Publiceret: 2/9/2014
Visninger: 256
Bedøm casen:

A&P Financial Credit scoring system on the cloud, Rush \u0026 Cash uses OSS on Windows Azure as a bridgehead into the Chinese

“We thought about sending developers to China, but changed our mind as NICE Information Service suggested that it would be possible to develop the system from Korea so long as we adopted Windows Azure. We also learned that doing so would eliminate our worries about in-service delays that would otherwise arise in the course of communication between our headquarters in Korea and developers in China, and the possibility of exposing our knowhow to local IT companies. It did not take much time to decide on a Windows Azure-based hybrid cloud model”
Director at A&P Financial

A&P Financial, which tends to be better known as the consumer finance service provider Rush & Cash, is stepping up its expansion plans into the Chinese market as part of its goal of discovering a future growth engine in order to cope with the increasing competition and tightening regulatory environment seen in the Korean market. With a branch established in Tianjin in 2012 and Sunjin in 2013, A&P Financial developed a credit scoring system that rates the credit of consumers, providing and managing loans according to the credit scores it calculates. What A&P Financial hoped to achieve with the new system was credit scoring optimization, particularly given the system’s integral status within the credit/loan business. Finding the right system set-up for local market conditions was seen as facilitating business expansion across the Chinese market down the line. For this, A&P Financial adopted JBoss-based R-CLIPS developed in Java by NICE Information Service, which has a wealth of experience and expert knowledge in the credit assessment field, and customized it in a way that suits the Chinese market. In addition, it used a self-developed data analysis solution named N-Builder that combines Excel with R. In doing so, A&P Financial decided to go with a Windows Azure-based hybrid cloud model. Given that each province in China has its own specific regulations and policies, for a credit scoring system to be successful, it would seemingly need to be newly developed whenever a new branch office is established in a different province. However, this would be to incur difficulties in terms of system development and maintenance. To avoid such problems, A&P Financial decided to run a credit assessment and analysis system on Windows Azure while managing operating data in the system located on local branches in China. Thanks to this development plan, A&P Financial can now provide, trouble-free, a unified credit scoring system to the 3-4 new branch offices which are scheduled to open in 2014 while also saving up to 50% when compared to what it would have cost to develop the system in each local branch.

Business Needs

A&P Financial, famous as Rush & Cash, the largest loan lender in Korea, is continuing apace with its ambitious plan to become a global full-service financial group. The group, indeed, has already expanded its business beyond Korea into China, South-East Asia and Central Asia in recent years. It was its first branch to be established in China (in Tianjin) in 2012 that marked the beginning of its expansion into the Chinese market. The experience of growing its consumer finance brand Rush & Cash in the face of mounting competition and tightening regulations in Korea has spurred the expansion into China of A&P Financial, which established its second Chinese branch (in Sunjin) in 2013.

Its new branch office in China represents a lot more to the company than the simple addition of one more branch office in the local market. It means that A&P Financial has established a bridgehead for its expansion into the larger Chinese market as a whole. To ensure its success in China, A&P Financial decided with the establishment of its Sunjin branch to implement a credit scoring system customized to local market conditions, as doing so would allow it to push forward with its plan to open 3-4 branch offices in 2014.

In the loan business, having the right credit scoring system in place is essential. The credit scoring system rates the credit status of consumers, enabling the provision and management of loans by credit rating. It seems to be simple in terms of logic, but it is sophisticated in terms of data analysis and corporate risk-management. In addition, much effort and time are required in order to implement a system that suits all different conditions, like local policies, regulations and market situations, particularly when a loan service provider expands its business into overseas markets.

The rationale for A&P Financial’s decision to focus on the implementation of the credit scoring system was that consumer credit status needs to be assessed and rated according to various, yet sophisticated criteria for its loan programs, credit limits and interest rates to be managed well and in a way consistent with best practice in China. The question left was to whether develop the solution with the help of a Chinese solution provider or optimize the solution developed in Korea. After much consideration, A&P Financial decided to adopt R-CLIPS, a credit scoring system developed by NICE Information Service with all-round expertise and experience in the credit assessment field.

