| The development of information technology and internet presents enormous challenges to traditional telecommunication.On one hand,development of mobile internet brings explosive growth in data usage,rendering major mobile service providers encountering the risk of being 'pipelined'.On the other hand,conventional services,such as voice calls and text messages,are being corroded by relevant APPs of similar functions,therefore abating the profits of conventional telecommunication services.As a result,seeking new profit growth and data usage quality is of vital importance.Mobile terminals,data flow and size of client groups have exerted great impacts on traditional telecommunication services.In the era of mobile internet,data-related service will become customers' core needs,thus represents core values of service providers.China Mobile Communications Corporation(CMCC)has established big data platforms in every provinces,hoping to utilize data management platforms to efficiently exploit on client information,optimize service functions,and provide services to our clients with high precision and therefore lowering the risks of new services.Currently,marketing of CMCC is hindered by subjective string filtering,targeting customers based on fixed presets,and marketing inaccuracy,recording a low marketing success rate at 1.+%.Observation and research on preferences of end-users can help CMCC target suitable clients with high precision,which in turn help improve marketing success rate and mobile data usage purchase.MIGU's current strategy focuses on enterprise development,facilitated by development of its provincial branches.This article answers to the calls of CMCC by analyzing behavioral preferences of mobile end-users,which promotes more efficient marketing of five core entertainment services of MIGU(MIGU music,MIGU video,MIGU readin,MIGU games and MIGU comics).Successful marketing of these services not only will boost mobile data usage of CMCC,but also lay the foundation of enterprise development.This work draws strengths from predecessors'accumulation and constructed a model that analyzes preferences based on data mining and situational recommendation theory.Our model has been applied in marketing and proved to be effective.This work established two versions of the model based on different data mining techiniques:a model that analyzes historical habitual preference of end-users(descriptive mining model)and a model that predicts potential preferences(predictive mining model).By applying the B/O/M triple-domain integration technology of the CMCC,data on social and telecommunicative attributes,and cellular behavioral characteristics were analyzed by data mining tools of IBM SPSS Modeler 14.1.Data were processed in a dimension-reductive and regressional-analytical fashion to devise a scoring scheme on preferences in reading,music,video,games,and comics.Scores obtained are used to predict possible APPs favored by clients.Furthermore,mobile signaling technology helps to study temporal and special preferences of clients,rendering situational recommendation of APPs possible.This article,in the last paragraph,presented four successful outcomes in marketing based on our model.Our model has practically proven to be effective to certain degrees,which arms commercial entities with a tool to precisely target clients and improve click and click-transformation rates.Contribution of this work lies centrally on its research on preferences of mobile end-users which has been consistently ignored by domestic scholars.This articles also introduces B/O/M triple-domain integration technology of the CMCC and its related techniques in data acquirement.It constructed mining models focusing on content preferences of users and proposed a methodology of recommending APPs to different categories of clients.Lastly,the paper integrates otherwise easily ignored factors,such as time and location of mobile web access,into service marketing,which is one step closer to fully customized and situational marketing. |
PDF Full Text Request