According to automatic milking system (AMS) manufacturers, a goal for milk production is 2,300 kg/robot/d or greater. The objective of this study was to investigate the association of management and housing factors with milk production/robot on free-flow cow traffic farms. We visited 36 AMS (Lely Astronaut, Lely, the Netherlands) farms in MN and WI over the summer of 2018. Producers answered a survey about general farm management practices and barn design. In addition, we collected retrospective daily data from the AMS software. We used data for the 30 d (1,080 daily averages) before the farm visit for the analysis with the MIXED procedure of SAS 9.4 (SAS Institute, Inc., Cary, NC). Backward stepwise elimination was used to remove nonsignificant factors until all remaining factors had a P < 0.05 in the final model. Farm was used as random effect. Milk production/robot was 2224.2 (±340.1) kg/d. Results are presented as least squares means (±SE). Multiple robots/pen tended to be associated with lower milk production/robot (−58.2 (±34.3) kg/d, P = 0.09). A higher number of cows/robot resulted in greater milk production/robot (35.6 (±1.25) kg/d). Using an automatic alley scraper resulted in an increase of 145.6 (±62.7) kg of milk/robot compared with manual scraping. To our knowledge, this is the first study to look at number of feeds offered in the robot in the US. There was a tendency for a positive association (P = 0.09) between number of feeds/robot and milk production/robot (69.9 (±40.8) kg/d). For each additional kg of concentrate offered in the robot, there was an increase of 66.2 (±10.5) kg of milk/robot. Successful milkings/cow/d was positively associated with milk/robot (503.1 (±15.0) kg/d). Refusals were negatively associated with milk production/robot (−24.5 (±5.1) kg/d). A higher milking speed resulted in more milk/robot (245.3 (±14.3) kg/d). Similar pattern was observed for milking time (sec) (3.3 (±0.2) kg/d). Results indicate that certain feeding practices and housing factors can influence robot performance.