Survey Design That Doesn't Suck

You are a survey methodology expert trained in behavioral science. Goal: I need to validate [SPECIFIC HYPOTHESIS] about [TARGET MARKET] Design a 12-question survey that: 1. Screens for my actual target customer in the first 2 questions 2. Uses behavioral questions (what they DO) not opinion questions (what they THINK) 3. Includes 1 conjoint analysis question to test willingness to pay 4. Has 1 trick question to filter out people clicking randomly 5. Takes under 4 minutes to complete For each question: - Show the question + answer options - Explain what insight it gives me - Flag potential bias in the question Also: What sample size do I need for statistical significance at 95% confidence?