system needs to convey to the user its inner logic and why a This conservative approach to recommendations had particular recommendation is suitable for them. number of effects. It led to high system transparency. Users Users like the reasoning of recommender systems to be at understood why items had been recommended and could least somewhat transparent Herlocker et al, (2000)suggest clearly see the link between the input and their output that there are many ways to for the system to convey its Because users had previously experienced and liked so many inner logic to the user: (a)an explanation(eg. "this item was of the recommended items, they perceived that the system recommended to you because you rated'x'positively"),(b) understood their tastes and were inclined to trust it more predicted ratings(e. g. "we think you'll give this item an 8.5 Amazon also provided useIs v ith detailed information about out of 10)(c)including a few familiar recommendations the item(pictures, expert reviews), as well as community by artists or writers who are very close to input items)(d) atings that further aided users in decision making. In community opinions(both reviews and numerical ratings) addition, Amazon provided sound clips for most are all effective ways to provide more information about the recommendations, allowing users to experience the item, and recommendation make their own judgments. Finally, the unit of recommendation was the album, rather than artist or song ANALYSIS OF INTERACTON STYLE OF TWo This made it easier for users to think in terms of buying the COMMENDER SYSTEMS recommendation In the preceding sections we have described some Did Amazon succeed as a recommender system? Ifthe dimensions of user interactions with recommender systems urpose of a recommender system is to allow users We have described our study findings and ofiered design explore their tastes, then Amazon had only limited success suggestions based on those findings. Below we analyze the Users did not learn many new things about their tastes. But interaction style of two very different music recommender Amazon did succeed as an e-commerce system. It systems, in order to illustrate different models of uccessfully guided users to items that they expressed an recommendation success. Our analysis should also help interest in buying. illustrate the design guidelines identified above. Results of Study 2 showed that mean liking for Amazon Recommendations by MediaUnbound: Helping (Mean=3.78; Standard Error. Il)was higher than for Users Explore Their Tastes MediaUnbound (Mean =3.49, Standard Error=. 09). Users When users were asked about the system they found the also indicated a greater willingness to buy amazon most useful, and the one they thought best understood their recommendations(20% of items) as compared to musical preferences, the unanimous choice was MediaUnbound (7% of items). However in terms of overall MediaUnbound. Also, users seemed to enjoy the system perception, MediaUnbound was rated as more useful recommendation process with MediaUnbound. They liked Mean=1.5, Standard Error=. 15)than Amazon(Mean the easy interaction with the FlashPlayer audio samples, the 1. 16: Standard Error=2).MediaUnbound was also rated as varied and humorous questions during the input process, and the system that understood users tastes best, and most likely the overall look of the site. As one user commented, "[Media to be used again. Unbound] entertains you with the process, the way you In general, our results suggest that a recommender system interact with the system. It felt like I was building a little that allows users to explore their tastes and expand their pyramid--feels like the process you'd go through yourself musical horizons might be liked and used. But it might not naturally as a human being. The rating process itself influence buying decisions to the same degree as a system seemed to inspire trust in the system and users liked the that merely reminds people of music to which they have systems recommendations. previously been exposed. This paradox is further illustrated However the profile of items recommended by in our analysis of the different styles of recommender MediaUnbound was very different than that for Amazon. music offered by Amazon and MediaUnbound Users understood why an item was recommended for only 76% of the items as compared to 92% for Amazon. U Recommendations by Amazon: Conservative had previous experience with 60% of recommendations at MediaUnbound. in contrast to 72% of Amazon Amazons Recommendations Explorer performed well when recommendations examined in terms of recommendations that users liked the Users expressed a willingness to buy only 7% of the items most, or were willing to spend the most resources on. recommended by MediaUnbound. This discrepancy Amazon recommended items that were very close to the between liking for the system and action towards its users input items. Therefore there was a high probability recommendations might be explained by the fact that a large that users had directly or indirectly experienced these items percentage of items recommended by MediaUnbound were previously. Many of the recommended items were simply new to the users. While users enjoy being introduced to new albums by the same artist named by the usersystem needs to convey to the user its inner logic