Which department in an enterprise is concerned with a certain big data objective? The most compelling answer is to combine the eTOM framework with big data use cases.
To understand how the big data applications are implemented in practices, we need to pinpoint the applications on the business processes in an organization. For such purpose, we use the TM Forum’s Business Process Framework (eTOM) which entails a comprehensive, industry-agreed, multi-layered view of the key business processes required to run an efficient, effective and agile digital organizations.
At the conceptual level, eTOM has three major areas, reflecting major focuses within typical organizations, i.e., 1) Strategy, Infrastructure and Product (SIP) which comprises of innovation and development activities; 2) Operations (OP) which deal with production activities; and 3) Enterprise Management (EM) which includes support functions such as risk, effectiveness, knowledge & research, financial & asset, investor relationship, and human resources management.
SIP includes three groups of processes, i.e., 1) strategy & commit which covers the organization’s strategy planning & execution; 2) infrastructure lifecycle management which deals with technology design, implementation, and retirement scenario; 3) product lifecycle management which entails the design, implementation, and retirement of products and/or services. OP consists of four groups of processes, i.e., 1) operations support & readiness which includes product/service deployment from development to production, processes to monitor the running infrastructure and product/services such as their health, growth, fault, etc., and to manage orchestration or choreography of multiple processes in the production; 2) fulfilment which ensures customer orders’ execution; 3) assurance which guarantees the organization’s services or products satisfying the SLA or contract and keeping organization’s infrastructure always meets the key performance indicators; 4) billing & revenue management which handles customer transactions.
Row classifications denote the flow of the organization from receiving raw materials from the suppliers/partners to providing product/services to the customers. The top row includes customer-facing activities such as marketing and customer relation; the bottom row includes supplier facing and support activities; and the middle rows consist of the organization’s core activities, i.e., product/service development, operations, and management, and resource development, operations, and management.
Assigning the big data applications to the eTOM matrices, as shown in the Figure above, reveals which process group in the organization is mainly responsible to execute the application. By doing so, finding the relevant process owners and the corresponding stakeholders for each BOLD application is easier.
Telecom exploited big data for various applications that could be categorized into six clusters, i.e., customer acquisition and retention, risk management, fraud & cybercrime detection, resource optimization & improvement, new product/service development, and online monitoring.
Customer acquisition and retention cluster includes: 1) customer profiling / segmentation; 2) personalized telecom products; 3) product/service recommendation; 4) proactive care; 5) personalized customer service; and 6) churn analysis & prediction.
Risk Management cluster includes: 1) predicting & ensuring compliance; 2) detecting unauthorized devices; and 3) predicting bad debt.
Fraud and cybercrime detection cluster includes: 1) detecting anomaly usage (e.g., too long calls, false answers); 2) detecting anomaly of cyber activities; 3) identifying network of organized fraud.
Resource optimization & improvement cluster includes: 1) demand forecasting & capacity planning; 2) service quality improvement; 3) computing & storage optimization; 4) routing & bandwidth optimization; 5) preventive maintenance.
New product/service development cluster includes: 1) revealing insights/patterns on customers; 2) recommending features based on usage/behaviour data; and 3) developing new product/service based on usage/behaviour data.
Online monitoring cluster includes: 1) monitoring brand reputation; 2) monitoring marketing campaign effectiveness; 3) revealing product feedbacks from sentiments and discussions; 4) monitoring influential/premium customers.