The Comparison
After this drill, you can build a structured comparison framework for any real decision — options, criteria, weights, and recommendation.
Why this matters
Decisions are hard when you have multiple options with different trade-offs across multiple criteria. AI is excellent at forcing that structure: making trade-offs explicit, weighting criteria, and cutting through the fog of vague preferences. This drill gives you a reusable pattern for any decision where you have at least two options and at least three criteria.
How to do it
- 1
Pick a real decision you are facing or have recently made
Best decisions to use: tool selection (which CMS to use, which vendor to hire), career choices, major purchases, strategic directions. Must have at least 2 options and 3 criteria.
- 2
Use the comparison framework prompt
Tell Claude the decision, list your options, and let it propose the criteria — then refine them based on what actually matters to you.
- 3
Assign weights to the criteria
After seeing the criteria, assign weights (1–5) based on importance to your specific situation. This is where your judgment matters — Claude doesn't know your priorities.
- 4
Review the weighted recommendation and ask "what am I missing?"
The final step is adversarial: ask Claude to argue for the option it ranked lowest. This surfaces blind spots in your criteria.
The prompt
I need to decide between these options: [LIST YOUR OPTIONS] The decision is about: [WHAT YOU ARE CHOOSING] My context: [2-3 sentences about your situation, timeline, and what matters most] Please: 1. Propose 6-8 criteria I should evaluate these options against (based on my context) 2. Create a comparison table with these criteria as rows and options as columns 3. Score each option on each criterion (1-5 scale with brief justification) I will then assign weights to the criteria and ask for the weighted recommendation.
Based on the comparison above, here are my criteria weights (1=low importance, 5=critical): [LIST EACH CRITERION AND ITS WEIGHT] Now: 1. Calculate the weighted score for each option 2. Give me the recommendation with the highest weighted score 3. Then argue for the option that ranked lowest — what would have to be true for it to be the better choice?
Success criteria
- ✓You built a comparison table with at least 2 options and 4+ criteria
- ✓You assigned weights based on your actual priorities
- ✓You reviewed the devil's advocate argument for the losing option
- ✓The recommendation matches or clarifies what you already believed — or genuinely surprised you
Common mistakes
Listing too many correlated criteria (e.g. "cost" and "price" are the same)
→ Each criterion should be independent. If two criteria always move together, merge them into one.
Assigning equal weights to everything
→ If everything is equally important, the comparison is useless. Force yourself to identify what matters most in your specific situation.
Skipping the devil's advocate step
→ This step consistently reveals criteria you forgot. The option you ranked lowest usually has a real advantage in one dimension you underweighted.