WISHING  EVERYONE IN THIS FORUM A VERY HAPPY NEW YEAR 2011  

Hi All,

The below is a new situation :

I am working a GB project and here is the situtation .The project is to reduce average closure time for Prio1 and Prio2 Tickets received from the customer end (Prio1 and Prio2 are the severity which customer assigned it to the tickets after the product is delived ).

Here is what i have followed  and please request all to confirm that the below steps followed is correct or not :

1.Samples collected seperately for Prio1  and Prio2 .

2.For each ticket there will be a start date  and end date for the ticket .That is when you got a ticket  and when u closed the ticket

3.So for Prio1 and Prio2 ,  for each sample , calcluated the difference (end date -start date ).

4.Here by i have big list of prio1 with closure days(end date-start date) and similarly for prio2

Lets take only prio1 for time being .

5.I executed noramlity test for 60 samples under prio1 and found that the days of closure(start date-end date) is NOT normal .

6.Since this is PASS/FAIL opportunity kind of sample .Either you solved it in define period or  you dont .So its a PASS or FAIL situation .

7.Used the Sigx Sigma Calculatior to calculate DPMO as below :

Total Size = 60

No of Defects =45 (for tickets which could not be resolved in prescribed time under prio1)

No of Opportunite =1 (Since either pass or fail )

I got DPMPO through this  and Sigma Level respectively

8.Hence i got Sigma Level for  both prio1 and prio2

PLease let me know if the above steps follwed by me are OK or not ,Are there any deviations.

Now , i am planning to draw the existing process in the form of flowchart to understand the process flaws in it .

Let me know if my direction is right or not .Your response will be highly appreciated .

 

Hello Pankaj, 

In your situation, the whole unit is taken either as PASS or a FAIL. This count of FAILED units should be taken as defectives and you can use the yield method for Sigma Level calculation. 

Regards, 

VK

 

Thanks Vishwadeep for replying .

Yes , its a Pass /Fail situation .Hence all those that does not meet SLA are fail  and are defects .However, i have use the the below to get DPMO and Sigma Level :

For Discrete data :

No of Total Units :

No of Defects :

No of Opportunites :

Got DPMO  and Sigma Level .

Is this correct ?

I got Sigma Level as 1.3 .I understadn that it is verly low but does it mean that there is no process in resolving the tickets .Is the process is  adhoc ?

Dear Pankaj,

• You should use the Yield method as you are working with defectives. 45 defectives out of 60 means that the yield % is 15/60 = 25%. Using sigma level calculator, you can get sigma level.
  
• If you get a low sigma level, you cannot conclude that process does not exist. There could be other reasons for low sigma level. For example, it is possible that the prescribed time is impractical.
  
• If you are able to find reasonable opportunities for improvement, it is a good project related to VoC.
  

Hi Vishwadeep ,

I have a reason to debate on your answer and the thing is the approach i have used i.e No of Units,No of Defect,No of Opportunities gives me the same sigma level as when used yield method to calculate sigma level .

1.

My approach:

Here are the data :

No of Units :82

No of Defcts :45

No of Opp :1

gives me "Sigma Level = 1.62"

2:

YEILD approach:

I plugged in the yield percenatge (i.e 45 out of 82 i.e 54.87%) in the YEILD section of Sigma Level Calculator gives me too 1.62.

Hence, the question is are they really different .If yes , are there any concrete examples ?

Dear Pankaj,

Let us consider an example. Each unit of an item has 50 opportunities for defect. 10 units are inspected. There are 5 defects found in a single unit out of 10 inspected items. 

DPMO = 5X1000000/10X50 = 10000

Yield uses number of defect free units. Yield = 9/10 = 90%

If you calculate sigma level, the two provide different values in this example. 

They can give same results sometimes (as in your example).

This happens in those cases where opportunities for error is equal to ONE per unit. (This leads to defect count and defective count becoming equal)

Best Regards, 

VK

Many thanks vishadeep , now the point is clear to me 

Now the thing is that i do not have measurement system in place .The data is generally taken from a tool and dump is taken to capture the start date of the ticket assigned and the close date when it is closed.Hence, i am taking dump to create this system in excel.The only system is the customer tool where tickets are assigned .The worklogbook are manually read to understand when the ticket was assigned as there are no date assigned fields to capture the data .However, i would say there is a manual interpreataiton of the assigned date and close date.

The thing here i want to ask is why do we measurement analysis ? In case if i use such manual system, will Measurement and Analysis would work ?

I read from "Six Sigma in Software Development " from Chrstine Taylor as below :

"Although the goal is to have all measurements to be repeatable and reproducible ,this goal is more difficult is to achieve with subjective data such as customer satisfaction ot pass/fail situation that can vary depending upon the rater or insepctor "

Hence what should be my approach as i am working pass/fail situations...??????

Isn't MSA a must in a situation like this particularly. Are there service desk personnel mentioning the Assigned/Closed Date along with the final closure status - Pass/Fail?

So in case its a manual system we must have a calibration mechanism.

 

Ya i too  have a opiniopn that MSA  is must , i have seen  that in case if you are taking data from a customer tool ,MSA implemenation is left to the customer .Hence i have seen some GB project presenation where MSA is not done saying that the data colected from the customer tool .

No there are no service desk personnel involved in the entire process .the worklog book is the history where you get when actually it was assigned   and there will be status attached to it when assigned .

I am not sure how can we have callibration mechnaism .What are the approaches is beging used in callibration ?PLease let me know if you have any idea on the same 

How exactly is the data entered in the worklog book? I guess what you are saying is that the customer is entering the assigned/closed date along with the final status of pass/fail, is that right?

As when the ticket is assigned to a person ,as when his name is selected from the list the worklog book is updated with date  and status as "Assigned to" .Now the person works on it  and update the worklog book with the findings .Once he resolved he change the status to "closed".The work log book is again updated as when the status is chnaged to closed .

 

But all this information is available in worklog book.There are no sepearte fields to capture this .Since worklog book being a text file , you have manually see  and take those date to external file may be excel .

the data came out to be non-normal.

1.i used "individul distributtion identification" .

2.among all distribution "lognormal" distribution had greater p value .fr others it was p value less than 0.05

3.i used capability analysis fr non-normal and used selected lognormal there

4.got zbench as -0.27

 

its negative,how do i interepret this . i havent done BB so not sure how to handle non-normal data

Dear Pankaj,

 

What you need to do in this case is "logical validation" which means that you need to ensure that the data you are capturing from the system has the same start and end points as defined on your Operational definition.

 

MSA needs to be done only when there is manual assessment involved. You are using the start and end time from when the ticket is logged to when it is closed in the system. You will still need to ensure that the person who works on it closes the status real time and not as per their convenience.

 

In TAT projects, it is typically seen that there can be a few outlier cases. If you remove the outlier cases, you could do a normality check again and use methods and tools that you would use for normal continuous data. But you should go this way only if your metric is TAT in number of days or hours.

 

If your metric definition is in pass/fail, you should calculate the sigma level as per the defectives method.