Conjoint Market Simulator

First choice (maximum utility) assigns each respondent entirely to their single highest-utility product, so a product's share is the fraction of respondents who rank it best; it is intuitive and free of the independence-of-irrelevant-alternatives problem but volatile and unrealistic, since people do not always buy only their favorite. Share of preference instead uses the logit equation to give each product a probability proportional to its exponentiated utility, producing smoother, less volatile shares that reflect divided patronage. The trade-off is that share of preference can overstate similar products' combined share, which is exactly the gap that randomized first choice was designed to close.

Randomized first choice (RFC) blends the realism of first choice with the smoothness of share of preference by adding random error to the utilities, including at the attribute level, and averaging many first-choice simulations. The attribute-level error is the crucial ingredient: it makes near-identical products correctly cannibalize each other rather than draw share proportionally from all alternatives, which fixes the independence-of-irrelevant-alternatives distortion that plagues plain logit. According to Sawtooth's simulator documentation, RFC has been shown to predict holdout choice shares better than the other Sawtooth rules, which is why it is the recommended default for most simulations.

Not directly. A conjoint market simulator outputs shares of preference, the relative probability that respondents prefer each product within the defined competitive set, not absolute sales volumes. Real sales depend on awareness, distribution, the size and definition of the market, and factors the conjoint study never measured. Practitioners therefore use the simulator for relative comparisons across scenarios and, when an absolute forecast is needed, calibrate the simulated shares against known market data, adjust for the 'none' option and reach, and treat the result as a directional estimate. Holdout validation and external calibration are what bridge the gap between preference share and a usable sales forecast.