 |
||||||||||||||
 |
||||||||||||||
|
Experian Appoints Country Manager for South African Operation Kim Jenkins is appointed to manage the South African office and to lead Experian's expansion into Africa. ABN AMRO: Improving collections productivity by more than 70% ABN AMRO selects the Tallyman Debt Management solution from Experian Decision Analytics to automate and streamline its collections process. Find out more about the decisioning workshops on offer this year. Features Scorecard Development Process – Building and testing a model The fourth article in a series relating to scorecard development and range of scoring models. Customer management for Retail Banking Find out how the implementation of a customer level decision-making approach can significantly increase profitability - An Experian Decision Analytics white paper. Fraud Focus Tackling the issue of bust-out fraud Bust-out fraud is a growing area of fraud for the financial services industry. This white paper concentrates on the methodologies associated, how it is perpetrated and how to predict bust-out fraud. Careers Exciting Career opportunities available now at Experian Find out what jobs are available. Experian appoints country manager for South African operationJohannesburg , South Africa , 14 January 2008 - Experian®, the global information services company, today announced the appointment of Kim Jenkins as Country Manager for Experian in South Africa . Kim will be responsible for managing the South African operation and also for leading Experian's expansion into Africa .
Kim has more than 10 years experience in the financial services industry, having held senior positions in some of the country's largest banks. Kim joins Experian from First Rand Bank Limited where she most recently pioneered the group's entry into India . Prior to this, Kim held senior management positions within First National Bank, including that of Head of Strategic Delivery and then Head of Consumer Segments. Earlier in her career, Kim was involved in product delivery across multiple portfolios within, the then, Nedcor Limited and Citi Bank.
Luciano Manzo , Senior VP Southern Europe, South Africa and Emerging Markets for Experian, said: “ Africa is a dynamic growth region for the group and an integral part of our long term investment strategy. As such, we were looking for a strong leader that not only had a good understanding of the market but also had a strong client focus, people management skills and cross-border experience. Kim's impressive management portfolio made her the ideal candidate to further grow both the Credit Services and Decision Analytics' lines of business in the region. ”
ABN AMRO: Improving collections productivity by more than 70%ABN AMRO has 20 million consumer clients, ranks eighth in Europe and 13th in the world based on total assets and has more than 4,500 branches in 53 countries. In recent times, banks have seen an increasing level of delinquency in the dynamic consumer credit market, and are having to increase provisioning as a result. ABN AMRO itself saw a 30% increase in loan impairment from 2005 to 2006.
This increasing default rate put pressure on the collections operation to improve recovery rates and reduce write-off values. As part of the bank-wide project to overhaul the credit and risk systems following implementation of the Basel II programme, the bank recognised that the collections operation could make a significant contribution to improving impairment levels and managing provisioning levels.
The current early collections operation had been centralised into regional centres from a branch based operation and later into one centre. Some portfolios were still managed in the branches which had led to a sub-optimal approach. Staff were highly motivated but lacked specific collections experience and training, and were using a generic workflow management system without specific collections functionality. Therefore, a collections programme was started, aimed at improving the early collections activity by implementing a dedicated debt collection system, enhancing processes and operations and centralising the remaining decentralised operations.
ABN AMRO selected the Tallyman Debt Management solution from Experian Decision Analytics. This specialist collections management solution enables ABN AMRO to automate and streamline the collections process to collect more debt from a complete insight in the entire customer relationship. The solution integrates with the existing customer management solution, Probe SM, for a customer-centric approach. Tallyman uses behavioural scoring to create an accurate single customer profile and, using this understanding of the behaviour and motivations of the customer, it segments customers and assigns the most effective action path. This sophisticated decisioning is dynamic and updated daily to reflect the most recent activity, promise to pay and change in status.
“Improving our collections operation was a key part of our credit and risk strategy but we were restricted by operational and IT restraints,” remarked Martin van Loon, Senior Vice President, ABN AMRO . “Experian was able to offer a dedicated collections solution which could be deployed rapidly and, just as importantly, deliver the business consulting skills and best practice experience to support its solid implementation capabilities. We have learned that IT projects can go fast but it is essential to have focus and a true partnership with the key vendors, as we have had with Experian. You just have to look at the results we have achieved, such as productivity up by 70% and provisioning levels reduced by more than 20%, to see the value that this project has given to ABN AMRO.”
