This paper considers the capacity rationing problem for a make-to-order manufacturer facing random demand from both a brick-and-mortar retailer and his own online store. As the marginal profit from the online store is higher than that from the retailer, the manufacturer is willing to set higher priority to the orders from the online store in case of capacity shortage. Assuming a compound demand process at each channel, we formulate the manufacturer's problem and identify the optimal rationing level of capacity below which the orders from the retailer will be rejected. The effectiveness of capacity rationing is investigated with numeric simulations under a wide array of conditions characterized by variations in factors including demand substitution rate between channels, prices in the system, and demand rates at the channels. The results show that the rationing policy outperforms the first-come-firstserve policy in most cases, especially when the available capacity is very tight. Thus capacity rationing could serve as an efficient management tool for capacitated manufacturers in multi-channel supply chains.