How to Evaluate Ideas: A Practical Shortlist Method

The short answer

To evaluate ideas, define the decision first, remove options that fail hard requirements, and compare the survivors against three to five criteria that matter for this project. Score only to expose trade-offs—not to manufacture certainty—then test the leading ideas with the cheapest useful experiment. Choose one, record why, and give it a fair trial before reopening the contest.

The crucial move is separating idea generation from evaluation. During generation, unusual answers need room to breathe. During evaluation, enthusiasm has to meet constraints, evidence, and consequences. Trying to do both at once produces either a timid list or a glorious pile of ideas that nobody can choose between.

Start by naming the decision

“Which idea is best?” is not a decision. Best for what, for whom, by when, and with what resources?

A useful decision statement has four parts:

• Outcome: what must improve or exist?
• Audience: whose needs decide whether it works?
• Constraint: what cannot change?
• Horizon: when must the idea show value?

For example: “Choose a workshop concept that gives first-time visitors a satisfying result in 90 minutes, works with our existing room, and can launch next month.” That is evaluable. “Choose our most exciting workshop” is an invitation to spend Thursday debating the spiritual meaning of exciting.

Write the statement at the top of the page. If people disagree with it, settle that disagreement before scoring anything. A polished matrix cannot repair a fuzzy decision.

Keep generation and judgment in separate rooms

If you still have only two obvious candidates, you may not be ready to evaluate. Generate first, using a deliberate method rather than squeezing the same thought harder. Our guide to brainstorming techniques includes solo and group formats for producing a wider field.

Then pause. Even a short break helps everyone change roles from inventor to editor. When evaluation begins, copy the ideas into a clean list without the name of the person who proposed each one if authorship is likely to distort the discussion.

Do not let the loudest advocate deliver a campaign speech before the first comparison. Clarifying an idea is useful; auditioning for its fan club is not.

Step 1: Remove ideas that fail hard requirements

Hard requirements are pass-or-fail conditions. They are not preferences wearing a serious expression.

Examples include:

• a legal, safety, or accessibility requirement;
• a fixed launch date;
• a maximum budget that truly cannot move;
• compatibility with a required format or platform;
• a promise already made to an audience or client.

Keep this list short. If every desirable quality becomes mandatory, all interesting ideas die at the gate and the safest beige rectangle wins.

For each rejected option, record the failed requirement. That prevents the group from rediscovering and re-rejecting the same idea at the next meeting. It also shows whether changing one constraint later could revive a strong concept.

Step 2: Choose criteria that reflect the outcome

Three to five criteria are usually enough for a working decision. More criteria create administrative fog and make weak factors cancel strong ones.

Useful criteria might include:

Avoid vague labels such as “impact” unless everyone can describe what evidence would count. Define each criterion in one sentence and, if needed, add examples of a strong and weak result.

Originality deserves special care. A strange idea is not automatically valuable, and a familiar idea is not automatically lazy. The question is whether the idea is different in a way that helps. If your list feels trapped by familiar assumptions, run one of these outside-the-box exercises before evaluating—not halfway through the scorecard.

Step 3: Weight only the differences that matter

Equal weighting is a good default. It is simple, visible, and hard to manipulate. Add weights only when one criterion genuinely matters more to the decision.

Use a small scale, such as:

• 3 — decisive: this factor can make or break the choice;
• 2 — important: it should influence the result substantially;
• 1 — useful: it matters, but not at the expense of the essentials.

Agree on weights before scoring the ideas. Otherwise, criteria have a funny habit of becoming “decisive” immediately after somebody's favorite concept performs badly on them.

Weights express judgment; they do not turn judgment into physics. Keep the discussion attached to the reasoning behind the number.

Step 4: Score with notes, not numbers alone

A simple 1-to-5 score can make comparisons visible:

• 1: clearly weak or unsupported;
• 3: plausible but uncertain or ordinary;
• 5: strongly supported for this criterion.

Every score should have a short note: what do we know, what are we assuming, and what remains uncertain? “Audience value: 5” tells the next reader nothing. “Five recent support requests ask for this exact capability” explains the judgment and can be checked.

Score independently before discussing if a group contains status differences or strong personalities. Then compare the spread. Averages are less interesting than disagreement. If one person gives feasibility a 2 and another gives it a 5, the useful question is not “Can we compromise at 3.5?” It is “What do you know that the rest of us do not?”

The matrix should reveal the conversation you need. It should not end the conversation by pretending arithmetic has become project manager.

Step 5: Look for fatal weaknesses and useful combinations

Before ranking totals, inspect each idea's shape.

An option with consistently good scores may be safer than one with spectacular audience value and impossible delivery. Conversely, a modest total can hide a brilliant concept whose only weakness is a solvable production problem.

Ask:

1. Does any low score create a fatal problem?
2. Could a small change remove that weakness?
3. Are two ideas strong in complementary ways?
4. Is the leading idea merely easiest, or genuinely most useful?
5. What would have to be true for this idea to succeed?

Combining ideas can work, but beware the committee chimera: the speed of one proposal, the depth of another, the audience of a third, and somehow no additional cost. A hybrid must be evaluated as a new idea with its own trade-offs.

Step 6: Run the cheapest useful test

When uncertainty, not preference, separates the finalists, stop discussing and create evidence.

The test should target the riskiest assumption. Depending on the work, that might mean:

• sketching three screens instead of building the product;
• reading one scene aloud instead of rewriting the whole script;
• making a paper prototype;
• showing two headlines to representative readers;
• pricing the difficult component with an actual supplier;
• running one small session before booking a season.

Cheap does not mean sloppy. The test needs to resemble the real question closely enough to teach you something. Asking friends whether they “like the idea” mostly measures friendship and their desire to finish the conversation.

Decide in advance what result would change the ranking. Otherwise every response can be interpreted as support for the idea somebody already wanted.

Make the decision and close the loop

Choose the leading option, name the decision owner, and record:

• the selected idea;
• the criteria that mattered most;
• the largest remaining uncertainty;
• the first review date;
• what evidence would justify changing course.

Then stop evaluating for a defined trial period. Constantly reopening the choice starves every idea of the time required to become good. A decision is not a wedding vow, but it should survive one inconvenient afternoon.

Keep the strongest rejected ideas in a parking lot with the reasons they lost. Some are wrong for this project and right for the next one. Others are simply wearing a magnificent hat and have no job.

A reusable 20-minute evaluation session

For a small decision, use this compact format:

1. Two minutes: read and agree on the decision statement.
2. Three minutes: confirm hard requirements.
3. Three minutes: choose and define four criteria.
4. Five minutes: score silently with one note per score.
5. Five minutes: discuss the largest disagreements and fatal weaknesses.
6. Two minutes: choose a winner or assign one specific test.

If the decision cannot fit that format, the issue may be genuinely complex—or the group may be trying to solve several decisions at once. Split them.

The final rule

An idea-evaluation method succeeds when it makes the trade-offs visible and produces a next move. It fails when it gives a decimal-point blessing to the option everyone preferred before the meeting.

Define the decision, use a few meaningful criteria, attach notes to scores, and test the uncertainty that matters most. Then choose. Good creative work needs imagination, but sooner or later it also needs someone to point at one door and walk through it.