Experian's Workshops for 2008Experian's decisioning workshops have become an integral source of learning for the South African credit industry. Its range of scoring workshops are positioned to meet the needs of both the seasoned credit risk professional as well as new staff across the banking, finance, and telecommunications industries. The detailed curriculum of all workshops are continually reviewed to ensure that the content is up-to-date with the latest topics, global trends and techniques in the credit industry.
Experian's workshops provide strategic guidance to improve decisioning across the organisation. The small class size and expert instruction from highly skilled and experienced presenters ensures that maximum value is received. The attendee evaluations allow for the incorporation of valuable feedback into the evolving format.
Experian's 2008 workshop calendar includes an Introduction to Scoring workshop and a scorecard monitoring workshop for the Scoring Analyst and Business Manager. Specialised training workshops can also be arranged at a client's site, at your convenience to address business-specific requirements.
Scorecard Development Process – Building and testing a modelThis is the fourth article in a series relating to scorecards. The approach to model building, whether for an application or a behavioural model, is based on the same principles. This article describes each of the stages, discussing the considerations that need to be made at each point.
Grouping of variables
Variables that might be included in a scorecard must be grouped into coarse bands. This is done for 3 main reasons:
• Model stability Variables consist of many values called attributes and some of these attributes may be populated with only a few record counts, if any at all. If these attributes are not grouped together it may cause instability within a model. The general rule of thumb to apply is to group at least 3% and generally no more than 15% of the population together. There may in some cases be good reasons to deviate from this, for instance: If only 2% of the population have judgments, they would still be classed together, as this group inherently displays very different behaviour to a group without any judgment information. It is also important to keep in mind that the number of groups for one variable should typically not exceed ten.
• Predictive strength The grouping of similar attributes with similar predictive strengths will optimise the overall predictive strength of the variable. The strength of a variable refers to the ability of the data to differentiate between 'good' and 'bad' re-payers and can be measured by looking at the characteristic strength.
• Logical trends Finally, the variables should display a logical trend. Variables with many attributes like ‘applicant age' may display some anomalies within the trend when looking at the raw attribute values and those trend reversals can be corrected by grouping attributes together.
Building the model
The next stage of the model build is to select the initial set of variables to be included in the scorecard. Many modelling techniques use a stepped approach where variables are prioritised and considered for the scorecard at different levels. There are typically a large number of variables and it is therefore prudent to group the variables into different categories such as positive and negative bureau data, and application variables. These can also be further categorised by the strength of the variable. These different categories of variables can be modelled with different degrees of priority using the stepped approach. After a series of iterations the most predictive variables will become clear and the optimum combination for the model can be selected. In addition to the predictive value of variables, it is also important to consider the balance of variables in the model, the operational considerations of using each variable as well as the business requirements for and the intention of the model. The reason for this detailed approach is to ensure that the scorecard does not rely heavily on a small number of variables and is a strong, robust and balanced model.
Validating the scorecard
To check that the model is representative and not biased by the dataset used, it is validated using the 80-20 rule. This means the scorecard is built on 80% of the population and the remaining 20% is used to validate the model. A scorecard validates if the 80% and 20% samples have similar score distributions for good and bad accounts and a number of statistical tests are met, for example, the KS test. The business impact of the model must also be assessed. This is done by comparing the acceptance rates and bad rates for each group of the variables with that of the previous decision process. For example, the acceptance rate should increase as the bad rate decreases for the variable ‘applicant age'. If the scorecard doesn't validate, further investigation needs to be done to determine the reason and appropriate changes should be made to ensure a valid scorecard.
After a model has been built it is advisable to do an out of time validation using more recent data. This is done to ensure that the scorecard works as expected on a sample from a different time period.
Read next months edition of Credinews to find out more about reject inference.
Written by Cezanne Gentle, Scoring Analyst, Experian Decision Analytics.
Customer management for Retail BankingRetail banks are continually striving to improve their risk adjusted return on capital. Bank profitability is driven by revenue, costs and bad debt losses whilst maintaining high levels of customer satisfaction and regulatory compliance.
