Using Big Data Analytics to Develop Marketing Intelligence Systems for Commercial Banks in Egypt

Nowadays, Big Data (BD) Analytics is receiving great attention in banking industry, considering the worthy data that have been stored for several decades, to reach the main targets of marketing by increasing the bank’s efficiency of studying their clients, knowing their feedback, in addition to promoting active and passive security systems. This study focuses on utilizing BD analytics to develop marketing intelligence systems. It aims to explore the big data as a valuable resource for Egyptian commercial banks, to improve the customer experience, customer segmentation and profiling, selling products based on profiling, and describing customer behavior. In order to develop the proposed system, data were collected from several banks of transaction performed in 2016, including a report on customer satisfaction, a procedure of analyzing customer satisfaction data, consisting of about 39,000 records of transactions for customers and a collection of about 4,000 records of transaction data for cardholders. These data were analyzed using Apache Hadoop to perform many tasks such as profiling the bank's clients to groups, customer segmentation based on client’s history, interest and habits, predicting customer behavior based on profiling, designing a new marketing strategy, and presenting the right offers to the bank's clients as individuals or as groups. It was concluded that BD analytics were very beneficial for achieving Marketing Intelligence in Banks.


Introduction
Nowadays Big data (BD) has got an extraordinary attention from the academic world and the IT business. In the automated world, data are created and collected with hight speed. At present, more than 2 billion people worldwide are connected to the Web yet more than 5 billion people own cell phones. By 2020, 50 billion gadgets are anticipated to stand associated with the Web. by then, expected data production will be 44 times bigger than in 2009 [1]. As data are exchanged and shared at light speed on optical fiber and wireless networks, both volume of information and the speed concerning market development increase. However, the expedient increment rate with respect to such large data produces various difficulties, such as the rapid growth of data, exchange speed, differing data, and security.

Big Data Definition
Big Data (BD) may be defined as "the amount of dynamic, large and scattered information generated by individuals, tools and machines" [2]. It needs innovative and extensible technology to collect, host and analyze a large amount of gathered information to capture insights into real-time practices that relate to consumers, risk, profit, achievement, production, and management. It can also be identified as "a collection of information from both traditional and digital sources within and outside the association that is a source of discovery and continuous analysis" [3]. As for the characteristics of BD, certain elements are common across the different definitions, namely the 5Vs Volume, Velocity, Variety, Veracity and Value.  Volume refers to the great preparation of information that is generated every second. This makes growing information sets huge for storage and analysis using traditional database technology for Velocity,it refers to the speed with which modern information is established and the speed at which information moves as for Variety,it refers to non-similar forms of information that we can use at this moment. With BD technology, we can now benefit from similar forms of information, including conversations, social networks, images, sensor data and audio and combined with more organized traditional structured data. concerning Veracity,it refers to confusion or reliability of information. With a lot of great information, quality, and accuracy, you can control. Value refers the ability to change the data to an amount. It is important that the institutions makeS a claim to any endeavor to collect and benefit from the great information. [5].

Applications of BD in Banking Sector
Following are some of the recent application Of BD In Banking Sector. Fraud Detection: Banks use analytics to detect patterns and determine their type if they are normal or abnormal. By analyzing large data, banks will have a fraud detection tool that will detect patterns, which will raise an early alarm making the analysis system suggest taking immediate actions that in turn will stop fraud before it happens. Real-time Risk Analysis : A real-time analysis of transactions will support banks against capital, credit risk and financial risks when they skip their initial beginnings and alert banks to take suitable action. BD could be an appropriate tool because of its high frequency in terms of business data and sheer volume. Client Satisfaction and Earning Loyalty : BD's ability to make the customer's DNA through the history of dealing, the rate of debt, and repayment habits, all tell a story and can facilitate the bank to distinguish customers which makes us understand the behavior of customers. Through this understanding, we can classify customers and enable large data to aggregate different clients into segments will create based on customer behavior. Personalized Marketing: Customer segmentation will be more utilized produce and create new schemes, aimed immediately to the particular needs of their customers. By learning behaviors return from BD sources and analyzing past and current cost and transactions, a bank will get a clear understanding of the way to get the best response frequency from their customers.The access to information of selection associated price creates a chance of providing customized goods to customers can cater to an This will raise satisfaction levels for customers , will improve profit for banks and this will offers banks the flexibility to make additional relationships with clients. Create New Product: By learning the direction of client needs and satisfaction through analytics you can create products according to the wants of clients.
Reduce costs: BD is best seen as a price cutting live within the banking sector. It's data management system that may scale back the necessity for storage making easy for data and lower the computational needs for the bank. The convoluted and learning operations of massive information technologies will be accessed and managed by average ability employees to avoid hiring specialist operators. The utilization of machine learning suggests that the system once set the suitable parameters, will learn for itself to relate and analyze information to create it additional helpful to the bank. Increase Bank Efficiency: The use of resources to attain outcome are often managed and even reduced thanks to BD. the major objective is to scale back the time employed in accessing, decoding and manipulating data and BD would cut back hiding and increase turnout. The banking systems would be economical and secure with abundant less time wasted throughout background checks, lines in banking halls, extended application procedure and redundant authentication transaction [6].

Problem Definition
The main problem in this research lies in the fact that Commercial banks are still using traditional methods and do not make use of BD. Exploratory studies showed that banks in Egypt are not using new technologies to develop marketing strategies. This makes them lose many competitive advantages in the marketing area. As a result, banks need to shift the paradigm for designing, developing, deploying, and maintaining BD solutions supported by a wave of emerging technologies that provide the flexibility and scalability required to support the shift.

