idea-evaluation

name	idea-evaluation
description	Evaluate these ideas, which idea is best, help me prioritize, compare these options, rank these, how do I choose. Structured evaluation via Impact/Effort Matrix, Assumption Mapping, Pre-Mortem, JTBD, weighted scoring.
argument-hint	[topic or problem]

Idea Evaluation — Structured Convergent Thinking

Evaluate, prioritize, and select the strongest ideas using proven evaluation frameworks. This skill is ONLY about evaluation — ideation should already be complete.

Required reading: Read references/facilitation-principles.md before facilitating. Its anti-anchoring and engagement-signal rules govern every framework below.

Routing: Solo scoring frameworks live here. High-stakes decisions that merit opposed expert lenses → /rad-council:convene.

Hard Rule: The User Scores First

In EVERY framework below that involves scoring, rating, or ranking: the user assigns their scores before you offer your perspective. Then share yours and discuss the differences — the conversation about WHY scores differ is where the value lives. If you co-assign the numbers, you end up ranking your own ideas and anchoring returns through the back door.

Mode Announcement

"We're in evaluation mode now. We'll look at each idea critically and constructively to find the strongest candidates."

Step 1: Gather the Idea Set

If ideas aren't already listed:

• Ask the user to enumerate all candidate ideas
• Ensure the list is complete: "Are there any other options we should consider before evaluating?"
• Number each idea for easy reference

Step 2: Select Framework

Reference references/evaluation-frameworks.md for detailed guidance on each.

Situation	Recommended Framework	Why
10+ ideas, need quick triage	Impact/Effort Matrix	Fast visual sorting into 4 quadrants
3-5 ideas, need deep validation	Assumption Mapping	Tests what needs to be true for each to work
Product/feature ideas	Opportunity Solution Trees	Connects solutions to user outcomes
High-stakes decision	Pre-Mortem + Assumption Mapping	Catches blind spots and untested assumptions
2-3 finalists, need to compare	Weighted Scoring (multi-criteria)	Transparent, structured comparison
"Who is this for?" unclear	Jobs-to-be-Done lens	Reframes around user needs

Recommend a framework and explain why, but let the user choose.

Step 3: Apply the Selected Framework

Impact/Effort Matrix

1. Define "impact" criteria with user (business value? user value? strategic alignment?)
2. Define "effort" criteria (development time? complexity? dependencies? risk?)
3. For each idea, score impact (1-5) and effort (1-5) — user's scores first, then yours, then discuss differences
4. Place on the 2x2 grid:
   • Quick Wins (high impact, low effort) — Do these first
   • Major Projects (high impact, high effort) — Plan carefully
   • Fill-ins (low impact, low effort) — When time allows
   • Hard Slogs (low impact, high effort) — Deprioritize or drop
5. Present the grid and recommended priority order

Assumption Mapping

1. For each top idea, ask: "What needs to be true for this to work?"
2. Generate assumptions across: Desirability, Feasibility, Viability, Adaptability
3. For each assumption, assess:
   • Importance: How critical? (High/Medium/Low)
   • Evidence: How much proof? (Strong/Some/None/Contradicted)
4. Plot on 2x2: Importance vs Evidence
5. Focus on top-left quadrant (high importance, low evidence) — these are the riskiest assumptions
6. For each risky assumption, suggest how to test it cheaply

Pre-Mortem

1. "Imagine it's 6 months from now. Idea [X] was implemented and failed spectacularly. Why?"
2. Generate 5-8 failure scenarios together (user first, then brainstormer adds)
3. For each: likelihood (H/M/L) and preventability (Y/P/N)
4. High-likelihood, preventable failures → these become design requirements
5. High-likelihood, non-preventable failures → these might kill the idea

Opportunity Solution Trees

1. Clarify the desired outcome (the metric to move)
2. Map opportunities (user needs, pain points) under the outcome
3. Map each idea under the opportunity it addresses
4. Check: are some opportunities over-served (many solutions) while others are empty?
5. Focus on the highest-value opportunity with the fewest good solutions

Jobs-to-be-Done Lens

1. For each idea, ask: "What job is the user hiring this to do?"
2. Explore three dimensions:
   • Functional: What task do they need to accomplish?
   • Social: How do they want to be perceived?
   • Emotional: How do they want to feel?
3. Ideas that serve all three dimensions tend to be strongest
4. Ideas that only serve functional jobs may be commoditized easily

Weighted Scoring (Multi-Criteria)

1. Define 3-5 criteria (e.g., impact, feasibility, novelty, alignment, urgency)
2. Weight criteria if needed (user decides)
3. Score each idea 1-5 on each criterion — user's scores first, then yours
4. Calculate weighted totals
5. Present as ranked table: | Idea | Impact | Feasibility | Novelty | Alignment | Total |
6. Discuss surprises — does the ranking match intuition?

Step 4: Challenge Top Candidates

For the top 2-3 ideas, optionally run a deeper challenge:

• Dispatch the idea-challenger agent for systematic stress-testing — it's a substantial dispatch, so offer it in its own message only when the stakes warrant it, never as an upfront menu item
• Or do a quick inline challenge: "Playing devil's advocate on Idea [X] — what could go wrong?"

Step 5: Present Results

Clear, ranked output:

## Evaluation Results

### Recommended: [Idea Name]
**Why:** [2-3 sentences on why this won]
**Riskiest assumption:** [What to test first]
**Next step:** [Concrete action]

### Strong Alternative: [Idea Name]
**Why:** [What makes this worth considering]
**Trade-off vs recommended:** [What you gain/lose]

### Not Recommended (for now): [Ideas that didn't make the cut]
**Why:** [Brief reason for each]

Step 6: Transition

Based on what the evaluated ideas are for:

• Software project → suggest design-sprint skill
• Business/product → suggest action plan or experiment design
• Content → suggest content strategy
• General → suggest next concrete steps

Ask: "Ready to move forward with [recommended idea]? If so, I can help with [appropriate next step]."