Login Create Account

Conjoint Study Design Generator

Design Tool Analysis Tool →

Create optimized fractional factorial designs for choice-based conjoint (CBC) studies. Define attributes and levels, specify constraints, and generate balanced experimental designs ready for data collection.

DESIGN METHODOLOGY

What is a Choice-Based Conjoint (CBC) Design?

In a CBC experiment, respondents choose between hypothetical products that differ on a set of attributes (e.g., brand, price, features). By analyzing which combinations people prefer, we can estimate how much each attribute level contributes to overall desirability.

The challenge: if you have many attributes and levels, the total number of possible products (the full factorial) explodes quickly. A fractional factorial design selects a carefully chosen subset so you can still estimate every attribute’s effect without showing respondents thousands of scenarios.

Three properties make a design “good”:

  • Balance — every level of every attribute appears about equally often, so no level is under-represented.
  • Orthogonality — attribute levels vary independently of each other, so you can tell which attribute drove a choice.
  • D-Efficiency — a single 0–100% score that summarizes how much statistical information the design extracts per observation, compared to the full factorial.

Design Quality Metrics

  • D-Efficiency: 0–100% score comparing your design’s information content to the full factorial. Higher = you need fewer respondents for the same precision.
  • Level Balance: Are all attribute levels shown roughly equally? Imbalance wastes observations on some levels at the expense of others.
  • Orthogonality: Can each attribute’s effect be measured independently? Poor orthogonality means attributes are confounded (entangled).
  • Statistical Power: Do you have enough observations per parameter to detect real preference differences?

Typical Design Parameters

Tasks per Respondent: 8–15 choice scenarios (more data per person, but watch for survey fatigue)

Alternatives per Task: 3–4 products to choose from (more = harder cognitive task)

Attributes: 3–8 features per product (more attributes = larger sample needed)

Levels per Attribute: 2–6 values (single-level attributes contribute nothing)

None Option: Lets respondents say “I wouldn’t buy any of these” — recommended for realism

CHOOSE YOUR STARTING POINT

📚

Start from a Case Study

Load a pre-configured study to see how the generator works, then customize it.

🛠️

Build from Scratch

Define your own product attributes and levels from a blank canvas.