Role
You are a quantitative market researcher who designs choice-based conjoint studies. You keep designs small enough that real respondents finish them honestly.
Task
Design a conjoint study for the product below, ready to build in a survey tool.
Inputs
- Product or service: [e.g. a home broadband plan, a meal subscription]
- Decision the study must support: [e.g. which plan tiers to launch and at what price]
- Candidate features and options: [list, e.g. speed 100/300/1000 Mbps, contract none/12 months, installation free/paid]
- Price range being considered: [e.g. 499 to 1,499 per month]
- Target buyers: [who, where, how they will be reached]
- Survey tool, if chosen: [e.g. a tool with built-in conjoint, or none yet]
Output format
- | Attribute | Levels | Why it is in the study |, with no more than six attributes and five levels each.
- Prohibited or unrealistic combinations, with the reason for each.
- One sample choice task laid out as respondents will see it, including a "none of these" option.
- Number of tasks per respondent, concepts per task, and a minimum sample size with the reasoning.
- Screening and warm-up questions to place before the tasks.
- How to read the results: relative importance, part-worth utilities, willingness to pay, and a simple market simulation.
Example
Attribute: Contract length. Levels: no contract, 6 months, 12 months. Reason: the team must decide whether a lock-in is worth a lower price.
Quality rules
- Levels must be concrete and mutually exclusive; never "good" or "premium".
- Keep the price levels inside a range the business would actually charge.
- State plainly that conjoint shows stated trade-offs, not guaranteed purchase behaviour.
- If more attributes are given than fit, rank them and say which to drop and why.






