editing disabled mbagroup5 Actions

* Join this WikiJoin this Wiki * Recent ChangesRecent Changes * Manage WikiManage Wiki

Navigation

1. Home 2. an article about a typical Japanese Entrepreneur I like 3. Deutsche Exec Summary

guest · Join · Help · Sign In · Wikispaces
Remove From FavoritesAdd To FavoritesDeutsche Exec SummaryProtected

* pagesubmenu o Details and Tags o Print o Download PDF o Backlinks o Source o Delete o Rename o Redirect o Permissions o Lock * discussion * history * notify me

Details last edit by willzhouchuan willzhouchuan Oct 10, 2006 8:26 pm - 23 revisions hide details
Tags

* none

Protected
COVERPAGE- NAMES, id#s AND NAME OF CASE(Please add your student number after your name/thanks)
Professor: Dr. Giovani J.C. da Silveira
Lodge Karen
Koupal Ondrej
Miller Alana
Ogunsola Oluseun
Will Chuan Zhou (328253)
Name: Case Analysis of Deutsche Allgemeinversicherung
Coverletter?- Addressed to kluck, from consultants
If someone can find a sample cover letter for case analysis then I can make it longer
Dear Mr. Kluck;
We are a 5 member consulting team from University of Calgary, Haykayne Business School and we did a very comprehensive study on Deutsche Allgemeinversicherung's PMV project and its problems, including the problem's background, the competition the company is facing, the firm's current state, Suitability of SPC use in the service industry,
Measurement Issues faced at DAV and specific recommendations, the following is our analysis.

Cheers
Group 4

TABLE OF CONTENTS
SYNOPSIS
Deutsche Synpnosis
Deutsche Allgegeinversicherung, the second largest insurance firm in Germany, had decided to adopt a new quality improvement initiative: PMV in order to maintain its prominent position due to the looming competition and increasingly customer demands.

Kluck, the architect behind Prozessmessung und Verbesserung, was head of Operation Development at Deutsche Allgemeinversicherung(DAV). The PMV project was a revolutionary effort to use manufacturing-style improvement techniques in insurance services, it will differentiate DAV in the industry and hopefully, maintain its prominent position. Kluck, however, was facing a number of difficult problems with the improvement phase of the project.

Priot to 1994, correct transcription from forms had been assured using a method called Double -key entry. There were two problems with this method of assuring accuracy. First, it was very expensive, since it essentially demanded that work be done twice. Second, it was found that first-pass quality actually deteriorated over time when the double-key method was used.

To begin to find out what accuracy levels were like throughout DAV, Kolb and Kluck selected New Policy Set-up as a pilot measurement project. The plan was to take a sample of the work carried out by the associates, and use that sample to infer what the general accuracy rate was in the New Policy Set-up process.

To carry out the experiment,

Kluck decided to use SPC which had traditionally been used for continuous variables such as the diameter of a piston. According the SPC practice in manufacturing in manufacturing, a different kind of chart called a p-chart is used. In such a process, a person would measure five components every few hours, and mark the sample average and range of the measurement on the chart. In the case of New Policy Set-up at DAV, group should sample each other's work at the end of every day.

The first two months of each team's project had been devoted to charting data, so that the natural variance of each of the processes could be estimated. This phrase also allowed people to practice keeping SPC charts. The next phase would involve improving the process.

The improvement gains achived after 12 weeks showed the project was a succes, but the problem is how to sustain the success in the long run.....

PROBLEM IDENTIFICATION
NEED TO MENTION THAT THEY NEED TO COMPETE ON THE QUALITY OF THEIR CUSTOMER SERVICE
NEED TO DEFINE WHAT ‘QUALITY’ means
It is difficult to measure in service industry. Training in standard procedures often used to improve quality. One way to measure quality of service is to use customer satisfaction surveys. Organizations should monitor customer defections to obtain feedback. Maybe mention customer expectations.
-quick response? Reliability? Perceived quality? humanity?
Problem Background
Deutsche Allgemeinversicherung (DAV), meaning ‘German General Insurance’, was founded in 1966. By 1996 it had become one of the world’s largest insurance companies writing nearly 48 billion DM in premiums in over 32 countries. The majority of their business was in Germany (51%) and of that market 60% was in retail insurance dealing with disability income protection and health, property, and life insurance.

In Germany, DAV is one of the giants of industry; in fact they were the second largest firm in the country. DAV’s dominant position was achieved by offering a wide array of insurance products, an excellent group of insurance risk managers, and outstanding customer service. DAV is concerned with keeping its prominent position as smaller insurance companies try to nibble away at their market share.

