This study compared output from mobile inventory data collection vehicles to manually collected data techniques with a focus on two-way communications primarily through the submission of a sample data set to be analyzed prior to the submission of a full data set. The interim submittal and feedback to the vendors based on that resulted in a marked improvement in data quality for 5 of the 28 assets studied. After feedback, it is apparent that highway data collection vendors can accurately locate the vast majority of assets, with the primary exception being those that are occluded by vehicles or surrounding landscaping, such as those assets in the median. Along with the locations of assets, vendors showed promise at collecting many of the feature descriptions such as asset type and condition. Many of the elements (location, type, etc.) for a particular asset type that created collection difficulty were only problematic for a particular vendor, which suggests that further improvements may be achieved through additional communication and more explicit definitions and examples by the contracting agency. Using measurement tolerances, the research team determined that measurements of height, grade, and azimuth were generally accurately obtained; however, measurements parallel to the direction of travel, such as offset and width, posed problems with accurate measurements. Last, the accuracy of data location was acceptable for finding assets in the field; however, many specific point features such as drop inlets or attenuators were not geo-located, but instead located from the vehicles position in the roadway. This appears to be standard practice for some vendors; therefore, if a specific location outside the roadway is desired, it should be clearly indicated in the instructions. Although mobile data collection is sufficient for most efforts, there is still room for improvement if more detailed location is necessary.