Quantifying merger effects and conducting other counterfactual analyses is often difficult in platform markets because direct and indirect network effects link demand within and across market sides. We develop a general empirical framework for two-sided markets that incorporates these effects within an otherwise standard empirical IO approach. The framework does not impose a particular demand specification: researchers can estimate flexible demand systems on each side, for example, using BLP-type demand on one side or both. We allow direct and indirect network effects to vary across groups. Standard onesided merger simulation emerges as a special case when network effects are absent. We provide sufficient conditions for uniqueness in the coordination game within and across market sides and for existence and uniqueness in the firms’ pricing game. These conditions are testable. When these conditions hold, the framework supports recovery of marginal costs from firms’ first-order conditions and can be used to conduct counterfactual analysis, including merger simulation.