Define problem
Insurance was becoming more ubiquitous and customers were having a hard time differentiating one firm from another. Of all their core capabilities, DAV’s customer service was seen as the best way of standing out of the insurance crowd. DAV management believed that customer service was a critical element in DAV’s strategy to maintain current customers and attract new ones.

One element of customer service that needed to improve at DAV was the ability to process information and data without mistakes, and the ability to retrieve it in a timely manner. The “Prozessmessung und Verbesserung (PMV), meaning ‘Process Measurement and Improvement’, project was launched and divided into two phases- Measurement and Improvement. Annete Kluck, head of Operations Development, spearheaded this process improvement project by implementing manufacturing-style improvement techniques to the insurance industry.

Inconsistencies in service quality require systematic monitoring to see if they are random, regular, or indicative of a problem.

However would these improvement techniques work, and if so, how will DAV use this imformation to actually improve the performance of the various processes and therefore improve customer service?

Problem background
The maturing of the insurance industry into a commodity and increase in the buyer-power presented a new set of challenges and demanded specific strategic responses in an attempt to differentiate the company’s product in the market place and maintain its competitive advantage. However, an emergence of a stronger competition signaled a potential shift in the industry. This necessitated the improvement in the processes at DAV, and the development of a company wide capability difficult for the competition to imitate. This move will differentiate DAV in the industry and hopefully, maintain its competitive advantage and guarantee its market share in the face of a stronger competition.

In achieving this goal, DAV set out to improve on the quality of its service by reducing variability in its process. Organizations typically move through four quality improvement cost stages:

Prevention cost: At this stage, the organization incurs these costs while preventing variations that leads to loss of quality of its service.
Appraisal cost: This is an assessment cost. Organization incurs this cost while trying to evaluate the current position of its process.
Internal cost of defect: These costs are incurred when organization

tries to correct an error in its process before delivery to its customers.
External cost of defect: These are damage control cost. Organization incurs these costs trying to deal with the consequences of delivery of poor products or customer service

Quality control means organizations must be proactive by constantly appraising their quality control processes with the goal of avoiding both the Internal and External costs of defects by preventing defects and errors in their processes. At the moment, in an attempt to maintain their dominant position in the industry and also to differentiate the company from the competition by developing a company wide capability, DAV is trying to reduce their internal and external costs of defects by incurring more prevention and appraisal cost. A shift in this direction is necessary for the company to achieve its goal.
An analysis of both external and internal factors that explain the current position of the company and the reason for process improvement follows

Competition
Until recently, DAV has enjoyed a dominant position in the industry but a shift in the industry dynamics has seen the emergence of smaller insurance companies capable of challenging the position of DAV. This development has increased the competition for the market share in the insurance industry. The consequence of this is a potential loss of market share and revenue. Improving on their process and developing effective capabilities will equip the company with the necessary tools needed to deal with the rising competition.

STRENGTH
The application

of leading edge technologies to various processes is considered on of the major strengths of the company. It is assumed that the use of these technologies will improve on the response capabilities of DAV to market demands. Furthermore, the company’s broad product base, knowledgeable management team and excellent customer service are also some of the strength the company can build on. An improvement on these processes has the potential of making DAV the largest and most efficient insurance company in Germany. In an industry where the product is becoming more of a commodity, the company can differentiate itself by improving on one or more of these processes.
WEAKNESS
As earlier stated, on the one hand, the improvement on any of the processes mentioned can be a source of competence for the company. On the other hand, a loss of efficiency in any of these departments can also be a reason for concern. The degree of inefficiency in the data entry process which is the reason of several errors is sending a wrong signal to the ever demanding customer. As a consequence, dissatisfied customers have the option of leaving the company.
THREATS
In addition to the external threat of a stronger competition encroaching on the market share of DAV, the over reliance of the company on the abilities of its insurance risk managers also brings about the potential threat of being hollowed out if the managers leave. This brings about the need for developing a company wide process based capabilities as against that based on individuals. Consequently, DAV decided to build on the strength

