proposes fewer form fields. Someone else asks whether people need an account at all. These ideas point in different directions, and that is useful at first. Divergent thinking expands the range of possibilities before the team commits to a solution. Convergent thinking compares those possibilities, finds what matters most, and selects a direction to pursue.
The modes work best when their jobs are kept distinct. If every suggestion is judged the moment it appears, familiar options can crowd out unusual but promising ones. If a team keeps generating ideas without evaluating them, it may leave with an exciting list and no decision. The goal is not to be endlessly creative or relentlessly critical. It is to give exploration and selection their own time, while allowing new ideas to emerge again when the evidence calls for them.
Start with a question worth exploring
Before generating ideas, define the problem without building a preferred answer into it. “How do we make the instructions shorter?” assumes that length is the obstacle. “Where do people get stuck during sign-up, and how might we reduce that friction?” leaves room to consider wording, form design, timing, and even whether each step is necessary. A useful question identifies who is affected, what difficulty they face, and which constraints cannot be ignored.
Next, create a short period in which the aim is variety rather than approval. People can write ideas individually before sharing them, so the first confident voice does not set the agenda. Invite alternatives at different scales: a small edit, a change to the process, and a more fundamental rethink. Combining two suggestions or asking “What would we try if this constraint changed?” can reveal options that a straightforward brainstorm misses. Record ideas in plain language so they can be understood later without their author present.
Narrow with criteria, not first impressions
Convergence begins by making the decision criteria visible. For the sign-up example, the team might consider whether an idea addresses a documented point of confusion, how difficult it would be to implement, whether it introduces a new barrier, and how easily its effect could be checked. Criteria should match the problem. A quick experiment may favor low effort and easy measurement; a lasting redesign may warrant more attention to accessibility, maintenance, and possible unintended effects.
Group similar ideas, then compare them against those criteria. Avoid treating a simple score as proof that one option is best: ratings often hide assumptions and uncertainty. Instead, ask what would have to be true for an idea to work and what might cause it to fail. A proposal to remove account creation, for instance, could simplify an early step but complicate returning later. Naming that trade-off makes the choice more informed than voting for the most appealing-sounding suggestion.
Turn a choice into a learning step
Choosing an idea does not require pretending it is certain to succeed. A team can select the option that is promising and practical to test, then define what it hopes to learn. It might try revised instructions with a small set of users and ask where the process remains unclear. If the feedback points to the form itself rather than the wording, the team has learned something useful without committing to a complete redesign.
This is where divergent and convergent thinking become a cycle rather than a one-way funnel. A test may confirm a direction, reveal a flaw, or open up a better question. Return to idea generation when the selected option fails to address the problem or when a newly discovered constraint changes what is possible. Return to evaluation when there are enough distinct options to make a meaningful comparison. The point is to alternate deliberately, not to follow a fixed number of rounds.
Make both modes easier for a group
Groups can improve the process by saying which mode they are in and how long it will last. During generation, save objections for later and make space for quieter contributors. During evaluation, invite specific concerns, including from people who did not propose the leading idea. A facilitator can keep a visible list of suggestions, criteria, open questions, and the reasons behind the final choice. That record helps distinguish a considered decision from a preference that happened to win the room.
The same approach works alone. Write several genuinely different responses to a problem before ranking any of them. Then choose a few criteria, identify the assumptions behind your preferred option, and decide on a manageable next step. Divergent thinking gives you alternatives; convergent thinking gives you a reason to act. Used together, they help prevent both premature certainty and endless possibility.
