Creative Thinking: How to Generate, Improve and Test Better Ideas

Creative thinking is often described as though good ideas arrive in flashes. Someone suddenly sees the answer, invents the concept or discovers a connection nobody else noticed. Moments like that happen, but they are a poor model for doing creative work consistently. Most useful ideas emerge through a more ordinary process: understanding the problem, producing alternatives, changing perspective, testing assumptions, improving promising options and exposing them to reality.

This is also close to how the OECD defines creative thinking in its PISA framework. Rather than treating creativity as originality alone, PISA defines creative thinking as the capacity to engage productively in the generation, evaluation and improvement of ideas that can produce original and effective solutions, advances in knowledge or meaningful expressions of imagination. For assessment, it separates this into generating diverse ideas, generating creative ideas, and evaluating and improving ideas.

That distinction matters because novelty by itself has little practical value. An idea can be unusual but impossible, entertaining but irrelevant, or original while creating a larger problem than the one it solves. Useful creativity therefore moves between two broad modes. The first expands the solution space so that the obvious answer is not accepted too early. The second narrows and develops that space so that promising ideas become workable.

Creative thinking is not the absence of discipline. It is disciplined variation.

Start by changing the problem before changing the answer

Many weak solutions begin with a poorly framed question.

If a team asks, “How can we make this form shorter?”, it has already decided that the form should continue to exist. A broader question—“How can we collect the information we genuinely need with less effort from the user?”—allows possibilities such as pre-filled information, automation, elimination of unnecessary fields or collecting information at a different stage.

Likewise, “How do I make this presentation more attractive?” pushes attention toward colours, fonts and animations. “How do I make the audience understand and remember the three decisions that matter?” creates a very different design problem.

Reframing does not mean making the problem vague. A useful creative brief still needs an objective, audience or user, important constraints and a clear description of what is currently unsatisfactory. The aim is to identify assumptions that have been embedded inside the question without being examined.

One practical test is to separate the goal from the existing method. A company does not ultimately need a weekly spreadsheet; it needs information for a decision. A customer does not need a five-stage approval process; the organisation needs sufficient control over a risk. A student does not need a prettier set of notes; they need to understand and retrieve information effectively.

Once the real objective is visible, the existing method becomes one option rather than the problem definition itself.

Constraints can help expose these assumptions. Ask what would happen with half the budget, half the time or no ability to hire additional staff. Imagine the process had to work without email, without a meeting or without specialised software. Ask what could be done if every change had to be reversible within a day.

Then reverse the exercise. What becomes possible if one important constraint disappears?

These questions are not predictions about what will actually happen. They are deliberately altered environments designed to reveal which assumptions are essential and which have merely become familiar.

Generate alternatives before deciding which one deserves to survive

Once a problem has been framed well enough, creative thinking needs room to diverge.

Evaluating every idea the moment it appears can make familiar options disproportionately attractive. Familiar ideas are easier to justify because the organisation already understands them. More unusual possibilities often begin incomplete and therefore look weaker when compared prematurely with polished conventional solutions.

A better sequence is to create a short period in which the objective is variety, followed by a distinct period of evaluation.

This does not require pretending every idea is good. Evaluation remains essential. The purpose of delaying it is simply to prevent selection from occurring before exploration has produced enough alternatives.

The OECD's PISA framework reflects this distinction by measuring both the ability to generate diverse ideas and the ability to generate creative ideas, before separately assessing evaluation and improvement. Generating several answers is therefore not enough if every answer belongs to the same narrow category.

Suppose a company wants to reduce customer waiting time. Five ideas involving slightly different queue displays may technically be five suggestions but represent very little conceptual diversity. A more useful exercise deliberately changes the mechanism: one idea removes a step, another changes who performs it, another changes sequence, another uses automation, another deliberately avoids new technology, and another changes the incentive creating the delay.

These categories should not become rigid brainstorming formulas. Their purpose is to force distance between alternatives.

Another effective method is to borrow structures from other domains. A hospital triage system may suggest a way to prioritise incoming support cases. Software version control can inspire a better method of tracking document revisions. Restaurant reservations may suggest capacity booking for shared equipment or internal specialists.

The important move is not copying the visible surface of another industry. It is identifying the underlying mechanism. What problem does the other system solve? How does it allocate scarce capacity, reduce uncertainty, identify priorities or coordinate people? Does the same logic transfer to the current problem?

