Diverge then converge: generate concepts that aren't obvious

Complete Pressure-test your brand against competitors first; with your difference clear, this lesson puts it to work generating concepts that are yours, not the obvious ones everyone reaches for.

The first idea Claude gives you is the idea everyone gets. The good one is usually hiding behind nineteen worse ones.

Real creative thinking happens in two separate moves, and most people skip the first. The two moves are diverge (generate lots of options, widely, without judging) and converge (narrow down and pick the strongest). Doing both at once is why ideas come out safe.

Why "give me the best idea" fails is that you are asking Claude to diverge and converge in one breath. It plays it safe and hands you the obvious answer. Asked to be excellent in one shot, it gives you the average, because the average is the safest bet.

Diverge first, on purpose. Ask Claude for twenty concepts, deliberately different from each other, including a few strange ones, with no judging yet. Quantity is the goal here, not quality. You are widening the field so the surprising idea has room to show up.

Then converge, as a second step. Now you switch hats. You pick the three that spark something, tell Claude why, and ask it to push those further or combine two into one. Judging belongs here, after the options exist, never before.

Here is the before and after: Someone asks "what is a good campaign idea for my tea shop" and gets "cozy autumn vibes," which is fine and forgettable. Someone else asks for twenty wild directions, spots "a tea for every bad day," and builds a whole honest, funny campaign from the one nobody expected.

Make the bad ideas first by diverging wide, then put your judgment to work converging, because the idea worth using rarely arrives first or alone.

I want to run a proper two-step idea session with you, not get one safe answer. We will diverge first, then converge.

What I need ideas for: [DESCRIBE, e.g. "a launch campaign for my handmade candle brand"].
My brand in one line (voice + what makes me different): [FROM MODULE 1, or describe].

Step 1 (diverge): give me 20 concepts that are deliberately different from each other. Include a few strange or risky ones. Do NOT judge or rank them yet. One line each.

After I read them I will pick a few, and then we do Step 2 (converge): I tell you which ones spark something and why, and you push those further or combine two into one stronger idea.

Here is the two-step session, for a candle brand:

**Step 1, diverge (20 quick directions, no judging):**
1. A scent for every room of a memory.
2. Candles named after specific rainy days.
3. "Burn this when you need to leave the party early."
4. Scents that smell like places that no longer exist.
5. A candle that is also a permission slip to rest.
... (15 more, including a few odd ones)

**Step 2, converge (after you pick):**
You loved 3 and 5. Both are about giving yourself permission. Combined idea: a small collection called "Off Duty," each candle a permission slip in scent form, with a one-line note on the label. That came from the strange options, not the safe "cozy candles for autumn" you would have shipped otherwise.

What good looks like

• You ran diverge and converge as two separate steps, not one combined request
• You got at least 15 to 20 genuinely different concepts before any judging happened
• You picked a few that sparked something and told Claude why, instead of grabbing the first
• Your final idea came from pushing or combining the surprising options, not the obvious one