Objective of The Study
The main objective of the study is to apply BD analytics on the valuable resources for Egyptian commercial banks in order to develop Marketing Intelligence Systems in banks b ased on cu st om er p r of i l i n g.
This should eventually lead to improving customer service and adding new products and services based on customer profiling.

Research Methodology
The study involves interviews with 5 CIOs and 5 IT managers from Egyptian commercial banks to find out if computer technologies affect marketing decisions in banking and how they affect marketing strategies of the Bank. It also makes use of a survey for customers to find out their satisfaction levels with the products' offers of the bank and their payment preferences (cash -credit card).
In addition, data we extracted from customers transactions.All of the previously -mentioned data are analyed.

Literature Review
Following are some of the most recent studies that dealt with BD in banking sector. A study explored ways on how to improve the efficiency of the banking system with BD.It covered customer satisfaction analysis, by identify the main customer of the bank with the analysis link, improve customer feedback, detected when the customer is about to leave and improve security enhancement. It found out that if the banking industry can effectively implement all analyzes and algorithms (feeling, aggregation, correlation, survival, decision tree, analysis and data file), it will increase the efficiency of the banking system [7]. Another study in the financial services industry dealt with a particular focus on the banking industry, highlighting many of the business problems that have been resolved through BD analysis. It also covered some of the latest technologies, which combined with large data analytics, have significantly changed the quality and productivity of the banking industry. The disciplines of the stock market and insurance have been explored, with predictive and directional analyzes playing a key role. The paper concluded with some challenges, which hamper the full implementation of BD analyzes in the financial services industry to enhance business value [8].
Third study made use of BD effectively to gain tremendous value in customer retention, targeted marketing, fraud reduction, internal operation, and many other areas. Customers ' data collected across diverse systems enable advanced sectors to drive new customer growth and improve marketing ROI. Excessive analytics facilitate the opportunities for mutual forecasting and growing sales to increase portfolio share and long-term customer value.It was found out that BD also plays an important role in improving operational efficiency. Interinstitutional data supports regulatory requirements, new requirements and early detection of fraud. [9]. A fourth study explained that BD analysis is being implemented across many areas of Banking sector, and helps them to provide higher services to their clients, all Internal and external, besides improving them. This study explains the impact and advantages of BD analytics on banking sector through analyzing transactional and sentiment analysis for the Banking Sector. There are other ways that banks and various monetary institutions can obtain the knowledge associated with the client in order to analyze emotions, ranging from social media sites for various marketing research channels [10]. A final study showed the relationship between Customer Relationship Management (CRM) and BD. Although CRM deals with understanding human behavior and interests, BD can be expected to improve customer relationships as it allows for multi-directional communication and customization. BD enhances CRM strategies by better understanding customer habits and behaviors so that business can offer CRM to be more personalized and customized for each customer. Finally, CRM will make huge data with tools and better strategies more personalized to customers because they understand the well targeted audiences and the intended messages to be sent The previous studies showed that BD is important to implemment in banks, especially in developing marketing strategies [11].

The Proposed Marketing Intelligence System
Following are the stages of Marketing Intelligence System implementation.

Data Collection Methods
1. Interviews with 5 CIOs and 5 IT managers of Egyptian commercial 2. Survey for customers to see their satisfaction with the products offers by the bank and the way they prefer to buy products (cash -credit card ). 3. Access to transactions for the customers.

Data sources
1. Customer satisfaction Report It was found that 20% of customers still don't know the products that can be purchased through the bank, 37% are not satisfied from the services provided by the bank through credit cards or customer service. 43% of customers have satisfaction with the services of the bank.

Tables
The data were collected from the bank for clients in 2016 the data are in CSV format dataset about (39,000 record) transactions for and (4,000 record) for cardholders. These data were analyzed by using Apache Hadoop because it is scalable, cost effective, flexible, fast, and resilient to failure [14].

Tool Used :Apache Hadoop
It is collection of open-source software utilities that facilitate using a network of many computers to solve problems involving massive amounts of data and computation.
It provides a software framework for distributed storage and processing of BD using the MapReduce programming model. It is originally designed for computer clusters built from commodity hardware still the common use it has also found use on clusters of higher-end hardware. All the modules in Hadoop are designed with a fundamental assumption that hardware failures are common occurrences and should be automatically handled by the framework. The base Apache Hadoop framework is composed of the following modules: Hadoop Distributed File System (HDFS) : a distributed file-system that stores data on commodity machines, provides high-throughput access to application data.
Hadoop YARN : is a platform responsible for job scheduling and cluster resource management. Hadoop MapReduce an implementation of the MapReduce programming model for parallel processing of large data sets [12].

Results and Discussion
The following transaction types have been taken into consideration: Allowance / salary -Account transfer -ATM -Cared card (Debit -Credit) contains (4,000 records) of transaction data for cardholders -Balance From the following analysis, the following results were concluded:  debit frequency, can be sold a credit card offering high credit capabilities.

4.1.5
Transaction data may also be used to estimate consumers who can sell types of financial products. It is frequently used by banks to divide and target potential customers.

4.1.7
This person has a capacity to spend, and though infrequently, he/she has spent more than what is credited to his account during certain periods.

4.1.8
The person is also a perfect candidate to use a credit card. A credit card can be provided or approved .If it already uses one, its credit limit can be increased. Offers associated with a credit card can be extended to this person as they are likely to use their cards.