When consumers have heterogeneous perceptions about product quality, traditional parametric methods may not provide accurate marginal valuation estimates of a product's characteristics. A quantile regression framework can be used to estimate valuations of product characteristics when quality perceptions are not homogeneous. Semiparametric quantile regressions provide identification and quantification of heterogeneous marginal valuation effects across a conditional price distribution. Using purchase price data from a bull auction, we show that there are nonconstant marginal valuations of bull carcass and growth traits. Improved understanding of product characteristic valuations across differentiated market segments can help producers develop more cost-effective management strategies.