Abstract: A method and online system for recommending gifts to a user, for a gift recipient, includes determining social relational aspects between the giver and gift recipient, determining personality traits of the giver or the recipient and calculating and then displaying to the user one or more recommended gifts for purchase. Personality, occupation and style traits are numerically rated in conjunction with product attributes of a population of potential gifts for accurate calculation of optimized gift recommendations. Gifts are rated by an online community associated with the system. The system may also learn by a feedback loop for successful and unsuccessful predictions of gift desirability, thus resulting in greater gift recommendation accuracy over time.