Application of large data in telecom industry

According to a 2013 survey by the Telecoms and media market research firm Informa Telecoms & Media, about 48% of the world's 120 operators are implementing large data operations. The research company said that the cost of large data operations accounted for an average of 10% of the operator's total it budget and would rise to around 23% in the next five years, becoming a strategic advantage for operators. Visible, from flow management into large data operations has become a trend.

Telecom operators have many years of data accumulation, with such as financial revenue, business development volume and other structured data, but also involves pictures, text, audio, video and other unstructured data. From a data source, telecom operators ' data comes from all the businesses involved in mobile voice, fixed-line telephone access and wireless Internet access, as well as public, enterprise and home customers, as well as collecting contact information from all types of channels, such as physical channels, electronic channels and direct marketing channels. Overall, the development of telecom operators ' large data is still in the exploratory stage.

General situation of application of large data in telecom industry

At present, domestic operators use large data mainly five aspects: (1) network management and optimization, including infrastructure construction optimization and network operation management and Optimization, (2) market and precision marketing, including customer portrait, relationship chain research, precision marketing, real-time marketing and personalized recommendations; (3) Customer relationship management, Including Customer service center optimization and customer lifecycle management; (4) Enterprise operation Management, including business Operation Monitoring and Operation analysis, (5) The data commercialization index is commercialized and profitable separately.

First, network management and optimization. This direction includes the optimization of infrastructure construction and network operation management and optimization.

(1) Optimization of infrastructure construction. such as using large data to achieve base station and hotspot location and resource allocation. Operators can analyze the distribution of the user's traffic in the time period and the position characteristics in the order and the signaling. Design 4G base Station and WLAN hotspot for high flow area of 2G and 3G, meanwhile, operators can evaluate the efficiency and cost of existing base station by establishing evaluation model, and find out the waste of resources in base station construction, In some areas, in order to complete the base station construction targets will build the base station in interpersonal rare to place.

(2) Network operation management and optimization. In the network operation level, the operator can analyze the traffic flow of the network through the large data, change the trend of the flow direction, adjust the resource allocation in time, and also can analyze the log, carry on the whole network optimization, improve the network quality and network utilization continuously.

Using large data technology to collect and process network signaling data in real time, monitor network condition, identify value community and business Hotspot District, and guide Network optimization more accurately, realize network, application and user's smart finger matching. Because of the different user groups, the contribution of different communities to operators is also different. The operator can analyze the data of the cell in multi-dimensional data, through to the Community VIP user distribution, the income distribution, and the correlation distribution model obtains the different community the value, again and the network quality analysis unifies, both together, may discover a certain community value is high, but the network coverage needs to ascend further, And then set the priority of network optimization, improve investment efficiency.

Deutsche Telecom set up a forecast of the city inside the regional wireless resource occupancy model, according to the forecast results, flexible deployment of wireless resources in advance, such as in the daytime to the CBD area to allocate more wireless resources, in the evening, the bar area to allocate more wireless resources, so that wireless network operating efficiency and higher utilization.

France Telecom through the analysis found that a segment of the network, the drop rate is persistently too high, the use of large data to diagnose the reason for the interruption of the call is caused by the network overload, and based on the results of the analysis to optimize the network layout, to provide customers with a better experience, access to more customers and

Second, market and precision marketing. This direction includes customer portrait, relationship chain research, precision marketing, real-time marketing, and personalized referrals.

(1) Customer portrait. Operators can be based on customer terminal information, location information, call behavior, mobile phone online behavior and other rich data, for each customer to hit demographic characteristics, consumer behavior, Internet behavior and interest hobby label, and with the help of data mining technology (such as classification, clustering, RFM, etc.) for customer clustering, Improve the customer's 360-degree portrait, to help operators understand customer behavior preferences and demand characteristics.

(2) Relationship chain research. The operator can analyze the communication circle by analyzing the customer address book, the call behavior, the network social line and the customer data. In particular, the use of a variety of contact records to form social networks to enrich the user's insights, and further use of map mining methods to discover a variety of circles, identify the key people in the circle, as well as identify family and enterprise customers, or analyze social circles to find marketing opportunities. As in a homogeneous circle of behavior, if this circle is mostly for high flow users, and in this circle to find users of the network, we can speculate that the user is also a high flow of the situation, it can be through marketing activities to the network of high flow of users to guide their networks, the promotion of its 4G package, improve marketing conversion rate. In short, we can use social circles to improve marketing efficiency, improve services, and increase the impact of products at low cost.

(3) Precision marketing and real-time marketing. Operators in the customer portrait based on the customer's in-depth understanding of the characteristics of the establishment of customer and business, tariff packages, terminal type, in the use of accurate network matching, and in the push channel, push time, push the way to meet customer needs, to achieve precision marketing. If we can use large data to analyze the user's terminal preferences and consumption capabilities, predict the user's time to change the machine, especially the contract machine expiration time, and capture the user's recent feature events, so as to predict the real needs of users to purchase terminals, through the SMS, call center, Business Hall and other channels to push the relevant marketing information to the user's hands.

(4) Personalized recommendation. Using customer portrait information, customer terminal information, customer behavior habits and preferences, operators can provide customers with customized services, optimize products, flow packages and pricing mechanisms to achieve personalized marketing and services to enhance customer experience and awareness, or in the application of the store personalized recommendations, in the electric business platform to achieve personalized recommendations, Recommend interesting friends on social networks.

Third, customer relationship management. This includes customer Service center optimization and customer lifecycle management.

