The only problem with that is that they'd have no idea the data would be bad. Even if they did, there's nothing stopping the marketers from taking the bad data as proof one or the other campaign was actually working.