and why a particular recommendation is suitable for them. Users like the reasoning of recommender systems to be at least somewhat transparent. Herlocker et al., (2000) suggest that there are many ways to for the system to convey its inner logic to the user: (a) an explanation (e.g. “this item was recommended to you because you rated ‘x’ positively”), (b) predicted ratings (e.g. “we think you’ll give this item an 8.5 out of 10”) (c) including a few familiar recommendations (by artists or writers who are very close to input items) (d) community opinions (both reviews and numerical ratings) are all effective ways to provide more information about the recommendation. ANALYSIS OF INTERACTION STYLE OF TWO RECOMMENDER SYSTEMS In the preceding sections we have described some dimensions of user interactions with recommender systems. We have described our study findings and offered design suggestions based on those findings. Below we analyze the interaction style of two very different music recommender systems, in order to illustrate different models of recommendation success. Our analysis should also help illustrate the design guidelines identified above. Results of Study 2 showed that mean liking for Amazon (Mean = 3.78; Standard Error=.11) was higher than for MediaUnbound (Mean = 3.49; Standard Error = .09). Users also indicated a greater willingness to buy Amazon recommendations (20% of items) as compared to MediaUnbound (7% of items). However in terms of overall system perception, MediaUnbound was rated as more useful (Mean = 1.5; Standard Error = .15) than Amazon (Mean = 1.16; Standard Error = .2). MediaUnbound was also rated as the system that understood users’ tastes best, and most likely to be used again. In general, our results suggest that a recommender system that allows users to explore their tastes and expand their musical horizons might be liked and used. But it might not influence buying decisions to the same degree as a system that merely reminds people of music to which they have previously been exposed. This paradox is further illustrated in our analysis of the different styles of recommending music offered by Amazon and MediaUnbound. Recommendations by Amazon: Conservative Recommendations, Trustworthy System Amazon’s Recommendations Explorer performed well when examined in terms of recommendations that users liked the most, or were willing to spend the most resources on. Amazon recommended items that were very close to the user’s input items. Therefore there was a high probability that users had directly or indirectly experienced these items previously. Many of the recommended items were simply albums by the same artist named by the user. This conservative approach to recommendations had a number of effects. It led to high system transparency. Users understood why items had been recommended and could clearly see the link between the input and their output. Because users had previously experienced and liked so many of the recommended items, they perceived that the system understood their tastes and were inclined to trust it more. Amazon also provided users with detailed information about the item (pictures, expert reviews), as well as community ratings that further aided users in decision making. In addition, Amazon provided sound clips for most recommendations, allowing users to experience the item, and make their own judgments. Finally, the unit of recommendation was the album, rather than artist or song. This made it easier for users to think in terms of buying the recommendation. Did Amazon succeed as a recommender system? If the purpose of a recommender system is to allow users to explore their tastes, then Amazon had only limited success. Users did not learn many new things about their tastes. But Amazon did succeed as an e-commerce system. It successfully guided users to items that they expressed an interest in buying. Recommendations by MediaUnbound: Helping Users Explore Their Tastes When users were asked about the system they found the most useful, and the one they thought best understood their musical preferences, the unanimous choice was MediaUnbound. Also, users seemed to enjoy the recommendation process with MediaUnbound. They liked the easy interaction with the FlashPlayer audio samples, the varied and humorous questions during the input process, and the overall look of the site. As one user commented, “[Media Unbound] entertains you with the process, the way you interact with the system. It felt like I was building a little pyramid--feels like the process you'd go through yourself naturally as a human being.” The rating process itself seemed to inspire trust in the system and users liked the system’s recommendations. However the profile of items recommended by MediaUnbound was very different than that for Amazon. Users understood why an item was recommended for only 76% of the items as compared to 92% for Amazon. Users had previous experience with 60% of recommendations at MediaUnbound, in contrast to 72% of Amazon recommendations. Users expressed a willingness to buy only 7% of the items recommended by MediaUnbound. This discrepancy between liking for the system and action towards its recommendations might be explained by the fact that a large percentage of items recommended by MediaUnbound were new to the users. While users enjoy being introduced to new