All of these are directly influenced by the daily decisions made in managing relationships with customers. These decisions are powerful levers in managing the ultimate profitability of the bank. In many cases growth strategies have sought to strengthen relationships by increasing the number of product relationships with each customer. However, this increases the complexity of the customer relationship.
Most banking organisations have some level of automated decisioning in place but few do this in an holistic way. Banking organisations have tended to develop on a ‘silo' or product-level basis, with risk (and other) managers within each silo being responsible for decision-making, scoring and strategies for their silo alone. Whilst this gives clear line of sight to roles, responsibilities and objectives at the product level, it has often led to a confusion of approaches to decision making for customers when looking across silos.
The fact is that a customer's performance on their current account gives the organisation meaningful insights into the customer's likely performance on their credit card or their personal loan, and vice versa. And yet, how many organisations take the full customer view into consideration when making credit risk and other decisions and interacting with their customers? Those that do report that the potential benefits can be realised across a range of profit drivers.
To make these decisions accurately, it is essential to truly adopt a customer focussed approach. In practice, this involves gathering together all the information on the performance of the customer and using analytics to drive customer focused strategies.
The implementation of a customer level decision-making approach is a major change programme, given the product silo legacy in many banks, but one with significant rewards. Overall, the bank is able to make better and more consistent lending and management decisions as well as delivering a high standard of customer service, because decisions are made based on the relationship the bank has with the individual customer, not the account.
Tackling the issue of bust-out fraudBust-out fraud is a growing area of fraud for the financial services industry. For organisations across the globe, bust-out fraud is a very topical issue and the losses incurred are becoming significant. In the UK, all the major banks have reported an increase in bad debt and provisions in the last two years, part of which could be attributed to this type of fraud.
Identification of bust-out fraud is difficult for many organisations, although Experian's research shows that there are many strong predictors including current account behaviour and transactional patterns, credit bureau trend data and 'event' trigger data.
This white paper concentrates on the methodologies associated, how it is perpetrated and how to predict bust-out fraud.
Exciting career opportunities available now at ExperianThe following jobs are available currently at Experian:
Solutions Analyst - The successful applicant will be responsible for: Solutions build, software and/or scoring, Impact analysis of changes, Unit and systems testing, Day-to-day project and client support, Problem solving and innovation, Sales support as and when required. Click here for more information
Technical Consultant - The successful applicant will be responsible for Technical consultancy and development on low to medium complexity delivery projects for all technical elements for selected products in the EDA product set. Click here for more information
Legal and Compliance Officer - The purpose of this job is to provide an administrative and legal and compliance service to both the EDA and ECS businesses under the direction of the Commercial Manager. Click here for more information
About Experian Experian is a global leader in providing information, analytical and marketing services to organisations and consumers to help manage the risk and reward of commercial and financial decisions. Combining its unique information tools and deep understanding of individuals, markets and economies, Experian partners with organisations around the world to establish and strengthen customer relationships and provide their businesses with competitive advantage. For consumers, Experian delivers critical information that enables them to make financial and purchasing decisions with greater control and confidence. Clients include organisations from financial services, retail and catalogue, telecommunications, utilities, media, insurance, automotive, leisure, e-commerce, manufacturing, property and government sectors. Experian Group Limited is listed on the London Stock Exchange (EXPN) and is a constituent of the FTSE 100 index. It has corporate headquarters in Dublin , Ireland , and operational headquarters in Costa Mesa , California and Nottingham , UK . Experian employs around 15,500 people in 36 countries worldwide, supporting clients in more than 65 countries. Annual sales are in excess of $3.8 billion. QUICKLINKS Introduction to Scoring Workshop Experian South Africa 5 March 08 Little Tuscany, Bryanston For further information click here
Experian Decision Analytics Forum Experian UK 6 March 08 The Nottingham Belfry For further information click here
Kim Jenkins
Click here for the workshop calendar
Click here to view the previous article in the Scorecard series
Click here to request white paper
Click here to request white paper
Johannesburg + 27 11 799 3400 Cape Town + 21 21 418 6454
Copyright © 2008 Experian South Africa |
||||||||||||||