of its customer service by measuring its current quality with the hope of identifying variations in the process. This will put the company in the position to improve this key process.
KEY CUSTOMER SERVICE ISSUES
An analysis of the company’s current process revealed several factors that potentially inhibit the achievement of desired quality. These factors also threaten to change this capability into core rigidity, working in contrary to, rather than in support of, the company’s quality goal.
TEMPORARY STAFF
The use of temporary staff as a means of cost saving alternative to permanent employees during business downturns could be contributing to the variations in the data entering process. Since these staff is not permanent, it may be hard to get the same attention to details that is required from them. It is also a common practice that companies do not invest as much in training temporary staff making it a bit difficult for the temporary staff to be fully integrated into the procedures of the company.
SAME-DAY PROCESSING
The corporate strategy of same day processing might have an excellent marketing appeal but was a hindrance to the provision of quality customer service. The need for workers to complete these transactions the same day brought about the pressure of working long hours. This will be counterproductive in the long run as tired workers are prone to errors. The work process and numbers of hours worked must be redesigned to ensure workers are not forced to go beyond their capacity. In responding to the customers’ demand and change in competition, a new strategy that will eliminate the long hours and achieve the same goal as the same-day processing idea must be developed.
CUSTOMER AND ASSOCIATE ERROR
Another source of the variation are errors by customers while filling out the application form and employees while entering this information into the data base. Since part of these are completely out of the company’s immediate control, an improvement in the process or a different approach to data collection will be needed. The errors by customers have the potential of passing through the DAV process undetected.

Current state
To improve upon the quality of DAV customer service, Kluck concluded that the area in which DAV was struggling, data input accuracy, needed to be addressed. She felt that unknown causes were creating the existing gaps between the process’s current and desired performance. This problem area was resulting in numerous customer complaints, and external failure rework costs were rising.

Contributing to the already existing human margin for error is a capacity management problem, largely due to the “same-day” mandate imposed by management. Despite ‘state of the art’ technology, those employees responsible for manual tasks, such as new customer and policy data entry often work long hours to get through the day’s volume.

Previous methods used by DAV to improve upon accuracy and quality included ‘Tastenbestatigung’, translated as ‘double-key entry’, and a less laborious alternative, 'sight verification’. The former involved a second associate retyping the transaction in order to cross-check accuracy with the initial entry, and the later method, involved a second associate visually inspecting the transaction. Disadvantages with these methods included excess internal quality failure costs, and the fact that first-pass quality actually deteriorated over time.

Industry benchmarking firms have placed DAV’s first-pass accuracy numbers at 99%, though DAV's management questioned the integrity of their performance data. Kolb thought that it was actually more like 90% for New Policy Setup. DAV needed to get accuracy in their data before proceeding with their improvement strategies. The management team decided to the use the strategies of SPC to measure and ultimately improve its process performance, as it would give the associates the tools necessary to improve upon the quality process themselves.
Kluck and Kolb decided to monitor the New Policy Setup as a pilot measure project of PVM.

Statistical Process Control (SPC), also known as statistical quality control (SQC), holds the basic idea that “it is extremely expensive to ‘inspect’ quality into a company’s outputs and much more efficient and effective to produce them right in the first place.” (Chap 4 text p132). SPC is a tool of the Six-Sigma Improvement Methodology, also known as Total Quality Management (TQM). Though originally developed for quality control in manufacturing, it is applicable to all sorts of repetitive activities in any kind of organization. Thus, TQM effectively switches appraisal emphasis from inspection to process control. SPC is a method for achieving and maintaining quality control. It is a set of methods using statistical tools such as mean, variance, and others to detect whether the process observed is in accordance to standards. The goal of statistical process control is to make a process stable over time and to alert management if the variation is excessive. In the case of Deutsche Allgemeinversicherung, SPC tracks the proportion of applications inputted correctly and incorrectly in each sample in order to measure, analyze and improve the quality of customer service.

The Six-Sigma Improvement methodology insists that there are five phases in quality improvement, and that if completed sequentially, it is likely the firm in question will be successful in improving the quality of a particular process.
Quality control involves 5 phases: (1) Defining the problem and goals for process improvement, and the project scope (2) identifying key process performance metrics, and collecting process and customer data, (3) analyzing and identifying cause and effect relationships of the process, (4) developing solutions to reduce the desired improvement gap, and (5) monitoring the process to control improved performance.

To improve the quality and accuracy of DAV data entry numbers, PVM was launched. The 'Process Measurement and Improvement’ project
There were advantages and disadvantages to implementing PVM. Advantages of PVM included the fact that it called for employees to breakdown their everyday procedures into checklists, allowing them to actually observe the complexity of some of the company's processes.

Employees themselves had to figure out what was important to measure, to measure it, and then figure out if there was a better way of doing things. The power of positive thinking was most definitely an advantage. Drawbacks of PVM included the fact that management commitment was invaluable for implementation, as uncommitted, disinterested management might directly affect the desire of the employees to commit to the project. And employee commitment was crucial. If employees feel that it’s just another ‘fad’, a waste of time, and another way of getting more work out of them, the project is doomed.

