In the Web 2.0 era, users share and discover interesting content via a network of relationships created in various social networking or content sharing sites. They can become for example contacts, followers or friends and express their appreciation of specific content uploaded by their peers by faving, retweeting or liking them depending on whether they are in Flickr, Twitter or Facebook respectively. Then they can discover additional content of interest through the lists of favorites of their contacts and so on. This faving (or favoring) functionality becomes thus a central part of content sharing communities for two purposes: (a) it helps the propagation of content amongst users and (b) it stimulates users' participation and activity. In this paper, we make a first step to understand users' faving behavior in content sharing communities in terms of reciprocity using publicly available datasets from Flickr and Twitter. Do users favor content only when they really appreciate it or they often feel the need to reciprocate when their content is appreciated by one of their contacts or even by a stranger? Do people take advantage of this process to gain popularity? What is the impact of the design, the social software, of a specific community and the type of content shared? These are some of the questions that our first results help to answer.