Deutsche Allgemeinversicherung was founded in 1966 by Andreas Steininger. It was one of the world’s largest insurance companies. Around 51% of its business was in Germany with a 60% of its business in the field of retail insurance. Its keys to success were the traditional insurance management and its remarkable customer service. In DA, a typical process flow of a new policy would start by either a customer going to one of the companies branches or contacting a company agent. The customer would fill out an application – sometimes a check was attached- and the application was then sent through the company mail to the Retail Transaction Processing (VEG) division in Hamburg. Also, some customers filled out applications at home and sent them directly to a number of DA locations which would then transfer them to the Hamburg operation. Once the application was received, VEG separated the various parts of the application then scanned it and digitized. The electronic image was then retrieved from a server and delivered to associates desktop client computer. The associate would enter the information on the form into an appropriate database. If the information available in the application was complete a confirmation would be sent to the customer. In case the customer find out something wrong on the notification notice he/she would either call a toll number or send a letter describing the problem. At that point the Customer Problem Resolution division would deal with the problem. On the other hand, if the information in the application was not complete another associate responsible of talking to customers would call the customer to obtain the needed data. The management of DA knew that they delivered a high quality service, but what bothered them was the percentage of the first time accuracy. This number was 99% which made the management believe that there was something wrong with the peoples’ data. Around 10% of the forms were entered into the system with some mistake. The management thought that it needed to get real accuracy numbers before it begun to improve, but shortly, it was then clearer that it was more important to improve the accuracy number than knowing exactly what it was. That is to say how to properly measure the accuracy of the process and transfer this measurement into an effective tool that leads to change which is the process continuous improvement. This is why the management decided to use SPC (Statistical Process Control). DA implemented SPC at this point because it aimed at improving the process itself. It used SPC to measure the process rather than the people who were involved in it. From a pure statistical point of view, the level of variance was low yet the process performance did not meet the specification limits. Accordingly, it was a process issue not an individual outlier issue. This is why DA used SPC at this point. In order to apply and obtain SPC results, a close look is needed (for example) at the first 12 weeks of data (Exhibit 4) that represent the diagnostic period for Policy Extension group in order to construct the 3-sigma control limits for the process and highlight in which (if any) of the subsequent weeks the process was out of control. In order to achieve this, certain steps need to be carried out:

1. Calculate the sample proportions, p (these are what can be plotted on the p-chart) for each sample Week 1 2 3 4 5 6 7 8 9 10 11 12 Sample Size 300 300 300 300 300 300 300 300 300 300 300 300 Errors 18 15 18 6 20 16 16 19 20 16 10 14 P 0.06 0.05 0.06 0.02 0.06667 0.05333 0.05333 0.06333 0.06667 0.05333 0.03333 0.04667

2. Calculate the average of the sample proportions p p= 188 = 0.05222 3600

3. Calculate the standard deviation of the sample proportion p (1 - p) = n

sp =

.05222 (1 - .05222) = .012844 300

4. Calculate the control limits (in this case z=3) UCL = p + z sp = 0.05222 + 3 x 0.012844 = 0.090752 LCL = p - z sp = 0.05222 – 3 x 0.012844 = 0.013688

5. Plot the individual sample proportions, the average of the proportions, and the control limits

Once the upper and lower limits are identified it becomes easy to recognize in which (if any) of the subsequent weeks was the process out of control. This could be carried out by calculating the p for the subsequent weeks and checking whether they were within the limits or not. In this case, in week 23 and 24, the process was out of control because the p of both weeks were higher than the upper limit (0.11 and 0.1533 consecutively) Now after coming up with these results and before going on with what Annette Kluck should consider as next steps in DA’s quality efforts. We need to highlight the challenges that faced Annette when applying SPC as an insurance company or in other words the challenges that could face insurance companies in general when implementing SPC. These challenges could be as follow: • Successful companies such as DA have processes with high accuracy value (like in this case 99%). Accordingly, the company has to increase the confidence interval in the sample hence resulting in an increased sample size which at some point can be either financially unfeasible or physically unfeasible. A company has to specify clearly what is a defective error that renders the full application as a defective application and the non critical error that could happen but could not deem the application as defected. In other words, identifying clearly the criteria based on which the sample of observation will be considered defective or non defective.

One measurement criteria is not enough nor sufficient to measure the different aspects when implementing SPC. Accordingly, it is crucial to develop multiple measurement criteria to accommodate the suitable measurement system for each case (Such as in the case when lawyers had a trouble measuring themselves and had a difficulty deciding what to measure.) Delivering the measurement results in an accurate and timely manner. The level of details shared with different managerial levels should be set and identified. For example, senior management should not be bothered by modestly operational processes and the operational level should not be informed with the details of the critical strategic decision making process of the top management.

Now after obtaining the results from SPC and after highlighting the challenges that might face insurance companies when applying SPC, it is recommended that Annette Kluck would consider the following as the next steps in DA’s quality efforts: • • • • Develop a balanced score card that integrates different KPIs reflecting different critical performance areas Create a team for management of change that would lead the change from pure management into management by numbers Assign a task force empowered by the top management to resolve the five challenges she is currently facing Annette used the services of an industry benchmarking firm only to judge the delivered accuracy against competitors but she doubted the correctness of the information. What is more important and highly recommended that she would benchmark the processes. This won’t happen by benchmarking with a competitive company or a company from the same industry. So since, DA depends mainly on customer service which is to a big extent similar to the customer service of banks. Annette could benchmark her process to banks and at the same time become an insurance provider of the bank itself. In the case, it was mentioned that DA had an accuracy level of 99% which make it difficult to identify the discrepancies. So, Annette needs to break the process into smaller more detailed processes to identify more easily where the problem is. Finally, one of the most important recommendations is that, Annette needs to focus more on the customer and what he/she really wants. For example, the team considered that in a form what is more important than the phone number is the address and that if the phone number is wrong or missing it is not a problem. But what if the customer prefers to be contacted by phone or instead of sending letters he would prefer to receive an e-mail. She needs to survey the customers’ preferences, and implement SPC on the processes that fulfill these preferences to improve them to the most.