Choosing the right set of new products to offer is a key driver of profitability. New products often share some design attributes with existing products, thus, firms need to decide which attributes to keep common and which to differentiate. We propose and empirically implement a new methodology that can help managers to navigate the complex decision of where to focus differentiation, using “looks-like” prototypes that typically become available in the later stages of the product-development process. Our methodology complements early stage product-positioning methods, such as conjoint analysis and perceptual mapping. It also offers a way to estimate the impact of context dependence on choice. Finally, our methodology provides a way to test empirically whether perceptual mapping based on pairwise similarity judgments is appropriate for a product category. Using data obtained from a major wristwatch manufacturer, we are able to suggest guidelines on how to differentiate the firm’s offerings and estimate the magnitude of context dependent effects. We also find that for wristwatches, attributes that drive perceptions differ from those that drive choice. Overall, our approach can help avoid falling into the trap of focusing variety on attributes that are costly to differentiate and have little impact on choice.