Research On The Design Of Marketing Decision Support System Of G Company

Company G as a global company,is more focus on using the traditional method and part of data mining work to manage,analysis and forecast the marketing data in the marketing decision-making phase.So they couldn't have the precise data,that caused the failure marketing strategies.It's very necessary to upgrade the marketing decision support system(DD-MDSS)which based on data-driven.Based on the analysis of company G's marketing decision support system,this paper proposed solutions to improve the marketing decision support system according to the current development of the theoretical community and the use of big data mining technology.The main improvements are from these four parts:(1)Market analysis.In this module,we have made improvements based primarily on the SWOT model.This is because the SWOT model has been proposed for a long time and has been limited by the times.Today's companies may place more emphasis on the organizational process.SWOT doesn't not consider the enthusiasm of enterprises to change the situations.This paper upgrades the existing SWOT,adding several dimensions to the previous opportunities,threats,disadvantages and advantages.It is divided into: opportunity cost,missed opportunity,opportunity overlap and loss of opportunity.These could help us to better position our target market.(2)Customer segmentation.This action could make marketing more accurate,thereby reducing business costs and improving user satisfaction.This paper uses K-means clustering algorithm based on the LRFMC model to segment customers and develop customized marketing programs for different customer groups.Firstly,based on the RFM model,this paper proposes to add 2 dimensions to the product features of G company,L(membership level)and C(single order quantity),and then provide the score,and then K-means clustering algorithm divides the customer points.It is important to maintain customers,important development of customers,important retention of customers,ordinary value customers and low-value customers,and then decompose and classify application programs for marketing activities.(3)Sales forecast.This is mainly based on a certain method and model of historical data analysis,and then estimate the future sales,which can guide enterprises in the direction of production and operation planning.In this paper,based on the ARIMA model,we use the predicted data as the input parameters of the BP neural network to obtain the predicted sales data.This proves that the optimization of the BP neural network based on the ARIMA model can better enhance the prediction.Accuracy.(4)Marketing performance assessment.This is mainly aimed at regular evaluation and assessment of sales results.The purpose is to find successful models based on marketing plans and actual results for replication.Auxiliary marketing activities make decisions based on marketing revenue and efficiency.This article is based on the existing foundation of G company,the unified caliber defined in the assessment,and the SMART assessment principle.It is the horizontal comparison of the marketing performance of each channel and the efficiency of marketing activities.