What you will learn
- Generate a wide range of ideas before judging any of them
- Use constraints to force more original options
- Combine unrelated ideas into something new
- Score ideas against practical criteria
- Turn a vague idea into a concrete next step
- Run a second round that improves the best ideas
Separate generating from judging
The most common brainstorming mistake is judging each idea the moment it appears. That kills the session early, because the first few ideas are usually the obvious ones. Ask Claude for a large set of ideas first, with no evaluation attached, and only start judging once the list exists. Quantity first is not a slogan; it is how you get past the obvious answers to the interesting ones.
Set a number. Ask for twenty or thirty ideas rather than "some ideas." A target forces variety, because Claude has to move past the first handful of safe options to reach the target. If the first list feels samey, say so and ask for a second list that avoids the patterns in the first.
Constraints make ideas better
A blank page produces generic ideas. A constrained brief produces specific ones. Give Claude the real limits: the budget, the audience, the time available, the channels you can actually use, and anything that is off the table. Then ask for ideas that work inside those limits. You will get fewer fantasy options and more ideas you could actually try.
Try changing one constraint at a time to see how the ideas shift. What would you do with half the budget? What if the audience were beginners instead of experts? Each variation is a fresh round of ideas, and comparing the rounds often reveals which constraint was shaping the answers most.
Combine, twist, and invert
When the list stalls, ask Claude to combine two ideas from the list into one, to take an idea and apply it to a different audience, or to invert an idea and see what the opposite looks like. These moves are simple, but they reliably produce options that a straight list would not reach. Name the two ideas you want combined rather than leaving the pairing to chance.
Another useful move is to borrow from another field. Ask how a restaurant, a gym, or a school would solve the same problem, then translate the answer back to your situation. The translation step is where the originality usually appears.
I run a small online shop selling handmade candles. Give me 20 marketing ideas that cost under $100 each, aimed at buyers aged 25 to 40, and usable within two weeks. No paid ads and no influencer deals. After the list, pick the three you think are most likely to bring repeat buyers, and say why in one sentence each.
Score ideas before you pick one
Once you have a list, judge it against criteria that matter to you: cost, time to try, likely reach, and fit with what you already do. Ask Claude to score each shortlisted idea against those criteria and to name the biggest risk for each. A scored shortlist is easier to discuss with a partner or a team than a long list of unranked options.
Be honest about the criteria. If speed matters more than perfection, say so, and let the scoring reflect it. If an idea scores well but depends on a skill you do not have, that dependency belongs in the risk column, not hidden.
Turn the winner into a next step
An idea is not a plan. For the idea you pick, ask Claude to write the first three concrete steps, what each step needs, and how you will know whether it worked. A test with a clear measure beats a grand launch: try the idea small, look at the result, and then decide whether to expand it.
Keep the ideas you did not pick. A shortlist that loses today may fit next quarter, when the budget, the audience, or the season has changed. Store the list with the criteria you used, so the reasoning is still clear when you return to it.
Tip: Run at least two rounds. The first round finds the obvious ideas; the second round, which bans the patterns of the first, is where the more original options usually appear.
Ideas Claude is good for
Claude works well for names, headlines, product features, content topics, workshop exercises, and business offers, because those are all lists that benefit from volume and variation. It works less well for ideas that depend on facts it cannot check, such as what your competitors charge or what your customers actually said last week. For those, bring the facts yourself and let Claude work from them.
Treat every brainstorm as raw material. Your judgement about taste, timing, and what your audience will accept is the part no tool can replace, and it is the part that turns a long list into a good decision.
Frequently asked questions
How many ideas should I ask Claude for?
Ask for a set number, such as twenty or thirty. A target forces variety and pushes past the first few obvious answers.
Why are my brainstorming results so generic?
Usually the brief is too open. Add real constraints: budget, audience, timeline, and what is off the table. Constraints produce specific ideas.
Can Claude help me pick the best idea?
Yes, if you give it criteria. Ask it to score shortlisted ideas on cost, time, reach, and fit, and to name the biggest risk for each.
What if I like an idea but cannot execute it?
Ask Claude to break it into smaller steps and to name the skills or resources each step needs. Often a smaller version of the idea is testable right away.
Should I keep ideas I do not use?
Yes. Store the shortlist with the criteria you used. An idea that loses today may fit next quarter when conditions change.