The DAV PVM core management team consisted of Annette Kluck, Erwin Weippert, and Hans-Jorg Schoss and Kerstin Kober (consultants), who, over eight weeks, developed the New Policy Setup process as a case study for training. Senior managers were prepped in the SPC background, and video and oral presentations, for first-line and middle managers to introduce SPC to associates, were scripted.
Organizational consultant Kerstin Kober was hired to help with training and any HR issues that might arise. She was very eager and excited to see the lessons learned from SPC be applied to other departments of DAV. Prior to the measurement phase of the pilot project beginning, more than 15 departments wanted to analyze their performances, and were included in the pilot project. Each particular department team, consisting of first-line and middle managers were responsible for getting the associates together and deciding upon the process that was to be measured using SPC, and building a sampling checklist.

Important points of emphasis for the senior managers at DAV included: (1) SPC measures a process, not the people (2) there’s a need to have people use a tool which would become part of their everyday job, so that quality management would also be part of that job, and (3) protect the people and teams from the wrath of the senior managers who, in their surprise of the real numbers, want to ostracize individuals at fault.

Samplers of the New Policy Set-up group were group members themselves. One member from each team was selected at random, on the same day that he/she would be recording the number of incorrect entries made by his/her team.
Data was recorded on a daily basis and P-charts were created and displayed. Senior management was known to survey the charts, and call attention to those departments who’s quality numbers were lowest.

Suitability of SPC use in the service industry
SPC, traditionally used in the manufacturing industry for the measurement of continuous variables, can be tailored to the service industry, but this can prove to be rather challenging. In comparison to SPC application in the manufacturing industry, the service industry is harder to measure. In the case of DAV, SPC implementation for the New Policy setup group wasn't clear-cut. Items that DAV needed to measure were not on a sliding scale- a new policy request was either entered correctly or incorrectly. The ‘Human Factor’, inescapable in the service industry, makes is difficult to determine how to measure, what to measure, what values are acceptable and when a problem actually is a problem. There is significant chance of motivational use of data and there are many sources of variation in this industry, which could possibly lead to dysfunctional measurement and the misuse of SPC. Methods were developed for measuring the quality of a number of process steps at DAV, such as the process for transcribing information from a customer application form into the computer. According to SPC practice in manufacturing, this type of data requires a different sort of control chart, called a p-chart. It is clear that proper planning and implementation was critical for success.

DAV encountered a number of issues in the deployment of the program. If these factors can be overcome, SPC can be made an effective tool for improving DAV productivity, consistency and service reliability, and customer relations.

Measurement Issues faced at DAV and specific recommendations

(1) Better teams do more sampling: Increased sample sizes were becoming a burden to those groups that had a high accuracy rate
Solution:
Keep the initial SPC experiment limited to the New Policy Set-up group.
Once news got out that DAV was running a quality improvement pilot project, 15 other departments, with a range of processes from sorting mail to the resolution of legal problems, expressed their interest in taking part. In the initial stages of a pilot experiment, 15 departments and processes are too many to monitor, as the initial learning curve is exposed to all processes and there is a larger risk of failure and creating a lack of interest for any future improvement ideas.. Kluck and DAV should first apply SPC to the New Policy set up department, and perhaps over time, implement SPC procedures in the other departments.
Streamline Sample sizes.
Each department ought to use its own sample size, depending on the accuracy of the processes sampled. Each department participating in the SPC pilot improvement program should be considered individually. Sample sizes need to be large enough for there to be at least three wrong items in each sample, otherwise there will be nothing to plot. Assuming that the benchmark of 99% accuracy is correct, initial sample sizes of 300 are the minimum requirement for the New Policy Set Up group.

(2) Problems measuring quality and defining problems
Solution:
Define parameters.
Determining when a mistake was not a mistake became tricky, as the parameters for right and wrong were not clearly defined in the beginning. Considering all scenarios of ‘right or wrong’ answers, mistakes should be those that affect the accuracy of important information. What is important and what is not important should be decided prior to the measuring phase. For example, for the New Policy Set-up group, phone number mistakes are not important, where as mis-entered addresses, are mistakes.

(3) Measuring lawyers: Measuring what was good and bad work in a legal department was difficult to define, and the sampling and measurement process was causing frustrations among peers.
Solution:
Use a different measurement tool to determine accuracy.
Initially, only test the New Policy Set-up group

in the pilot experiment. SPC tool of measurement should only be applied to those groups that can clearly define mistakes in their processes. Customer satisfaction surveys per lawyer should be used as a measurement of ‘accuracy’.

