The company is in the finance and accounting BPO with services ranging from payroll to cash disbursements and claims processing. The data below is the actual tally of the responses we received from our clients. I inherited this from the last six sigma guy who was fired because of his failed attempts in deploying for the last six months.

                     A. Overall        B. TAT      C. Accuracy rate     D. Adherence

                              satisfaction

1 - Very satisfied   4                    5              3                           9 

2 - Satisfied          15                  13             16                         13

3 - Neither            4                    5               4                           2

4 - Dissatisfied      1                    1               1                           0

TOTAL                 24                   24              24                         24

Standard Deviation 0.717            0.776         0.680                     0.624

Acceptable Average 1.717           1.776         1.680                     1.624

1. How do I make sense of this data?

2. How do I proceed in doing statistical analysis based on the likert scale data? The stdev function in excel was used to come up with the standard deviation and the standard deviation is std dev + 1. How do I compute for the correct std dev? Is the concept of the acceptable average wrong? Is there a benchmark acceptable average for the industry?

3. Is there an improvement story or a LSS project that can be derived from the data?

4. What sort of questions should I be asking with regard to likert scale data? What six sigma tools can I use? How can I make a histogram or a control chart for this type of data? Is the DPMO calculation possible for this data?

5. Am I asking the right questions in the first place?

 

Dear Mr. Rellora,

                             I am from manufacturing specilisation so I am new to the terms you are refering to.

What is B TATA? when the accuracy rate is given what the limits considered? What are the limits for adherence?

Have you any customer specified or business specific limits?

What are the targets of company interms of satisfied customers,Very satified customers.

Once you set the targets or goals, verify the raw data of each customer like attending time/response,appropriateness of data. Take those parameters and make your analysis.

The above table is only starting point to decide on how to set targets and go back to the base data for improvements.

Regards

C V rajulu

Thanks for the reply.

A. Overall satisfaction on the services (payroll services, accounting, investment research etc.

B. TAT = Turn Around Time or cycle time

C.Accuracy in terms of no miscalculated payslips, no erroneous reporting etc.

D. Adherence to policies and procedures

The clients were asked to rate each likert item from a scale of 1(very satisfied)-5(dissatisfied) and the dataset I posted is the actual tally sheet. e.g. 4 out 24 respondents are very satisfied overall, 5 out of 24 are very satisfied with how long we process a claim, run a payroll, or prepare inverstement analysis: our Turn Around Time (TAT) and so forth.

regards,

Don

Dear Mr.Rellora,

                            From the explanation and the data, the first step is to find out what is the present Zst value of the process.

The Likert data numbers cannot be taken as values, so one way of analysisng is that the values other than satisfactory and very satisfactory as rejections means we have two conditions. 1.passed and other failed.

By using this we can calculate the process yield>Yrty=+Y of OS*Y of B TATA*Y ofAcc*Yof Adhrnc

=(19/24)(18/24)(19/24)(22/24)=0.430881

Yna= Yrty *exp1/4 = 0.553262

DPMO = 446737  So ZLT=0.1   Z st=1.5+0.1 =1.6

Which means the process is not satisfactory and needs improvement.

You have to set your targets for improvement and look at the base data and analyse the customer data to improve process Zst.

You may use proper tools for analysing the data.