In this paper we describe our submission to the TREC2011 MicroblogTrack. Our run combines different methods namely customized scoring function, query reformulation, and query expansion. We apply query expansion from dataset with different weighting scheme. Furthermore, we do an initial experiment to incorporate timestamp of the tweet document in order to improve search performance. We found the query expansion utilizing external search result combined with re-tweet value in the customized scoring function was the most effective.