Analogy works similarly. Complete the sentence, “This problem is like...” and then examine the relationship rather than the object. A project pipeline might resemble traffic flow because both involve capacity, bottlenecks and queues. A training programme might resemble progressive physical exercise because both involve increasing difficulty, feedback and recovery.

Analogies are powerful precisely because they import relationships from one mental model into another. They are also dangerous when similarities are emphasised and differences ignored. Use analogy to generate possibilities, then return to the actual evidence and constraints.

Externalise the problem and step away when fixation appears

Ideas are easier to compare when they are outside the mind.

Working memory is limited, particularly when a problem contains several conditions, stakeholders and alternatives. Sketching a process, writing ideas on cards, creating a rough flow diagram or building an intentionally crude prototype reduces the amount of information that has to be mentally maintained at once.

Externalisation also exposes ambiguity. A proposed workflow may sound coherent until somebody tries to draw who sends what to whom. A product concept may appear obvious until someone sketches the interface and discovers three unanswered decisions. A strategy may look comprehensive until its assumptions are placed side by side.

The representation does not need to be attractive. Its first job is thinking.

When exploration repeatedly returns to the same solutions, temporarily leaving the problem can also help. Creativity researchers describe this as incubation: setting a problem aside before returning to it. A meta-analysis of incubation research found an overall positive effect on problem solving, with the benefit varying according to the kind of problem, preparation and what people did during the break. Divergent-thinking tasks showed particularly useful incubation effects, while cognitively demanding interruption tended to reduce the benefit.

That finding does not justify waiting passively for inspiration.

Incubation appears most useful after genuine engagement with the problem. If the problem has never been understood or explored, taking a walk is unlikely to substitute for analysis. The useful sequence is usually to work seriously enough to build a mental representation, recognise that thinking has become repetitive, shift temporarily to something else, and then return.

A practical variation is to return without immediately rereading previous ideas. Generate again first. This creates a chance for new associations to appear before the earlier solution set pulls attention back toward the same paths.

Modern neuroscience also gives little support to the popular idea that creativity is simply a “right-brain” ability. Research instead points toward interactions among large-scale brain systems involved in internally generated thought, cognitive control, memory retrieval and inhibition. Reviews of creative cognition particularly highlight cooperation between the default network and executive-control systems during creative performance.

That fits the broader practical lesson: creativity involves both freer associative processes and controlled evaluation rather than one mental mode operating alone.

Improve ideas instead of merely choosing winners

Once alternatives exist, the process should change.

Evaluation is often treated as elimination: score the ideas, select the highest-ranked one and discard the rest. That can be efficient, but it misses a major part of creative thinking.

Promising ideas are frequently incomplete.

Instead of asking only “Which one wins?”, ask what is valuable inside each option. Which assumption makes it weak? Could the strongest part be combined with another idea? What change would make it cheaper, safer, easier to use or more reversible? Could a technically difficult idea be simplified without losing the mechanism that made it interesting?

This is why the OECD framework explicitly includes evaluating and improving ideas, not merely generating them. Creative quality can emerge through revision rather than appearing fully formed in the first proposal.

Explicit criteria become useful at this stage.

The correct criteria depend on the problem, but they might include user value, originality, feasibility, cost, implementation time, risk, reversibility, strategic fit and expected impact. A safety-critical system will weight reliability differently from an advertising concept. A temporary pilot can tolerate uncertainty that a permanent infrastructure decision cannot.

Criteria should therefore come from the objective rather than from a generic creativity scorecard.

Originality is one criterion, not the entire definition of quality.

A spectacularly novel idea that violates a legal requirement or creates unacceptable safety risk is not a strong solution. Conversely, a highly feasible idea that changes almost nothing may not justify implementation.

The task is to find or develop an idea that creates enough value under the actual constraints.

Different perspectives can improve this evaluation when they contribute different information. The employee closest to the customer may see usability problems that senior managers miss. An operations specialist may notice dependencies. A risk professional may identify failure modes. A newcomer can question conventions that experienced colleagues stopped noticing.

The useful principle is not “invite more people”. Large groups can create coordination costs of their own. Invite people whose knowledge, experience or vantage point changes the information available to the problem.

Prototype before discussion makes the idea expensive

Ideas become psychologically expensive when people spend large amounts of time defending them before reality has tested them.