(4) Automatic charting: Manual charting was becoming a burden and some groups wanted their results to be charted automatically.
Solution:
Listen to employee concerns
After the pilot “New policy set-up” group experiment has reached the control phase, and when the method is ready to be applied to other groups, DAV should look into the option of automatic charts, to make the process less of a burden for those groups with suited processes.

(5) On the prowl: Sr. managers were reacting negatively to charted data. Employees were becoming leery to the testing process, as they were concerned about remaining employed.
Solution:
Strictly enforce “no ‘punitive’ elements” rule
Sr. management comparing and reacting negatively to SPC data is not conducive to a positive and encouraging environment. The fact that they react this way implies that they did not fully understand the purpose and meaning of SPC data. More time needs to be spent explaining the methodology of SPC, what the charts actually display, what is being measured, and the importance of the managerial roles.

Other Suggestions for improvements to Kluck’s methodology

Relay the Importance of Each Department’s role in the ‘big picture’
It is important to relay the reasons behind the new procedure and well as the importance of each department in reaching DAV’s overall goal of improved customer service. By doing so, people feel included and are given a sense of importance or worth, and are more likely to welcome the new program. Clearly displaying how the particular department’s activities are involved in the process under investigation, such as in the flow chart for the New Policy Set-up in Exhibit 1, results in a better understanding of the overall process as well as displays the importance of the group’s role in achieving improvement in the company’s core competency. This inclusion will also likely reduce the likelihood of employees ‘fudging’ the accuracy numbers, as the selected employees who were measuring the accuracy of groups from which they are members, understand that reporting truthful data is in the best interest of the company as a whole.

Address employee concerns
Rather than believe that negative sentiments about PVC will diminish over time, Kluck and her team ought to hold a question and answer period with employees as well (not just managers). In such a forum, concerns could be addressed and reduced.

Ensure that the program feels ‘finished’. Be able to answer all questions.
Diving into the pilot program, despite the surrounding enthusiasm, is not in the best interest of DAV

Management Team
Have a six sigma black belt on the management team, as they are especially dedicated to quality improvement projects. They will work as mentors, coaches and project champions. Ideal rate of black belts is 1/100 employees. One master black belt is need for every 100 black belts.

Develop a process checklist

for each group.
Managers and departments ought to work together and develop a process checklist for each group.

Specific Recommendations for New Policy Setup Measurement phase: SPC implementation

1. Select the process improvement team: Annette Kluck, Consultants- Schoss and Kober, and cross-level management team.
2. Define the total process flow, outline the critical components that are required in a New Policy Setup.
3. Identify critical process parameters, like what is considered a mistake (phone number vs. address)
4. Select a sample size (300)
5. Select a measurement period. (2-3months?)
6. Construct a process-sampling plan, to ensure consistent, accurate sampling
7. Create a process control and monitoring system, such as an automated charting system that charts the data as it is collected, that all associates can use and understand.
8. Define a corrective decision making process, to help identify the common and special cause issues with appropriate corrective actions. A decision tree could be used here.
9. Continuously monitor and identify any potential improvements for this process.

Once the necessary improvements are made to the measurement phase of the customer improvement program, Kluck and her team will be able to use the information recovered from the improved experiment to identify factors affecting data entry, and develop ways to improve the performance of the processes involved. DAV will use the information recovered as a basis for performance improvement.

Performance Improvement Phase

This second phase of PVM focuses on identifying the causes behind good or poor performance, creating and implementing process improvements, and the continuous monitoring of improvement.

The original teams should be restructured to include cross-functional and cross-level members, dedicated to the enhancement of a process/department. These teams will provide diversification and insight, improving the likelihood that root causes and their solutions will be identified, as well as suggest ways to improve the process. These new teams would be responsible for analyzing the charts and brainstorming to hypothesizing as to why a particular point might fall outside of the normal range, indicating that the process has been affected positively or negatively. These hypotheses should be tested to ensure that the true cause of the problem has been identified. Once the cause has been identified, then possible solutions need to identified and implemented. Implementation of the improvements is dependent on managerial support, for without support, it is likely that the process will not be improved as much as it could be.

The core management team should shift and be a more focused, central steering committee that would conduct additional training of the teams, monitor improvements, and facilitate information sharing between the teams and departments. It is likely that the sharing of information will allow the numerous departments seeking to improve the quality of their services, to be timely and efficient in the successful improvement of their processes.

