We often remember in the company of others. In particular, we routinely collaborate with friends, family, or colleagues to remember shared experiences. But surprisingly, in the experimental collaborative recall paradigm, collaborative groups remember less than their potential, an effect termed "collaborative inhibition". Rajaram and Pereira-Pasarin (2010) argued that the effects of collaboration on recall are determined by "pre-collaborative" factors. We studied the role of 2 pre-collaborative factors--shared encoding and group relationship--in determining the costs and benefits of collaborative recall. In Experiment 1, we compared groups of strangers who encoded alone versus together, before collaborating to recall. In Experiment 2, we compared groups of friends who encoded alone versus together, before collaborating to recall. We found that shared encoding abolished collaborative inhibition in both Experiments 1 and 2. But prior relationship did not influence collaborative inhibition over and above the effects of shared encoding. Regardless of encoding condition, collaborative group recall contained fewer intrusions than nominal group recall, and these benefits continued in subsequent individual recall. Our findings demonstrate that pre-collaborative factors--specifically shared encoding--have flow-on benefits for group and individual recall amount, but not recall accuracy. We discuss these findings in terms of self- and cross-cuing in collaborative recall. (Contains 2 tables and 5 footnotes.)