This paper empirically investigates consumer response to a traffic light advisory for environmentally sustainable seafood, which was implemented in the seafood department of a regional supermarket chain in the United States. Green meant 'best choice'; yellow meant 'proceed with caution'; red meant 'worst choice'. Using a unique product-level panel scanner data set capturing sales information for 2 treatment stores and 8 nearby control stores, we apply a difference-in-differences identification strategy to estimate the impact of color-coded labels on consumers' purchases. We find that the advisory leads to no significant difference in total seafood sales. Green sales significantly increase an average of 29% per week; yellow sales significantly decrease an average of 27% per week; red sales show no significant difference in sales. Green products on a mercury safe list had the greatest increase in sales whereas yellow products not on the mercury safe list had the largest drop in sales.