Score each idea on three questions: is it New, is it Useful, and is it Feasible?
An idea can be exciting but useless, useful but impossible, or practical but exactly what everyone already does. The NUF Test catches all three problems with three quick questions. Is it New: different from what's been tried? Is it Useful: does it actually solve the problem? Is it Feasible: can you really make it happen?
You score each idea from 1 to 5 on each question and add up the total. The total helps you rank, but the pattern teaches more. A high-New, low-Feasible idea needs development, not the bin. A high-Feasible, low-New idea is safe but may not be worth the effort. The weakest score tells you exactly what to work on next.
Use it when you have a shortlist and want a fast, fair comparison, when a group keeps arguing about which idea is best, or when you want to know what to improve in one promising idea.
The best ideas sit where all three overlap. A low score on any one shows what to work on next.
Choose a part of the diagram or a tab to explore each question.
How different is this from what's already been done, here or elsewhere?
How well does it fix the real problem, and for how many people?
Can you make this happen with the time, money, skills and support you have?
The NUF Test takes minutes, yet it guards against the most common ways that good-looking ideas go wrong.
Ideas usually fail in one of three ways: they aren't really different, they don't solve the problem, or they can't be done. Checking each one separately stops a single strength, like novelty or ease, from hiding a serious weakness.
Teams tend to drift to one extreme: chasing exciting ideas that never happen, or approving safe ones that change nothing. Putting New next to Feasible keeps both kinds of value in view.
A total tells you which idea is ahead. The separate scores tell you why, and what to do next: make it more useful, find a cheaper way to deliver it, or push for something more original.
There's no scoring system to design and no long debate about how to judge. Three plain questions that anyone can learn in a minute mean you can score ten ideas in a coffee break, so people actually use it.
Write each idea on the shortlist as one clear sentence. The test works best on three to ten ideas; thin a longer list first with Negative Selection.
Before scoring, agree what 1 and 5 mean in your situation. For example, Feasible 5 means “we could start next week with what we have”, and Feasible 1 means “we'd need money or approval we don't have”.
Each person scores every idea from 1 to 5 on New, Useful and Feasible, alone and without discussion. Score one criterion across all the ideas at a time; it keeps judgments consistent.
Put the scores side by side. Talk about the biggest disagreements first, since they often reveal different assumptions. Agree on a final score for each idea, then total them.
Rank by total, then look at the shape. High New, low Feasible: develop it. High Feasible, low New: fine for a quick win. Low Useful: rethink it. Choose what to pursue and what to improve.
Your shortlist, a simple grid with columns for New, Useful, Feasible and Total, a pen for each person and 15–30 minutes. With AI, use the prompt below.
Problem: food waste in a 600-student middle school cafeteria. Scorers: four student council members and the cafeteria manager. Scale: 1–5 per criterion, scored alone, then discussed.
A table where students leave unopened food for others. N 2 · U 3 · F 5. Simple, but plenty of schools already do it.
Weigh each day's scraps and post the results by grade. N 3 · U 4 · F 5. Turns waste into a visible, friendly competition.
Play first, eat second, so students arrive hungry and don't rush. N 4 · U 5 · F 2. Needs the principal's support and a new timetable.
Students taste-test new menu items before they're served. N 4 · U 4 · F 3. Needs the district menu team on board.
Students run a compost bin for the school garden. N 3 · U 2 · F 4. Good for learning, but it doesn't reduce waste at the source.
Recess before lunch scored highest on Useful but lowest on Feasible, so the council chose to build a case for it rather than drop it.
AI makes a useful second scorer for a NUF Test. It can spot that your “new” idea is common elsewhere, list the practical obstacles you've skipped, and suggest how to lift a weak score. But it doesn't know your budget, your people or what's been tried locally, so score first yourself and treat its numbers as a second opinion.
Score first yourself, then have the AI score without seeing yours, and compare.
Score these 5 ideas 1-5 on New, Useful and Feasible, with a one-line reason each. Don't ask for my scores yet.
Find out whether your “new” idea is already common somewhere else.
Where have ideas like this been tried before? What happened, and should that change my New score?
List what it would really take to make the idea happen.
What would it actually take to move recess before lunch? List the costs, approvals and the biggest obstacle.
Turn the weakest letter into a design challenge instead of a reason to reject.
This idea scores 2 on Feasible. Give me three ways to make it more doable without losing what makes it useful.
List your ideas, pick a mode, then copy the prompt into Claude, ChatGPT, Gemini or Copilot. In coach mode the AI asks for your scores first, one criterion at a time, then compares them with its own and helps you read the pattern.
The “Open in” links put the prompt in the web address. If your challenge includes anything sensitive, copy and paste instead.
After the session, click one to copy it, then paste it into the same chat.