The 2010 Nonresponse Followup (NRFU) operation was the most expensive operation in the 2010 Census and is therefore a primary focus of cost-saving research. One proposal to reduce NRFU costs is through the use of crowdsourcing - information contributed by the general public via the internet - to help identify vacant housing units sooner in the Census and survey lifecycle. This paper presents the findings of a usability experiment that assessed which type of address display (satellite map versus road map, each with an adjacent address list) allowed participants to more accurately identify vacant housing units, as well as how many residential units to include on the display. Participants were randomly assigned to use either a satellite or a road map that had either 10 or 15 neighboring housing units displayed. Using a mixed methods approach, results show that participants overwhelmingly preferred the satellite map to the road map and those who used the satellite map reported less difficulty identifying vacant units in their neighborhood and were more confident in their accuracy. Participants also preferred the maps to display 10 units rather than 15, and those who saw 10 also reported greater confidence in their task-based accuracy. Measures of eye-tracking gaze patterns corroborate participants’ spontaneous and probed comments about using both the map and the address list next to the map in order to identify housing units of interest. Together, the results suggest that a 10-unit satellite map with an adjacent address list may elicit the most accurate crowdsourced data, reducing follow-up fieldwork and saving costs.