In a field experiment involving online workplace lunch orders, this study examines the impact of numeric and traffic light calorie labels on calorie intake. Employees of a large corporation ordered lunches through a website of the authors’ design, on which they were presented menus with numeric calorie labels, traffic light labels, or both together, and the authors compared the calorie content of the ordered lunches with that of diners randomized to receive no calorie information. Each label type reduced lunch calories by approximately 10%. Nutrition knowledge was not improved by any menu format. Traffic light labels achieved meaningful reductions in calories ordered even in the absence of numeric information, and the authors found no apparent benefit or detriment of combining label types. These findings suggest that consumers may benefit most from help in identifying relatively healthier choices but rely little on information about the exact caloric content of items.