The learning curve relates the performance of a specific task to the number of repetitions of that task and is used as a planning and control tool. Several mathematical models have been devised to describe the learning curve, each of them based on certain parameters which have to be estimated using performance data of previous repetitions. In many real-life situations, when several products are simultaneously produced by the same organizational unit, only aggregated data are available. The objective of this study is to present two models that are able to derive the learning-curve parameters of each product from such data. The models developed are for the linear and the power learning curves.