With the solution chosen, A&P Financial worked out a plan to optimize R-CLIPS in a way that would best suit local market conditions, regulations and policies. In the course of establishing a plan of action, A&P Financial studied ‘hybrid cloud’, a model that had not yet been attempted anywhere else in the world. “We thought about sending developers to China, but changed our mind as NICE Information Service suggested that it would be possible to develop the system from Korea so long as we adopted Windows Azure. We also learned that doing so would eliminate our worries about in-service delays that would otherwise arise in the course of communication between our headquarters in Korea and developers in China, and the possibility of exposing our knowhow to local IT companies. It did not take much time to decide on a Windows Azure-based hybrid cloud model”, said Director at A&P Financial.

Solution

The hybrid cloud model chosen by A&P Financial was designed to run the credit scoring system on Windows Azure and store related data on the system operated at the local branch in China. In short, data would be managed at local branches while the analysis service is run on the public cloud.

The primary reason NICE Information Service proposed this model was that it would have been too difficult to develop and maintain the various systems that would exist if each branch in China was to build its own. Whatever system was chosen, it would have to take into account the different policies and regulations that exist in each province in China. On a practical level, having different systems in each branch would pose logistical problems as to simply develop and maintain each system would require at least several hours spent on a plane in transit between branches.

Plus, it is difficult to outsource the development work to local developers. In general, companies go through many trials and errors when establishing IT systems in overseas branches. There seem to be no difference between home and aboard in terms of searching developers, engineers and service providers and ordering the various different equipment required. When it becomes a reality, however, difficulties specific to overseas become much clearer. In China, the turnover rate of IT engineers is high and local IT service providers are often slow in responding to requests related to development. Another problem was security. If R-CLIPS is implemented, installed and maintained in local branches, there is the distinct risk of A&P Financial’s credit scoring model, a valuable piece of intellectual property, being leaked to a third party. The best choice in all aspects was Windows Azure.

There were good reasons why NICE Information Service recommended Windows Azure. First, NICE Information Service developed R-CLIPS in Java with JBoss as WAS (Web Application Server). Second, N-Builder, a data analysis solution, is built on Microsoft Excel equipped with GNU R, a widely-used functional language for data analysis. Though they are built on different platforms, Java and open source, there is no problem running both on Windows Azure. Moreover, R-CLIPS was structured to allow separation of the management and execution modules. From its experience of running CSS 2.0, a credit scoring service provided in a form of ASP (Application Service Provider) in Korea, NICE Information Service discovered that with Windows Azure capable of running R-CLIPS, it is possible to develop and operate the system on the public cloud. In other words, Windows Azure solves all the problems that A&P Financial was concerned about - efficient development and maintenance. What’s more, it also eliminates the risk of its credit scoring model being leaked to a third party.

A&P Financial began building the system on Windows Azure. This project was carried out on a global-scale – with the system developed in Korea, China and Hong Kong. The service architecture was designed in a way that the execution module of R-CLIPS is controlled by the system in the data center located in Sunjin, China, whereas the module, which manages the credit scoring model, runs on Windows Azure operated by the data center in Hong Kong. In addition to this architecture, A&P Financial designed a data flow in which the credit scoring mode and strategy created in R-CLIPS management module running on Windows Azure are transferred as an encrypted metafile to the system in its local branches in China. The intended purpose here was to deliver an executable form of a file to local staff members in order to prevent information being leaked.

“The biggest advantage of the hybrid model is that it complies with regulations, but carries no security risk. A&P Financial’s credit scoring system has two different sets of data - analysis data managed by the Windows Azure data center in Hong Kong and operational data stored in the system placed in local branches in China. Many countries forbid storing consumer data on systems located overseas. Indeed, when it comes to compliance with differing security regulations, the hybrid model is ideally suited to enabling the expansion of A&P Financial across Asia, including China, because in this model, only data that can be delivered across countries, like ID numbers, is stored on the public cloud”, said Lee Seowon, Director at NICE Information Service.

