Recent research finds that many analyst recommendation revisions take place shortly after earnings announcements.Altinkilic and Hansen (2009) attribute the clustering of recommendations to analysts strategically piggybacking on earnings information to improve the perceived performance of their recommendations. This study proposes an alternative view: I find that analysts issue recommendations when they face greater demand from investors, when the relative supply of information available on earnings announcements is higher and when they detect mispricings. These results are consistent with analysts striving to meet the demands of investors by providing useful recommendations after earnings announcements.