XML element retrieval aims at finding the best elements satisfying a user's information need. Elements spanning only a few words, like titles or italicized phrases, are not in themselves useful results, but they can support the relevance of their enclosing elements. For example, if a section's title contains the key words from the user's query, the title itself is unlikely to be a useful result, but the section is very likely to be useful. This paper provides an overview of methods for exploiting small elements for better retrieval results, highlighting their respective advantages and disadvantages. Using the INEX testbed, we show that small elements can indeed provide useful retrieval hints, and we evaluate the trade-offs.