Benefits

No risk of credit scoring model-related information and knowhow being leaked

The credit scoring method varies by financial group – depending on the diversity of the information it obtained and the service it provides. One recent trend in credit scoring systems is for financial service providers to develop models specific to their business. Though there are different models, each shares a common goal: to customize a comprehensive model with everything from future profit to marketing plan, while also encompassing proactive measures against possible risks and incorporating both internal and external information. Thanks to Windows Azure, A&P Financial has been able to protect its credit scoring model, which is of great strategic value, and prevent it from being leaked to a third party. “We analyze data on Windows Azure and distribute the results as an executable file to branch offices. The upshot of doing this is that the local service providers that manage the legacy system cannot see the source code”, said Director at A&P Financial.

Maintenance takes only one day

With Windows Azure making remote development and maintenance of R-CLIPS possible, the handling of requests related to the credit scoring model has become much faster and happens now almost in real time. “The logic of the credit scoring model does not change much, but small modifications are needed whenever the service strategy changes. In general, we make big modifications from the whole-system view every six months and small ones with almost every month. If we had to develop and maintain the system at each local branch, it would have taken at least one week to make each small modification one by one and spoon-feed those into system management. Thanks entirely to Windows Azure, it now takes less than a day to, for example, distribute new rules using the R-CLIPS management module from Korea”, said Director at A&P Financial.

Development expense savings of as much as 50%

As it implemented the hybrid cloud model, A&P Financial does not need to send developers and IT engineers to local branches in China. The only thing developers have to do is access Windows Azure and remote control the system. As a result, A&P Financial was able to save TCO. “If we didn’t have Windows Azure, we’d have to fly at least 2-3 developers to China to do the work there. Not only that, but we’d need to send IT managers on regular business trips to check on the progress of the work. Windows Azure makes possible remote development, meaning we did not have to send developers and/or IT managers abroad, except at the very outset, when they were sent to analyze local requirements. Considering air transportation costs and the time involved, we saved about 50% in development expenses”, said Director at A&P Financial.

Momentum to speed up expansion into overseas markets

Windows Azure is seen at A&P Financial as providing the necessary platform for its rise to prominence as a global full-service financial group. A&P Financial now plans to open 3-4 more branch offices in 2014. What tends to happen is that, when expanding in China, companies are weighed down by needing to establish separate systems for each province in which it operates due to the differing local regulations and policies, customizing IT systems, operating strategies and even the credit scoring model. Whilst the same is true for A&P Financial, the adoption of Windows Azure means it is now able to support the development and maintenance of the credit rating system remotely, without having to send developers and/or engineers abroad, no matter where its new branch offices are established. Its plan is to make the most use of Windows Azure in expanding the credit scoring model service while leveraging its advantage in development cost-reductions and efficient maintenance in its rise to dominance in the Chinese market.

Microsoft Windows Azure

For more information, see the Microsoft website.

www.windowsazure.com/en-us

Løsningen - kort fortalt



Organisations størrelse
1000 medarbejdere

Organisationsprofil

A\u0026amp;P Financial is a full-service financial group whose consumer financing brand Rush \u0026amp; Cash is particularly well-known. Indeed, Rush \u0026amp; Cash was founded in 1999 and, since then, the group has diversified its business across various fields and expanded into overseas markets like China and Vietnam. Business Challenge


Løsningen
A\u0026amp;P Financial chose to run R-CLIPS (Rule based Credit \u0026amp; Lending Integrated Planning System), part of NICE Information Service, on Windows Azure together with N-Builder, a self-developed data analysis solution that combines Excel and R. For reference, R-CLIPS is written in Java and is based on JBoss.

Fordele
No risk of credit scoring model-related information and knowhow being leaked
Development expense savings of as much as 50%
Maintenance takes only one day
Momentum to speed up expansion into overseas markets

Software & Services
  • Microsoft Online Services
  • Microsoft Azure Platform

Industri(er)
Banking

Land
Korea

Forretningsbehov
Cloud & Server Platform

Udfordringen
  • Cloud Services
  • Virtualization

Languages
English