SOLVING: Three Dot Four Capital Management John Galt was recently promoted to Senior Vice President of Consumer Lending at Three Dot Four Capital Management. Three Dot Four is a large financial services organization ranked among the top 20 financial institutions in terms of total assets. A key consideration in John's promotion was his past success leading the bank's quality and productivity group. In particular, there appears to be a significant amount of dissatisfaction among both the bank's customers and the bank's loan officers with its online mortgage application process. The online mortgage application process is initiated when a customer clicks on the Apply for a Home Mortgage Now link on Three Dot Four's homepage. This link takes the user to an Instruction page that overviews the application process and provides a checklist of the information the applicant will be asked to supply on subsequent Web pages. Clicking on the Continue button at the bottom of the Instruction page takes the user to the first of four Web pages, each containing a Webbased form to collect the required data. The first Web page, Personal Information, solicits information regarding the applicant and requires entering information in 33 fields. Information collected on this page includes the applicant's and coapplicant's names, full addresses, previous addresses, dates of birth, Social Security numbers, phone numbers, email addresses, and so on. To continue on to the next page, the user selects the Continue button located at the bottom of the Personal Information page. When the Continue button is selected on a given page, a check is made to ensure that none of the required fields has been left blank. If the validation check is passed, the next Web page in the application process is displayed. In cases where the validation check fails, the blank fields are highlighted and the user is asked to enter the information in these fields. Once information has been entered for all required fields on the Personal Information page, the second pageProperty, Loan, and Expensesis displayed. This page is used to collect information about the property the loan will be used to purchase, the type of loan the applicant desires, and information about the applicant's monthly expenses. In total, the Property, Loan, and Expenses page contains 10 data fields. Once the information has been entered for all required fields on the Property, Loan, and Expenses page, the Employment page is displayed. This page captures information about the applicant's and coapplicant's employment history, including salary and other income information. The Employment page contains 16 user fields. Finally, the last Web page in the application process captures information about the applicant's Assets and Liabilities. In particular, the user is asked to supply information about checking accounts, savings accounts, credit card accounts, investment accounts, car loans, and so on. In total, this page contains 22 data fields. When the applicant clicks on the Submit Application button at the bottom of the Assets and Liabilities page, a final validation check is performed and the information is transferred to one of the bank's loan officers. The loan officers subsequently print out the information and then add the application to their backlog of other inprocess applications. To even out the work across the loan officers, all loan officers process loan applications submitted via the Web, as well as applications received via the mail and applications completed at one of the bank's branch offices. Initially, John identified two areas in need of improvement: the fairness of loan approval decisions and the accuracy of the information in loan applications submitted online. In terms of the fairness of loan approval decisions, over the last couple of years the company has received numerous complaints from applicants questioning the organization's fairness in making loan approval decisions. To begin understanding this problem, John initiated a study in which 25 loan applications were randomly selected. These loan applications were then evaluated by a panel of three experts to determine whether the loan should be approved or rejected. Next, three loan officers were selected and asked to evaluate each of the 25 loans two times. The data collected from this study is summarized in Table 1. To investigate the issue related to the accuracy of information in online mortgage applications, John formed a process improvement team. The team began by collecting data on the total number of hits each page in the Web application process received as well as the number of times the page was actually completed during the month of January. In addition, the team performed a detailed audit of all the information that was submitted during January and tallied the number of fields that contained errors across all submitted information. A summary of the team's preliminary results is given in Table 2. Table 1 Summary of Loan Approval Fairness Study Loan Officer 1 Loan Expert Panel Loan Officer 2 Loan Officer 3 01/01/20 02/01/20 01/01/20 02/01/20 01/01/20 02/01/2005 05 05 05 05 05 1 A A A A A A A 2 R R R R R R R 3 A A A A A A A 4 A R R R R R R 5 R R R R R R R Loan Officer 1 Loan Expert Panel Loan Officer 2 Loan Officer 3 01/01/20 02/01/20 01/01/20 02/01/20 01/01/20 02/01/2005 05 05 05 05 05 6 R R R A A R R 7 A A A A R A A 8 R R R R R R R 9 A A R R R A R 10 R R R R R R R 11 R R R R R R R 12 A A A A A A A 13 A A A A R A A 14 R R R R R R R 15 R R R A R R R 16 A A A A A A A 17 A A A R A A A 18 A A A A A A A 19 R R R R R R R 20 R R R R A R R 21 R R A R A R A 22 A A A A A A A 23 A R R R R R R 24 A A R A R R R 25 A A A A A R A A = Loan approved R = Loan not approved Table 2 - Online Mortgage Application Submissions, January 2005 Web Page Number of Number Number of Hits Submitted Errors 108.571 68.400 45.144 68.400 62.928 22.025 Employment information 62.928 59.781 28.695 Asset and liability information 59.781 52.009 51.489 Personal information Property, loan, and expense information 1. What is the DPMO for the loan applications submitted via the Web? 2. What could be done to improve the DPMO? 3. Approximately, what is the process sigma of the loan application process? 4. Develop intuitive measures of repeatability and reproducibility' for the loan approval process. What do the results of this analysis tell you about the fairness of the loan approval process? 5. make? Regarding the fairness of the loan approval process, what recommendations would you