What Is Conjoint Analysis?
Conjoint analysis is a survey-based market research method that measures consumer preferences by requiring respondents to make trade-offs between product attributes rather than evaluating each feature in isolation. This forced-choice structure reveals the relative importance of each attribute and the implicit value — called part-worth utility — that consumers assign to each feature level.
In practice, a respondent might compare two industrial pump configurations: one priced at $500 with a two-year warranty and standard support, and another priced at $750 with a five-year warranty and dedicated support. By observing which option respondents choose across many such tasks, analysts can back-calculate precisely how much each attribute — price, warranty length, support tier — is worth to the buyer.
How Conjoint Analysis Works
- Define attributes and levels. An attribute is a product dimension such as price or warranty. A level is a specific value of that attribute — for example, $500 or $750. Practitioners typically cap attributes at six to eight to prevent respondent fatigue.
- Design the survey instrument. An orthogonal or D-optimal experimental design ensures that attribute combinations are statistically independent, preventing multicollinearity from distorting preference estimates. Poor design at this stage undermines the entire study.
- Present choice tasks. In the dominant variant — Choice-Based Conjoint (CBC) — respondents select a preferred profile from sets of two or more complete product descriptions per task. This mirrors real purchase decisions more closely than rating exercises.
- Estimate part-worth utilities. Responses are modeled via multinomial logit for aggregate-level results or Hierarchical Bayes (HB) for individual-level estimates. HB is preferred when the goal is segmentation or individual-level willingness-to-pay analysis.
- Simulate market scenarios. Derived utilities feed a market simulator that predicts how share of preference shifts when one attribute level changes — for example, raising price by $50 or adding a premium support tier. This simulation output is where conjoint results connect directly to pricing decisions.
Conjoint Analysis vs. MaxDiff Analysis
Both are survey-based preference methods, and practitioners regularly conflate them. The table below clarifies where each method applies.
DimensionConjoint AnalysisMaxDiff AnalysisPrimary purposeQuantify trade-offs between full product configurationsRank importance of individual items or featuresWhat respondents doChoose among complete product profilesIdentify the best and worst item in a small setKey outputPart-worth utilities; willingness to payRelative importance scores across a long listBest used whenPricing or configuration research across attributesFeature prioritization or message testing
Use conjoint analysis when you need to quantify trade-offs between product configurations or price points; use MaxDiff when you need to rank a long list of features or messages by importance without defining full product profiles. For simpler price-range research — when a full attribute design is unnecessary — the Van Westendorp Price Sensitivity Meter offers a lighter-weight alternative focused specifically on acceptable price boundaries.
Common Variants
The right conjoint variant depends on attribute count, decision complexity, and whether pricing is a study objective.
- Choice-Based Conjoint (CBC) — Respondents choose among complete product profiles per task; the most widely used variant for pricing research because it closely replicates real purchase behavior.
- Adaptive Conjoint Analysis (ACA) — The survey adapts in real time based on earlier responses to reduce respondent burden; suited to studies with nine or more attributes but generally less reliable for price estimation.
- Menu-Based Conjoint (MBC) — Respondents build a custom product by selecting from optional features, each with an associated price; best suited to configurable or engineered-to-order products.
- Full-Profile (Traditional) Conjoint — Respondents rate or rank complete profiles; the earliest form of conjoint, largely superseded by CBC for pricing work due to susceptibility to social-desirability bias.
Limitations and Risks
Conjoint analysis is rigorous, but it carries meaningful limitations practitioners should understand before commissioning a study.
- Hypothetical bias. Survey choices may not match actual purchase behavior. This is the core limitation of all stated-preference methods and tends to inflate willingness to pay relative to real-market results.
- Attribute omission bias. Excluding a relevant attribute inflates the apparent importance of the attributes that are included. Careful attribute selection during the design phase is the primary mitigation.
- Sample size requirements. CBC with Hierarchical Bayes estimation requires an adequate sample to produce stable individual-level utility estimates. Under-powered studies yield unreliable part-worths and should not be used to set prices.
- Scope limits. Conjoint is inappropriate for products with 30 or more attributes, very low-involvement commodity purchases, or entirely novel product categories where respondents lack reference points for making meaningful trade-offs. Most introductory treatments skip this "when not to use" context, but it is critical for avoiding costly research design errors.
Related Terms: Price Sensitivity Analysis | Van Westendorp Price Sensitivity Meter | MaxDiff Analysis | Willingness to Pay | Price Optimization
