We quantify the benefits of better firm-to-firm matching in an aggregate diffusion model where individuals acquire profitable knowledge from others. We estimate the model using a small-scale program in Kenya that creates new opportunities for firm managers to interact. Aggregate gains from the program depend on the relative importance of meeting a high-knowledge firm versus learning within that meeting. We show how moments from the intervention identify these parameters, demonstrating how easily estimated moments beyond the average treatment effect inform at-scale gains. We provide sufficient conditions extending the identification procedure to a broader class of experiments and aggregate diffusion models.