Many purchase and consumption decisions require that consumers estimate the quantity of a product that they have in inventory. Despite the frequency with which these estimates are made, the authors do not know if they are accurate, whether they influence consumption, or whether they are systematically biased by the quantity and salience of the products in inventory. To address these questions, the authors develop a quantity-salience model of inventory estimation and test it in six lab and field studies. They find that inventory estimations consistently exhibit an asymmetric contraction bias, where consumers slightly overestimate low levels of inventory and strongly underestimate high levels of inventory. They find that this bias is highest when products are not perceptually salient in the pantry, and that consumers with a tendency to overestimate inventory level consume stockpiled products at a faster rate. They also find that the salience of a product in inventory increases the likelihood of recalling it, whereas its quantity only influences its estimated numerosity (through a process of anchoring and adjustment) once the product itself has been recalled. Finally, they show that these results can account for known biases and seemingly inconclusive findings in other estimation tasks.