Hedonic models are commonly used to quantify the value of characteristics implicit in a product’s price. However, when products are heterogenous across quality levels, using traditional parametric methods for estimating characteristic values may provide poor inferences about quality effects. We propose using a quantile regression framework for estimating the value of characteristics in quality-differentiated products. Semi-parametric quantile regressions allow the data to flexibly identify and estimate quality effects across a conditional price distribution. Using purchase price data from a bull auction, we show complementary non-linear relationships exist between quality and bull carcass and growth traits. Improved precision in understanding consumer valuation of product characteristics across quality market segments can help producers tailor products for each segment.