The Target Oriented Network Intelligence Collection (TONIC) problem is the problem of finding profiles in a social network that contain information about a given target via intelligent crawling. Such profiles are called leads. Two best-first search frameworks are proposed for solving TONIC and several heuristics are proposed for each framework. Based on the assumption that leads tend to cluster together, the first proposed framework, called the Basic TONIC Framework (BTF), limits the search for new leads only to immediate neighbors of the leads previously found. The second proposed framework, called the Extended TONIC Framework (ETF), relaxes this limitation and extend the scope of the search to a wider neighborhood, including the possibility of crawling to proles that are not leads. We analyze experimentally the pros and cons of the two frameworks on two large social networks: DBLP co-authorship graph and the popular Google+ social network. The results show that ETF can reach more leads and in some cases even faster, but BTF might be preferred in cases where the reward of finding a lead is moderate and the available budget for finding leads is large. Our results also provide insight into the maximal leads that can be found in each framework, topological characteristics that may effect the leads finding process and analysis of the TONIC problem from different perspectives.