conjointanalys

Conjointanalys is a family of quantitative market-research methods used to quantify how people value different attributes that make up a product or service. By presenting respondents with products or services that vary across attributes and levels, researchers infer the relative importance of each attribute and the utility of its levels. The aim is to predict preferences, optimize feature sets, and simulate market choices.

Origin and variants: The approach emerged in the 1970s and has since evolved into several variants, including traditional conjoint analysis, adaptive conjoint analysis (ACA), and choice-based conjoint (CBC) sometimes called discrete choice experiments. CBC asks respondents to choose their preferred option from sets of profiles; this has become the dominant form in marketing research.

Methodology: A researcher defines attributes (e.g., price, size, brand) and levels (e.g., $10, $15). An experimental design creates profiles by combining levels, then respondents complete a series of tasks. Statistical methods estimate part-worth utilities (preference weights) for attribute levels; common models include multinomial logit, hierarchical Bayes, and latent class approaches. Outputs include attribute importance and market simulations to predict share of preference.

Design considerations: Full factorial designs test every combination; usually, fractional factorial orthogonal designs reduce task count. CBC tasks require careful balance and realism to avoid respondent fatigue. ACA tailors questions based on previous answers.

Applications and limitations: Used for product configuration, pricing, feature prioritization, portfolio optimization, and segmentation. Limitations include cognitive burden on respondents, potential design bias, and assumptions about independence between attributes; complex interactions may be missed without larger samples or advanced models.