In today’s connected world it is possible and very com-
mon to interact with unknown people, whose reliabil-
ity is unknown. Trust Metrics are a recently proposed
technique for answering questions such as “Should I
trust this user?”. However, most of the current re-
search assumes that every user has a global quality
score and that the goal of the technique is just to pre-
dict this correct value. We show, on data from a real
and large user community, Epinions.com, that such an
assumption is not realistic because there is a significant
portion of what we call controversial users, users who
are trusted and distrusted by many. A global agree-
ment about the trustworthiness value of these users
cannot exist. We argue, using computational experi-
ments, that the existence of controversial users (a nor-
mal phenomena in societies) demands Local Trust Met-
rics, techniques able to predict the trustworthiness of
an user in a personalized way, depending on the very
personal view of the judging user.