During the 2020 election cycle, numerous national, state, and local organizations mounted campaigns designed to counter mis- and disinformation about election activities and to foster public trust in election processes. Local and state election offices focused intently on creating and disseminating accurate messages about when, where, and how to vote. Despite these efforts, trust in the 2020 election remained confoundingly low. This research analyzes this disconnect between messages from election offices and public trust by testing messages collected from election officials around the country using focus groups and a national panel survey experiment. We find that in focus groups, neutral messages that evoke local connections tested better than other types of messages. Further, messenger characteristics influenced whether participants trusted the messages. Using messages based on these findings, we fielded a survey experiment during the 2022 midterm election cycle, finding that the interaction between baseline trust, racial identification, and identification with the messenger moves trust.