Recently, both theoretical and experimental literatures have incorporated the ability of strategic players to communicate verbally prior to choosing their actions. We design an experiment to show how and why presence and type of communication matters. We use a multilateral bargaining setting, and focus on the specific strategic setup of the "divide-a-dollar" game between more than two players under different voting rules (majority versus unanimity). Relative to a situation where no communication is allowed, an unrestricted communication protocol leads to more inequality and closer-to-equilibrium outcomes under the majority rule, while it removes all inequality and leads to further-from-equilibrium outcomes under the unanimity rule. Under the majority rule, subjects use the communication tool to engage in back-room deals, while under the unanimity rule, subjects communicate publicly. Private communication channels are used by non-proposers to compete for favors from proposers while public communication channels are used to express preferences for fairness and equality. We further show that in order to reach close-to-equilibrium outcomes in the majority setting, private communication channels are a necessary component of communication protocols.