Feed readers have emerged as one of the salient applications that characterize Web 2.0. Lately, some of the available readers introduced social features, analogously to other Web 2.0 applications, such as recommendations and tagging. Yet, most of the readers lack collaborative features, such as the ability to share feeds in a community or divide the reading task among community members. In this paper we describe CoffeeReader, a web-based feed reader, which combines social and collaborative features, and is deployed in a small community within our company. CoffeeReader provides awareness of other users' feed lists and read status; it enables information sharing such as tags and recommendations; and aims to support coordination of filtering through feeds to locate important items. We compare these group collaboration features of CoffeeReader with emerging features in publicly available feed readers; present the outcomes of using CoffeeReader within our community; and discuss our findings and their implications on making feed readers more collaborative.