An indirect measurement involves collecting measures of one quantity (or a set of quantities) and using a statistical model to gain insights into another (or others) quantity of interest that cannot be directly observed. An example is to gain information on (unobservable) inner variables of a system (physical, chemical, or biological) by observing its response to known stimuli. In this scenario, differently from the direct-measurement one, the measurement equation (whether linear or non-linear) maps the quantities of interest to the observed ones, is not invertible, and can potentially be inconsistent. This chapter addresses the problem of inverting it, making optimal estimates of the quantities of interest, and minimising the propagation of the uncertainty. The Gauss-Markov theorem and Cramer-Rao lower bound are guidelines.