The Pandora music site attracts about 70 million listeners, who tune in to playlists Pandora creates based on their initial preferences for certain artists. The site uses a music intelligence algorithm to dissect the characteristics of favorite songs and serve up others that are similar. Pandora’s engineers constantly tweak the playlists as they experiment with variations of the experience. For example, do listeners want to hear mostly familiar songs, or do they want to discover new music? One of the biggest issues they wrestle with: How frequently should Pandora repeat the same song or artist in a playlist? The site constantly tries new variations to arrive at the optimal number of repetitions, but it turns out a lot depends on other factors, such as the time of day and where listeners are when they tune in. For example, Pandora’s data show that users welcome new music instead of the same old, same old, but when they’re at work not so much. The company continues to tweak its algorithm as it tries to answer the elusive question, “Can you have too much of a good thing?” What do you think— should music sites like Pandora focus on songs we love now, songs they think we will love, or both?