SOCIALLY ENHANCED ACCOUNT BENCHMARKING IN APPLICATION MANAGEMENT SERVICE (AMS)

1 SOCIALLY ENHANCED ACCOUNT BENCHMARKING IN APPLICATION MANAGEMENT SERVICE (AMS) Ying Li IBM T. J. Watson Research Center yingli@us.ibm.com Abstract This paper describes our recent work on designing and developing a socially enhanced benchmarking system to be applied to AMS (Application Management Service) domain. Benchmarking allows an AMS client or account to understand where it stands in relative to others in terms of its operational performance, and helps it set a realistic target to reach. We elaborate the entire system process in the paper which consists of the following seven modules: account data collection, cleansing, sampling, mapping and normalization; account social data mining; benchmarking pool formation and data range selection; KPI (key performance indicator) design for account performance measurement; KPI implementation, evaluation and visualization; benchmarking outcome visualization; and finally, post-benchmarking analysis. So far, we have applied this benchmarking tool to a dozen of real AMS accounts, and have received some initial yet encouraging feedback.