Continual monitoring of the process for improvement is critical to the success of PVM. By making improvements part of everyday business, DAV will maintain momentum in this process improvement initiative.

Ways to interpret data (Alana)

- consider exhibit 2, as well as create p-chart for exhibit 4 data.
-explain that ‘assuming the measurement phase included the above recommendations’…. We can see that….
-after identifying weeks that data entry accuracy was higher/lower than mean, need to identify reasons WHY.
-what was the difference those weeks?
-should be suggest that groups keep journals chronicling how they felt overall, any big changes/differences?
Once we identify the reasons WHY, then what?

Analyzing and comparing alternatives:
--suggest a standard method for choosing those ‘errors’ to be measured, as each checklist was different, and each department developed its own set of measures.
-need to ask- is what we’re measuring related to customer satisfaction- is it contributing to the errors dealt with by customer service.
-maybe deal only with those departments that are causing these errors.
-is test truly objective? Have peers ‘marking’ peers- result alteration?
-suggest a variation of SPC, and then use the info gathered to implement changes to reduce errors

Error Prevention

1) Bold the most important sections of instructions: this increases the clarity of the forms for the customers, as well as draws the attention of the data input associate while information is being transferred to the appropriate database.
2) Create application forms in different languages so that most customers

can complete the form in their first language.
3) Manually scan application forms for missing information such as address before they are accepted at branches and subsequently sent to VEG in Hamburg
4) Implement an information system that allows customers to input their own information into database. The system will present an error message if information is missing or in the wrong format.

Evaluation of Statistical Process Control

Once the data has been gathered in a methodical method, the data can be used to compare error rates based on different variables.

The variables being compared should be:

- % accuracy of Temporary workers vs % accuracy of full time workers
- % accuracy of morning input vs % accuracy of afternoon input
- Number of breaks during the day vs Rate of errors

To improve on the rate of errors:

1) Create groups within departments that are periodically rotated and compensate those groups with low error rates
2) Implement job rotation so that employees are not required to be accurate all the time.
3) Further open the lines of communication between Kluck, managers and employees so that suggestions can be offered at all levels.
4) Allow employees to choose their own hours. Perhaps some employees will choose to work through the night to clear the applications received late.
5) Present overall project to employees in an understandable format so that employees don’t feel their jobs are in jeopardy.

To improve customer satisfaction once an error has occurred:

1) Offer customers a discount off next year’s insurance

fees

Possible Reasons for inaccurate statistics:
- Employees are rushed when inputting data
- Employees are tired when inputting data
- Employee hours are too long to maintain quality work
- Unmotivated employees make careless errors
- Skepticism of experiment creates errors
- Anonymity may increase errors because there is no accountability
- Employees are not taking enough time during breaks to perform at an optimal level
- Stressful times of the year such as Christmas increase the rate of errors

Implementation and Recommendations:
1) Implement the six sigma system for recording and analyzing data.
Implementation of Six Sigma will create framework to improve and sustain quality and provide a standard measure of process performance.
Six Sigma is a widely used business improvement process that focuses an organization on customer requirements and process alignment. The system operates under the philosophy that it’s better to produce the right item the first time than have to rework.
Five Phases of Six Sigma
1) Define
Clearly specify the problem and goals.
Problem: Customer service needs to improve
Goals:
- To do same day processing of insurance application forms
- To reduce the number of input errors that cause dissatisfaction among customers
2) Measure
Choose process performing metrics that provide insight into how the process is meeting customer requirements.
Deutsche should measure the number of errors in applications because the error rate is directly correlated to customer satisfaction.
INSERT BELL CURVE HERE!!
Process performance

measures:
Six Sigma assumes that customers consider output to be acceptable as long as it is within plus or minus six deviations from target value. This means that the process mean can shift by as much as 1.5 standard deviations and the process itself will produce at most 3.4 defects per million opportunities.
Three drivers of Process Sigma:
a) Actual customer requirements
b) Variation in the process as measured by the process standard deviation
c) Stability of the process as measured by how much it can shift over time

3) Analyze
Goal: To identify cause and effect relationships that can be leveraged to improve performance
Possible causes for errors:
- Employees are rushed when inputting data
- Employees are tired when inputting data
- Employee hours are too long to maintain quality work
- Unmotivated employees make careless errors
- Skepticism of experiment creates errors
- Anonymity may increase errors because there is no accountability
- Employees are not taking enough time during breaks to perform at an optimal level
- Stressful times of the year such as Christmas increase the rate of errors

Identify relationships between variables
- Temporary workers vs Full time workers
- Morning input vs Afternoon input
- Number of breaks during the day vs Error Rate
- Length of time taken during break vs Error Rate

INSERT FISH BONE DIAGRAM HERE!