Prototyping reduces that problem.

A prototype is simply the smallest representation capable of answering an important question. A sketch can reveal whether users understand a concept. A mock landing page can test whether a message makes sense. A spreadsheet can test the logic of a proposed workflow. A temporary manual process can test demand before software is built. A one-week pilot can show whether changing a meeting structure actually improves decisions.

The prototype does not have to resemble the final product closely. It needs to produce useful information.

This changes the purpose of creative work from proving that an idea is brilliant to discovering what is true about it.

Suppose a team proposes an automated client-reporting system. Months of discussion could focus on whether customers will find it useful. A rough prototype containing sample reports shown to ten representative users may answer the central question far more quickly.

Testing can also reveal that the idea is solving the wrong problem.

Users might value faster access rather than more sophisticated analysis. Staff might resist a proposed workflow not because they dislike change but because the new design removes information they need later. A seemingly technical obstacle may turn out to be an incentive problem.

The earlier this information appears, the cheaper revision becomes.

Prototypes therefore connect creative thinking with learning. They transform assumptions into questions that can be tested.

The same principle helps prevent excessive attachment to a single solution. If a prototype is deliberately small, failure becomes information rather than the collapse of a six-month investment.

Creativity is neither brainstorming nor unrestricted originality

Several popular beliefs make creative thinking harder to understand.

Creativity is not simply a right-brain activity. Contemporary neuroscience describes creative cognition through interactions among multiple large-scale systems rather than locating creativity in one hemisphere or mental faculty.

Creativity is also not the opposite of expertise. Knowledge supplies material that can be recombined, compared and transferred. A skilled engineer can notice technical analogies that a novice cannot because the engineer possesses richer mental models. Expertise can also create fixation when familiar solutions are accepted too quickly, which is why deliberate reframing remains useful.

Brainstorming is not synonymous with creative thinking either. It is one family of techniques for producing ideas. Creative work also includes problem framing, analogy, research, sketching, experimentation, criticism, improvement and testing.

Nor is creativity measured simply by how strange an idea appears.

The OECD's formulation deliberately links originality with effectiveness and useful improvement. An idea can violate convention and still solve nothing.

The more useful distinction is between variation and value.

Variation creates alternatives. Evaluation asks which alternatives matter. Development improves them. Testing determines whether the assumptions survive contact with reality.

All four are necessary for practical creative work.

A repeatable creative-thinking cycle

A useful creative process can therefore be understood as a cycle rather than a mysterious event.

Frame: define the actual objective, audience and constraints while identifying assumptions hidden inside the current problem definition.

Diverge: produce multiple genuinely different possibilities rather than variations of the first answer.

Disrupt: change constraints, use analogies, borrow mechanisms from other fields and seek perspectives that introduce different information.

Pause: when serious effort has produced fixation, temporarily leave the problem and return after an incubation period.

Externalise: sketch, map, write, model or prototype so that ideas can be compared and assumptions become visible.

Converge: evaluate alternatives against explicit criteria connected to the real objective.

Develop: improve promising ideas, combine useful elements and repair weaknesses rather than simply selecting an initial winner.

Test: create the smallest realistic experiment capable of exposing an important assumption to evidence.

The cycle can repeat. Testing may reveal that the problem itself was framed incorrectly, returning the process to the beginning.

That is not creative failure. It is how creative work becomes more accurate.

Creative thinking is disciplined variation

The opposite of creative thinking is not seriousness, expertise or evaluation. It is accepting the first plausible model too quickly.

Creative thinkers create room for alternatives before committing. They ask whether the question itself is too narrow. They deliberately change constraints, transfer useful structures from other fields, externalise ideas and step away when fixation appears.

Then they become demanding.

They compare ideas against evidence and constraints. They improve rather than merely admire originality. They prototype early enough that reality can contradict them cheaply.

This combination of expansion and selection is why creativity can be developed as a practical capability rather than treated as a personality trait available only to unusually imaginative people. The OECD's PISA work explicitly frames everyday creative thinking as a malleable capacity that can be developed through practice.

Useful creativity rarely depends on producing one perfect idea in a flash.

It depends on generating enough variation to escape the obvious answer, applying enough judgment to improve what emerges and testing early enough to learn before an idea becomes too expensive to change.

Originality opens the possibilities.

Evaluation and revision turn those possibilities into something worth using.