Supply chain planning in global production networks is a very difficult task due to the high diversity of products and machines and the immense number of possible configurations. An important task in this area is capability matchmaking: finding machines that are capable to produce specific parts. Today, this is done by experienced engineers who have the necessary knowledge to assess the feasibility and efficiency of solutions. However, they have limited knowledge of all available machines in the network and are strongly influenced by their personal familiarity with specific products, machines and locations. We present a decision support system that expands the search space, thereby facilitating the process and ultimately improving the implemented solution. To this end, we built an implicit recommender system that predicts possible machine types for parts based on their historical production patterns. This approach constitutes an effective and lightweight option for capability matchmaking in brown-field settings.