Service Identification is one of the most important phases in serviceoriented development methodologies. Although several service identification methods tried to identify services automatically or semi-automatically, various aspects of business domain are not taken into account simultaneously. To overcome this issue, three strategies from three different aspects of business domain are combined for semi-automated identification of services in this article. At first, the tasks interconnections within the business processes are considered. Then, based on the common supporting requirements, another tasks dependency has been determined and finally, regarding the significant impact of data in business domain, the last tasks relations are specified. To combine these three strategies, task-task matrices are used as a common language and eventually services are identified by clustering the final task-task matrix.