If a user realized that the "cost" of viewing the image was free, wouldn't the user disregard the cost on future instances of the experiment?
Did the experiment limit a user's participation to one time? Otherwise, the results might be skewed and the benefits exaggerated.