Managing the evolution of a single product working with a small number of teams is somewhat straightforward. Working from a single backlog, once the teams have established a predictable velocity, the product roadmap becomes relatively easy to visualize, whether by timeframe or feature set. As we increase the complexity of the product, things become harder. Different teams require different backlogs. Different products require work from different teams. Before you know it, there are lots of independent moving parts, with the risk that the coordination cost becomes higher and efficiency falls. In this talk, we consider some ground rules for visualizing work across multiple teams and discuss how dependencies are coordinated across different teams and product lines.