4) Improve
The Improvement phase is challenging because variables analyzed difficult to control. One challenge is that some variables are not able to be held steady to evaluate another. Another challenge is the fact that interaction and joint variation between variables is difficult to measure.
DIAGRAM (Pg 165) in text to illustrate
One solution to the difficulties is to use design of experiments techniques (DOE). DOE uses statistics to design experiments to investigate multiple variables simultaneously.
Interviewing employees may identify relevant factors and possible correlations.

5) Control
Develop procedures to monitor the process. The goal of this phase is to ensure the process’s higher level performance is maintained.
The p-chart is used to control the process system. The p-chart will distinguish between chance variation (values between UCL and LCL) and assignable variation (values outside UCL or LCL).

One of the benefits of using the p-chart is the ability to use different sample sizes. Since the calculated values are based on proportions, any representative sample size can be used.
P-Chart

A p-chart is a statistical tool to monitor a process and track the proportion of nonconforming items in subgroups of varying sizes. By observing the proportion of nonconforming items, the p-chart can be used to determine whether the process is stable and predictable and can access the effects of process improvements.

Creating the P-Chart

In order to draw conclusions regarding a process variable, control data must be gathered to determine long term behavior. From the long term data the average (u) and the standard deviation ( ) should be calculated.

Center Line:

U- The long-term average value of the process variable (Number of correct applications)

Standard Deviation:
(Std Dev)2 = (x1 – U)2 + (x2 – U)2 + (x3 – U)2+ …
N- 1
Upper Limit Mean:

ULM = U + 3 (Std Dev)

Lower Limit Mean:

LLM = U – 3 (Std Dev)

The upper and lower limits mark the natural variation in the process. The items outside the upper or lower limits are considered nonconforming and those inside
Once the long term data is collected the sample can be graphed in a p-chart.

The Plotted Statistic

The percent of items in the sample meeting the criteria of interest.

P = (Number of Correct Applications)/ number of items in sample

One of the benefits of using the p-chart is the ability to use different sample sizes. Since the calculated values are based on proportions, any representative sample size can be used.

Sample Calculations:
Center Line (Average Number of Correct Applications in Sample):
U = (total number in sample) - (Number of defects in Sample)
Total number of units Sampled
Sample Calculation from Data in Exhibit 4 p.12
U = 9000 - 516
9000
U = .94266
Standard Deviation:
(Std Dev)2 = (x1 – U)2 + (x2 – U)2 + (x3 – U)2+ …
N- 1
Where N is the total number of samples
(Std Dev)2 = (.94 - .942667)2 + (.95 - .942667)2 + (.94 - .942667)2 + …
30 – 1
(Std Dev)2 = .000575019
Std Dev = .023979557
Upper Limit Mean
ULM = U + 3(Std Dev)
ULM = .942667 + 3(.023979557)
ULM = 1.0669
Lower Limit Mean
LLM = U – 3(Std Dev)
LLM = .942667 + 3(.023979557)
LLM = .87072
The Plotted Statistic
P = (Number of Correct Applications)/ number of items in sample

Week 1:
P = 282/300
P

= .94
Long Term Use of the P-Chart - Control

In order for the evaluation of quality to have value in the long term, p-charting should be performed continuously. As the data is inputted into the p-chart managers should be ensuring that the proportion of correct applications is between the upper and lower limit. If a value is outside those margins, the source of the error should be identified and eliminated if possible.

Analysis of P- Chart

Analyzing a p-chart requires observation of the proportion of correct applications relative to the upper and lower control limits. The process being analyzed is functioning well when the proportion of correct applications is within the outer boundaries. Otherwise the system is out of control and in need of analysis.

In the case of the data from Exhibit 4 on page 12, the upper control limit is above one so the UCL is impossible to exceed. The control limit that is of concern in this project is the lower control limit which is set at .087. Once the proportion value drops below .087, customer service has dropped below a satisfactory level.

Week number 24 was a week of concern for Deutsche because the proportion of correct applications was .8466667. At this point the managers should have evaluated the application inputting process to determine the reason for the drop in performance and made the appropriate improvements.

All the other weeks of samples fell between the upper and lower control limits which means that the input process was successfully operating.

2) Take steps in Error Prevention

1) Bold the

most important sections of instructions so that directions are clear.
2) Create application forms in different languages so that most customers can complete the form in their first language.
3) Manually scan application forms for missing information such as address before they are accepted at branches and subsequently sent to VEG in Hamburg
4) Implement an information system that allows customers to input their own information into database. The system will present an error message if information is missing or in the wrong format.

