Frame the challenge
What exactly are you designing, and what must any good answer do?
- What is the one thing this session should produce?
- Who is it for, and what would delight them?
- Which limits are truly fixed: budget, time, rules?
Map every possible answer in a grid, then combine options across it to find designs nobody proposed.
Morphological Analysis breaks a design problem into its independent dimensions, such as place, format, price or audience, and lists several options for each. Laid out as a grid, those few short lists hold a surprising number of possible solutions.
Five dimensions with five options each give you 3,125 combinations. You won't read them all. Instead you pick one option from each column, sometimes logically and sometimes at random, and see what the combination suggests. Many paths are useless. A few are ideas nobody on the team would have proposed.
Use it when you're designing something with several moving parts (an event, a product, a service or a program) and want to see the whole space of options instead of polishing the first idea that came up.
One option per column makes one party. This path became a detective clue trail ending in a picnic in the park.
What exactly are you designing, and what must any good answer do?
Which independent choices does every possible solution have to make?
What different answers could each parameter take?
What happens when you join one option from every column?
Which combinations are workable, and which deserve a closer look?
Morphological Analysis looks mechanical, but it fixes several habits that quietly limit creative work.
Teams usually develop the first decent idea and polish it. A grid lays out every dimension at once, so you can see how much of the space you never looked at, and where the unexplored corners are.
Joining options that don't normally go together, like a picnic and a detective story, creates ideas that free thinking rarely reaches. The grid does the forcing, so you don't have to wait for a flash of inspiration.
You only need a handful of options per column, which is easy. Multiplication does the rest: five short lists can hold thousands of distinct solutions.
Because every idea is described in the same terms, you can compare combinations side by side, rule out clashing pairs in one move, and explain to others why one path won.
Write one sentence saying what you're designing and what it must achieve. “A fundraiser that raises $4,000 without burning out our volunteers” gives the grid a clear purpose.
List the choices any solution must make, then keep the 3–6 that matter most and are independent of each other. Fewer, sharper columns beat many fuzzy ones.
Under each parameter, list 4–6 options. Include the obvious ones, then push for at least one unusual option per column. This is a creative step, so don't judge yet.
Draw paths through the grid, one option per column. Try a few logical paths, a few random ones (roll a die for each column), and a few that keep one bold choice fixed while the rest vary.
Cross out pairs that clearly clash. Describe the 5–10 most promising paths in a sentence each, then shortlist with a quick Concept Screening or NUF Test.
Turn your favorite path into a real concept: give it a name, sketch how it works, and test its riskiest part first.
A big grid on a whiteboard or spreadsheet (parameters across the top, options below), sticky notes, a die and about an hour. With AI, start with the prompt below.
Subject: the annual fundraiser for a volunteer-run youth soccer league. Goal: raise $4,000 for uniforms and field fees. Constraint: each volunteer can give about ten hours.
Activity, place, who gives, timing and money source, with four options each: 1,024 possible fundraisers on one whiteboard.
Skills contest + home field + far-away relatives + during games + pledges. Players try five skills at halftime; relatives pledge per success.
Trivia night + downtown café + local shops + one Saturday + sponsorship. Each shop sponsors a round and writes its own questions.
A random roll: car wash + online + far-away relatives + raffle. Odd, but it sparked a twist: relatives pay online to gift a wash to someone local.
Silent auction + during games + neighbors. Ruled out fast: volunteers can't run an auction from the sidelines.
14 paths described, 4 ruled out as clashes, 3 shortlisted. The board chose the Halftime Challenge.
The grid multiplies fast, which is both its strength and its weakness: nobody can read 1,024 combinations. AI can suggest options for every column, walk dozens of paths in seconds and flag the clashes. You bring what no model knows: which options are realistic in your world, and which combination people will get excited about.
Ask AI whether your columns overlap or miss a key dimension.
Here are my 5 parameters. Which overlap, which are missing, and which could I merge?
Push one column well past the obvious options.
Give me 8 options for 'Where' for a youth soccer fundraiser, from normal to strange.
Let AI roll the dice so you meet combinations you'd never choose.
Pick 6 random paths through my grid and turn each into a one-sentence concept, however odd.
Have AI rule out incompatible pairs so you can focus on the live paths.
List the option pairs in my grid that clearly can't work together, and say why in a few words.
Describe what you're designing, pick a mode, then copy the prompt into Claude, ChatGPT, Gemini or Copilot. In coach mode the AI builds the grid with you column by column, then walks paths through it with you.
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.