(1) Customer Service Center optimization. Customer Service Center is the operator and customer contact more frequent channel, so the customer service Center has a large number of customer call behavior and demand data. We can use large data technology to deeply analyze customer service hotline inbound customer behavior characteristics, select a path, wait for a long time, and related customer history contact information, customer package consumption, customer demographics, client-type data, set up a customer hotline intelligent path model to predict the next customer inbound demand, The risk of complaint and the corresponding path and node, this will shorten customer service incoming processing time, identify the risk of complaints, help improve customer satisfaction; In addition, can also through semantic analysis, the problem of Customer Service hotline classification, identify hot issues and customer sentiment, for the occurrence of large and serious problems, To timely alert the relevant departments to optimize.

(2) Customer care and customer lifecycle management. Customer lifecycle management includes five phases of new customer acquisition, customer growth, customer maturity, customer decline, and customer departure. In the customer acquisition phase, we can excavate and discover the higher customer through the algorithm, in the customer growth stage, through the association rule algorithm and so on cross-selling, promotes the customer per capita consumption amount, in the customer maturity period, may carry on the customer grouping (RFM, clustering and so on) and carries on the accurate recommendation, At the same time to different customers real-time loyalty program, in the customer recession, need to carry out the loss of early warning, the early detection of high-risk customers, and to make corresponding customer care; In the customer departure stage, we can use large data mining high potential return customers. Domestic and foreign operators in customer lifecycle management applications are more cases. SK Telecom, such as the new establishment of a company SK Planet, specialized in dealing with large data-related business, through the analysis of user behavior, before the user made the decision to leave, the introduction of the business in line with user interest to prevent the loss of users, and t through integrated data analysis of the causes of customer churn Halve the wastage rate in a quarter.

Four, enterprise Operation management. Can be divided into business operations monitoring and management analysis.

(1) Business operations monitoring can be based on large data analysis from the network, business, user and business volume, business quality, terminal and other dimensions for operators to monitor the pipeline and customer operations. Construct the flexible and customizable index module, construct the index system such as QOE/KQI/KPI, and the Intelligent Monitoring system, and control the operation and the reason of the movement from macroscopic to microcosmic.

(2) Business analysis and market monitoring. We can through the data analysis of business and market operation status of summary and analysis, mainly divided into operating daily, weekly, monthly, quarterly and thematic analysis. In the past, these reports were written by analysts. In the large data age, these business reports and thematic analysis reports can be automated to generate Web pages or app forms, through machines. The data source is the business and user data within the enterprise, as well as the external social network data, technology and market data collected by the large data means. The analyst is transformed into a reporting product manager, developing a reporting framework, analysis and statistical dimensions, and the rest of the work is done by the machine.

On the other side, data commercialization. Data commercialization refers to the commercialization of the large data assets owned by the enterprise to obtain benefits. The data commercialization of operators at home and abroad is at the exploratory stage, but relatively speaking, foreign operators are developing faster in this respect.

(1) Provide marketing insights and accurate advertising.

Marketing Insight: Verizon, the US telecom operator, set up the Precision Marketing department precision Marketing Division. The department provides precision marketing insights (Precision harsh Insights), which provides business data Analysis Services. In the United States, baseball and basketball games are the most popular marketing occasions, after the Super Bowl and NBA games, Verizon for the audience's source of accurate data analysis, the team to understand the audience's preference for sponsors; The US telecom operator Sprint uses large data to provide consumer and market insights to industry customers, including demographic, behavioral, and seasonal analysis.

Precision Advertising: Verizon's precision marketing department offers accurate advertising services based on marketing insights; AT&T provides alert services, and when users are close to the business, they are likely to receive a large discount of the electronic coupons offered by the merchant.

(2) based on large data monitoring and decision support services.

Passenger flow and location: In October 2012, Telefónica set up a dynamic Insight department dynamicinsights to carry out large data business, to provide customers with data analysis packaging services. The Department, in collaboration with market research institute GfK, has launched its first product name, smart tournaments, in the UK and Brazil. Smart footprint based on fully anonymous and aggregated mobile network data, retailers are helped to analyze customer sources and the flow of people from various shops and booths, as well as consumer characteristics and consumption capabilities, and provide insight into the customer flow analysis and retail location services.

Public service: France Telecom, France's largest operator, and its communications solutions Department, Orange Business Services, has built IT systems for many of France's public services, such as the construction of a French highway data monitoring project, which produces millions of records a day, The analysis of these records can provide accurate and timely information for the vehicles travelling on the expressway, and effectively improve the road patency rate.

Overall, the telecommunications industry's big data is still in the exploratory phase, in the next few years, whether the internal large data applications or external large data commercialization has a lot of room for growth. But the biggest obstacle of large data of telecommunication industry is the data island effect is serious, because of the regional operation of the domestic operator, the data of Telecom enterprise is stored in the branch of each region separately, and even the data of different business of branch office may not get through. The internet and large data have no boundaries. NTT Docomo, Japan's largest mobile communications operator, began planning for large data use before 2010, and NTT Docomo has a big advantage over domestic operators to collect and integrate data across the country, so the NTT Docomo can easily get the country's system data. DoCoMo not only focus on collecting the user's own age, sex, address and other information, but also the production of fine forms, require users to fill out more detailed information business. For domestic telecom operators, to really use large data, data unification and integration is the most important step. We have seen China Mobile has begun to prepare for this work, and believe that in the next few years, under the pressure of internet companies ' competition, the big data of Chinese telecom industry will develop faster and the change will be more thorough.