Mostly e-business and e-applications rely on the Service Oriented Computing paradigm and its most popular implementation, namely Web Services. When properly implemented and described, Web Services can be dynamically discovered and reused using Internet technologies, pushing interoperability to unprecedented levels. However, poorly described Web Services are rather difficult to be discovered, understood, and reused. This paper presents heuristics for automatically detecting common pitfalls that should be avoided when creating Web Service descriptions. Experimental results with ca. 400 real-world Web Services, empirically show the feasibility of the proposed heuristics.