Why the first idea is usually the wrong one

Creative teams are routinely asked for twenty concepts instead of three. That instruction is not arbitrary. It comes from a well-documented account of how thinking goes wrong. 

Three thinking errors

Chris Griffiths organises the obstacles into three kinds of thinking. All three are useful at the right moment, and all three are damaging when you need to think expansively. 

Selective thinking is the tendency to validate certain ideas and discount others — typically favouring our own "pet" ideas. Griffiths treats this as the single most harmful pattern in brainstorming and decision-making, and links it to confirmation bias: we notice evidence that supports what we already think and filter out the rest. He quotes the French philosopher Emile Chartier: nothing is more dangerous than an idea when it is the only one we have. 

Reactive thinking is reacting to existing influences, events or ideas too quickly — Kahneman's fast, automatic System 1 running when the slower, deliberate System 2 is needed. Fast reaction is essential for physical survival and useless for a strategic creative decision. 

Assumptive thinking is accepting a belief or convention as true with no proof, usually on the basis of past experience or "common knowledge". Adair adds a sharp distinction here: assumptions made consciously and temporarily are a legitimate tool — like making a chess move while keeping your hand on the piece. Assumptions made unconsciously are, in his phrase, hidden sandbanks outside the harbour mouth. 

A team that proposes "ten NFT pictures about wheelchairs" four minutes into its first session has run all three at once: it reacted fast, selected the first available idea, and assumed that the way to depict a disability cause is to depict the equipment. 

Defining the problem is a creative act in itself

Adair is emphatic that analysis, not inspiration, is where most creative work is won or lost. His argument, compressed: 

• The object of analysis is clarity of thought — and the commonest failure is treating symptoms rather than the disease. 
• There are often several equally valid but not equally obvious ways of defining any problem. Different definitions are worth collecting; they are signposts to different avenues of thought. 
• Every problem definition is, in effect, a general statement of a potential solution. Choose the definition carelessly and you have already chosen the answer. 
• Therefore: do not be over-hasty. Play with alternative formulations until one emerges that commands your support — and if a definition leads nowhere, try another. 

This is exactly why the lesson asks you to write your problem three different ways before generating anything. Magistretti, Pham and Dell'Era's recent review of the field treats problem framing as a creative process in its own right rather than a preliminary step — a position that matches Adair's almost exactly, seventy years later. 

Griffiths operationalises the same insight in Step 1 of his four-step Solution Finder process (Understanding → Ideation → Analysis → Direction). His Understanding stage asks a team to state the challenge, record initial opinions and obstacles, mark the boundaries, and then explicitly restate and reframe it. He recommends what he calls invitational stem language — openings like "It would be great if…" or "How might we…?" — precisely because they prevent you closing in on the problem prematurely. 

His Understanding canvas also asks a team to separate three columns that are usually blurred: what we know (facts only), what we think we know (assumptions), and what we need to know (research). When you are working on a social cause, that third column is often the most revealing. 

Widening the span of relevance

Adair's principle of the span of relevance explains why teams get stuck inside their own field. People with a narrow span think within the tramlines of their own industry; the creative contribution often comes from transferring a solution across a boundary. His examples include Jethro Tull deriving the seed drill from the mechanism of a church organ. 

He goes further, and it is worth sitting with: too much knowledge may be a disadvantage. Expertise in a field builds strong patterns, and strong patterns are hard to see past — so, he argues, we must learn to unlearn. In a mixed team, this means the youth worker who has run the same campaign for fifteen years and the seventeen-year-old who has never seen it are each holding half of what is needed. 

Goller and Bessant reach the same conclusion from the research side, noting Dorothy Leonard's "deep smarts": deep domain knowledge lets you see patterns others miss, and simultaneously narrows the range of variations you consider. Expertise is an asset and a liability at once. 

Divergence before convergence — and why quantity comes first

Griffiths' Solution Finder alternates Guilford's two modes deliberately: Understanding and Ideation are divergent, Analysis and Direction are convergent. The mistake almost every team makes is running them simultaneously, so that ideas are judged before they are grown. 

Adair's term for the fix is suspending judgement — and he defines it precisely within his own model of thinking. He describes effective thinking as three functions: analysing (taking apart), synthesizing (building a whole from elements) and valuing (assessing worth). Suspending judgement means erecting a temporary barrier between the first two and the third. His warning is two-sided: we post guards on our own minds through premature self-criticism, and we should also beware of over-critical others. 

His practical instruction is one line worth putting on a wall: do not confuse evaluation with idea fluency — be as prolific as you can with ideas until you find one that satisfies you.

Where the twenty comes from

Putting these together gives the reasoning behind the practical exercise. Selective thinking means the first idea will be defended disproportionately. Reactive thinking means it will arrive too fast. Assumptive thinking means it will encode the stereotypes already in the room. A narrow span of relevance means it will look like everything else in the sector. And judgement applied too early will kill the unusual candidates before they have grown. 

Twenty concepts, generated silently and shared without comment, is a structural answer to all five at once.