3) Improve on the rate of errors:

1) Create groups within departments that are periodically rotated and compensate groups with low error rates
2) Implement job rotation so that employees are not required to be accurate all the time.
3) Further open the lines of communication between Kluck, managers and employees so that suggestions can be offered at all levels.
4) Allow employees to choose their own hours. Perhaps some employees will choose to work through the night to clear the applications received late.
5) Present overall project to employees in an understandable format so that employees don’t feel their jobs are in jeopardy.

4) Hire a Six Sigma “Black Belt”
A black belt is a highly trained individual who has experience and talent when dealing with quality improvement projects. The black belt will make suggestions to improve the five phases of Six Sigma as well as aid with the logistics of the project.

To improve customer satisfaction once an error has occurred:

-Offer customers a discount off next year’s

insurance fees

According to industry benchmarking, Deustche’s first-time accuracy numbers were quoted at 99%. Ninety nine percent is a number that came from comparing Deustche’s proportion of delivered accuracy with the best practices to be found in the services industry.

Ninety nine percent was the assumed accuracy rate that determined that 300 samples were necessary in order record at least three errors in the first 30 weeks of data gathering.

Number of Errors= Number of Samples (Error Rate)
3 = Number of Samples (1-.99)

Number of Samples = 300

In reality, from the data shown in Exhibit 4 (page 12) of the case, the average proportion of correct applications during the first thirty weeks was .942667 and after the diagnostic period was .939259. Therefore the error rate after the diagnostic period was
1- .939259 = .060741

In accordance with Schoss’ request to have at least three errors per sample we can calculate the number of samples required:

Number of Errors = Number of Samples ( Error Rate)
3 = Number of Samples (.060741)
Number of Samples = 49

As a result of requiring only 49 samples the burden of sampling is significantly reduced which may cause increase job satisfaction amongst employees.

Recommendations and implementations to improve customer service
The above analysis showed how SPC processes and techniques can be used in the DAV environment to gauge customer service. Having been armed with this statistical information, how can we now begin improving performance of customer service?

First step it to baseline the

current state of customer service. A year’s worth of data can than be used in subsequent years to benchmark improvements, in this case less errors. For example, after imposing a change in customer service process, the YTD errors are improved, than this is quantifiable evidence that the change has worked.

Once you have this baseline, you can analyze the data for trends that might indicate potential improvements. Abnormal spikes in errors around a certain date might provide evidence of a process improvement opportunity. For example, if a certain error spike occurs each year at the same date, and it happens to coincide with a German soccer finals event, it might indicate that your staff might be preoccupied with an external societal event outside of the company’s control. Some options to consider are giving the employees a mandatory day off, or giving employees the option to take a ‘no questions asked’ day off, leaving only employees with little interest in soccer to work undistracted.

Moreover this same information can be used to monitor spikes throughout the day and not just the year. You will than be able to monitor whether more errors happen during the morning or afternoon. If employees are more prone to make mistakes in the morning than the work day may be staggered so that workday occurs when the employees are least likely to make errors.

DAV can also look at the error rate of full time employees versus part-time employees and their work shift length. Does the error rate go up as more part time employees are utilized? Do longer shifts coincide with more errors? Or is it the opposite effect? These types of trends can than provide DAV evidence with which to customize their workforce mix and capacity management. For this type of data entry work, having more flexible part time employees working 4 hour shifts might actually be more beneficial than the traditional full time, 8 hour a day, workforce.

Again, this type of analysis can only work once DAV has this baseline from which to compare future customer service modifications. Therefore, the quantifiable evidence for some improvements might take two years (as in the case of the German soccer finals distraction) before it can be determined that a change has actually improved service. Management must be fully aware that this type of process improvement needs sufficient time to gathering, evaluation, process change and than evaluation again. For this type of quality control to work patience must be fully communicated to senior management and full buy in must be solidified before time is invested in this process. Management must know that this is no quick fix but in the end they will be rewarded with quantifiable evidence for this soft science called ‘customer service’.

help on how to format text

About · Blog · Pricing · Privacy · Terms · Support · Upgrade
Contributions to http://mbagroup5.wikispaces.com are licensed under a Creative Commons Attribution Share-Alike 2.5 License. Creative Commons Attribution Share-Alike 2.5 License
Portions not contributed by visitors are Copyright 2011 Tangient LLC.
Home > ...
Loading...
Home Turn Off "Get