Propensity modeling and marketing evolution a powerful combination at Yapi Kredi
When its cross-selling campaign to promote telephone banking began to falter, Yapi Kredi used SAS Enterprise Miner to build propensity models of customer behavior. The results were positive – but analysis established that they had to do more than select the right customers.
Established in 1944, Yapi Kredi Bank has introduced products and services that changed the face of banking in Turkey. Today the company has $15 billion in assets and $9 billion in deposits. Yapi Kredi's 9 million customers are serviced by 10,000 staff at more than 420 domestic and subsidiary branches, affiliated companies in leasing, factoring, investment banking, insurance, brokerage and the new economy. Complementing its wide domestic network, the bank also maintains an important international presence with a subsidiary bank in Düsseldorf and Amsterdam, a bank in the Russian Federation, an off-shore banking unit in Bahrain and four representative offices in Moscow, Munich, Cologne and Stuttgart. Yapi Kredi is also Turkey's leading credit card issuer with some 17 percent of the market in cards issued and 28 percent in business volume.
The company's recent success has been built largely on an integrated, multi-channel customer relationship management strategy. In addition to the retail branch network, Yapi Kredi interacts with customers through direct mail, e-mail, WAP, Web sites and digital television. So the company has built up a lot of customer and product information in its legacy systems, which is extracted into a data warehouse. Since July 2001 Yapi Kredi has been exploiting this data for campaign management and customer relationship management, with a CRM group reporting directly to the executive vice president of retail banking.
"We value one-to-one communication with our customers very highly in our organization," says Arzu Umur, marketing analysis manager at Yapi Kredi Bank. Following an evaluation of SAS Enterprise Miner, Yapi Kredi began implementing the solution. "With the help of SAS we could finish our customer segmentation and start using more sophisticated profiling models," says Umur.
In 2002, Yapi Kredi ran cross-selling campaigns to recruit customers to telephone banking, using basic supervised selection methods, and achieved a response rate of 12 percent. But when these campaigns started to falter, Yapi Kredi decided to apply SAS Enterprise Miner technology and the SAS data mining methodology to breathe new life into the strategy. "We were having difficulty identifying new target customers to register for telephone banking services," explains Umur. Yapi Kredi tried loosening the targeting criteria, but this only reduced the response rate, making the campaign activity less profitable. So the Market Research Team looked at historic customer behavior to try to predict which customers would be most inclined to register for telephone banking within a three-month period.
The team developed seven alternative propensity models with SAS Enterprise Miner, selecting the regression model as yielding the best results with a lift factor of 21 and a captured response rate of 21 percent for the top percentile. Based on these findings, in the second half of 2003 Yapi Kredi made further improvements to its campaign strategy, applying SAS data mining to the selection procedure and running a major promotion campaign.
But the data mining methodology does not end with deployment of the model; it involves monitoring and analysing results of the campaign, further developing the model and adjusting the strategy accordingly. Yapi Kredi found that selection of the right customer was only one aspect of the challenge; selection of the right channel was also critically important. "In 2003, data mining raised the response rate from 13 to 23 percent for call center campaigns, and from 11 to 15 percent overall," says Umur. Yet it had very little impact on cross-selling efforts in the branches.
"So in 2004, we stopped sending the customer to the branch," says Umur. "Previously we sent all payroll customers there because we assumed that the staff at the branches would know these customers better than anyone else. But our analysis using SAS showed that there was no significant difference in uptake among payroll customers, so we diverted them to the call centers and an online procedure for signing up for telephone banking." In 2004, Yapi Kredi ran two parallel campaigns using supervised and data mining selection methods. Call center response rates increased significantly with both the supervised selection procedure (from 13 to 27 percent) and the data mining selection procedure (from 23 to 28 percent – compared with control group responses of around 1 percent). "Shifting customers to the online membership procedure and calling payroll customers had a major impact," says Umur. Overall response rates using data mining increased from 15 percent in 2003 to 17 percent.
Umur concludes from the experience that "analytical success is important, but it is not enough in itself to guarantee success. SAS Enterprise Miner provides an easy procedure for improving customer selection for direct marketing campaigns. But you have to look at and change many factors and lower barriers to entry. Customers are more responsive when they are contacted via a channel that is suitable for giving the appropriate advice." Sometimes business assumptions do not work in practice, and Yapi Kredi has learned how SAS Enterprise Miner can help establish why that is the case.
"The result of the analysis is not only an increase in revenues, but also an overall improvement of the banking relationship," concludes Umur.
The results illustrated in this article are specific to the particular situations, business models, data input, and computing environments described herein. Each SAS customer’s experience is unique based on business and technical variables and all statements must be considered non-typical. Actual savings, results, and performance characteristics will vary depending on individual customer configurations and conditions. SAS does not guarantee or represent that every customer will achieve similar results. The only warranties for SAS products and services are those that are set forth in the express warranty statements in the written agreement for such products and services. Nothing herein should be construed as constituting an additional warranty. Customers have shared their successes with SAS as part of an agreed-upon contractual exchange or project success summarization following a successful implementation of SAS software. Brand and product names are trademarks of their respective companies.
Copyright © SAS Institute Inc. All Rights Reserved.
Marketing Analysis Manager at Yapi Kredi Bank
Reverse the trend of decreasing response rates to direct marketing campaigns
SAS Enterprise Miner provides intelligence that increases cross-sell rates, improves revenues
“ The result of the analysis is not only an increase in revenues, but also an overall improvement of the banking relationship. ”
marketing analysis manager