YAPI KREDI, a leading Turkish financial institution, enhances revenues, profitability and operational efficiency via introduction of advanced analytical CRM capabilities.

The Challenge

The mission of YAPI KREDI’s executive management team is to transform YAPI KREDI into Turkey’s leading financial institution and to maximize its shareholders’ value.

Accordingly, among the top strategic priorities YAPI KREDI has established revenue enhancement through focusing on growth in retail banking by increasing operational efficiency and by aligning performance management to value maximization. YAPI KREDI aims at becoming the fastest growing retail bank in Turkey in terms of assets under management and retail profitability. It targets to maintain a banking relationship with 60% of the segmented retail customer base.

To this end, YAPI KREDI introduces a new retail banking approach that enables it to serve all customers according to their specific needs through individual product packages. The new capabilities required to achieve these goals are

• Advanced analytical customer segmentation
• Segment specific offering of product bundles
• Conversion of customers to more profitable segments via campaign management through advanced CRM tools
• Increasing contract per customer ratio to 5
• To enhance operational efficiency YAPI KREDI will encourage the usage of alternative delivery channels among customers.

The Strategy

YAPI KREDI chose to systematically implement a new CRM department in the retail department to execute the above strategy. This group relied on the excellent skill set of the acquired team members and on new leading technology and customer analytics. After completing Data Warehouse and campaign management infrastructure YAPI KREDI decided to start a project designed to provide significant impact and with the following objectives:

• Put a complete CRM process in place
• Provide a clear understanding of the customer base
• Identify homogeneous segments
• Keep valuable customers and serve others with less cost
• Increase revenues with predictive model based cross-selling
• Increase operational efficiency
• Obtain data mining knowledge transfer from external consultants

The project was divided in two sub-projects with the following goals:

1. Advanced customer segmentation:

Better understanding and analysis of the customer base and consequent business process reorganization

• Clear definition of profitable customer segments
• Facilitate new product planning and pricing
• Release new product bundles
• Improve the accuracy of cross-selling prediction

2. Multi-channel predictive model based cross-selling campaigns:

Primary target was to acquire new money from customers
Secondary target was to have customers transfer their existing investments into mutual funds
YAPI KREDI chose SAS Institute as the partner and the software vendor in this project and Dr. Frank Block, then in his previous role as Chief Technology Officer and co-founder of Predict SA, supported YAPI KREDI over the time frame of approximately one year in his role as business project leader. His main tasks were to ensure seamless introduction of leading analytical CRM capabilities and maximum leverage of the new data mining technology by the relevant business processes. In this function he

• Led interviews with business departments at YAPI KREDI to assess business needs and priorities
• Gave planning support for project milestones and deliverables
• Controlled analysis outcomes and reported to YAPI KREDI executive management
• Carried out customer segmentation analyses and interpreted results in terms of actionable business knowledge
• Led cross-sell campaign planning including definition of objectives, set-up and evaluation processes

“Dr. Frank Block played an important role in leading this project, and helping YAPI KREDI achieve impressive goals very rapidly”, says Mehmet Kirca, head of CRM at YAPI KREDI. “His long track record in business and analytical CRM consulting were critical in setting the right business focus for this strategic project right from the beginning. With his help we could rapidly obtain added business value and important knowledge transfer to our department’s staff. Dr. Block always listened and very efficiently translated our business needs into technical and system requirements.”

Results

Segmentation

It was essential that segments fulfilled certain conditions in order to be of value, i.e. they had to be:

• Identifiable: precise description of the members of each segment
• Accessible: the descriptive variables had to corresponded to easily accessible and addressable features of the segments, so that each segment member could easily be recognized and contacted.
• Actionable: all gained customer insight on segment membership had to be transformed into business action.
• Stable: customers should not too easily jump from one segment to another and back over short time frames.

For the segmentation two sets of customer variables were used:

• Active variables used for finding the segments related to usage of savings and investment products, automatic payment behavior, loan usage, customer’s tenure, customer’s frequency of usage, and customer’s life cycle.
• Profiling variables used for describing the segments found in an intuitive way such as demographic, product ownership, channel usage, assets, and profitability variables.

Two types of segmentation techniques were chosen and carried out in parallel for quality check purposes:

• Unsupervised segmentation analyses with K-means clustering
• Supervised segmentation according to customer’s total assets with a pre-defined 5-level scale

Several segments were identified and named as follows: Simply the best, golden seniors, payroll customers, young investors, light users, and credit cards only. The results obtained were impressive. Among some customer, segments that corresponded to established knowledge at YAPI KREDI some unexpected segments appeared too. These could easily be associated with immediate cross selling and product bundling opportunities.

Cross-Selling

After a deep analysis of potential products to be offered during this first predictive model based cross-selling campaign it had been decided to choose YAPI KREDI’s B-type mutual funds, characterized by being low risk investment instruments based on fixed income securities and by being easily bought and sold via ATM, Web, and Telephone.

A small two-step pilot cross-sell campaign was developed targeting only customers with a high propensity to invest into these funds: first, these customers were contacted either by branch personnel or through the YAPI KREDI call centre. Then, positive responders received a follow up call if they had not purchased until one week later.

Based on six months of customer data history, various predictive models were developed to estimate a customer’s propensity to invest into a B-type mutual fund during the following three months period. All customers were scored using the best model and a set of 4200 customers with highest propensity to purchase was selected for the pilot campaign. A set of 16 branches was selected for contacting about 3000 customers and the remaining 1200 were contacted by the call centre.

An impressive response rate of 6.5% was obtained with the branch based and of 12.2% with the call centre based part of the campaign. The pilot campaign could acquire more than € 1 million into B-type mutual funds. A large-scale rollout to a larger part of YAPI KREDI’s customers holds the promise for overwhelming results.

Key Benefits

1. Discovery of new segments with new needs and business opportunities
2. Increased response rates of cross sell campaigns
3. More new money
4. Cost savings through migration to alternative delivery channels
5. Acquired critical knowledge on analytical customer modeling
6. Reorganized business processes to leverage potential of new technologies and capabilities introduced with analytical
7. CRM
8. Increased frequency of direct marketing campaigns