Pallab
@PallabMember since October 15, 2008
was active Not recently activeForum Replies Created
Forum Replies Created

AuthorPosts

April 15, 2009 at 7:41 am #59587
Hello RBK,
Since you are in finance and accounts field, let me give you some quick examples of meaningful green belt project in your own space. Standard DMAIC methodology to be followed for green belt project. Google search will give you loads of training materials on DMAIC.
1. Cycle time time reduction and defect reduction for Account Payables : starting from claim submission by employee or vendor, till receipt of cheques. This is a very popular and much troubled area to focus and improve. Kind of low hanging fruit.
2. Monthly payroll defect reduction and improve timeliness : Capture cases where pay outs were incorrect or late. Do root cause and improve. This is a very critical and sensitive area for any firm.
Similarly you can now think of a big list of project opportunities in finance and accounting space.
Best of luck. Share your ideas with me also. I have done both of these projects myself as Master Black Belt.
Regards.
Pallab.
ASQ certified CSSBB, CMQ/OE.0April 15, 2009 at 7:31 am #59586Hello Mike,
A lot of global banks and investment banks are practicing Lean sixsigma for many years, in fact they are saving in millions. I would like to make three points here that may help you :
1. Operation is the best place in financial services companies for lean sixsigma, for other departments it may prove to be a fad. I have myself achieved millions of $ savings for investment bank simply through automtion and waste elimination.
2. In financial services, operations people dont prefer statistics for continuous improvement. So a light weight approach heavy on lean and light on sixsigma would help. This has helped tremendously. Also, here the ‘non standard/brain work’ is more common as compared to manufacturing and hence you would surely require less statistical data analysis tools (DOE, hypothesis testing etc).
3. Go for online courses that does not cost you much to start with, so that you judge whether this is your cup of tea or not without risking much, hence going for black belt directly I would not suggest. I have seen many banking operations jobs requesting for a green belt awareness. The continuous improvement and the problem solving skill is the focus and not the statistics. In fact, my black belt knowledge is rarely used to what I do as Master Black Belt in my company.
It is the mindset that matters, seeing the opportunity and asking the right questions.
Best of luck.
Pallab.
ASQ certified CSSBB, CMQ/OE.
0October 15, 2008 at 4:39 am #59463Hello Gregory,
I would like to take part in this discussion as well. I am working for a large investment bank and heading the Lean sixsigma program. Would like to share my experience and learn from your experience on how to deploy LSS and what type of projects are generally applicable in investment banking space.
It would be great if you could share one or two case studies with us.
Thanks and regards.
Pallab.
0July 30, 2008 at 5:30 am #65133Venkat,
I suggest you look at the basic core definition of sigma instead of looking at at what DPMO level sigma start going negative, thats not the right approach.
The basic definition of sigma level of a process is very statistical, it is :
If you have single spec limit :
‘The number of standard deviation that a process can accomodate between the process mean and nearest customer spec limit.’ (USLMu)/SD or (Mu – LSL)/SD. Now simply, when on an average(mean) your process has gone above the upper spec limit or come below the lower spec limit(which ever is applicable to your process). This is the area left of USL or right of LSL.
Both spec limits are available :
Then we calculate the Z bench : that is we calculate both are areas and add them up.
Z=0 essentially means, the yield <= 50%, it cant be more.
Also sometime you will see on articles or case studies, that mean is more than USL (or vice versa), still they are reporting a positive sigma level, dont be confused, in that case, they have reported short term sigma which long term + 1.5 (generally).
Note, all your sample data/calculation always give you long term sigma value.
Hope this helps.
Pallab.0May 25, 2004 at 9:02 am #100716Kris,
As Gabriel rightly explained, there are two behavioral parameter of a process.
1. Stability(caused by process variation or consistency issues and measured by control chart)
2. Capability(response location with respect to the specification limits, caused by process centering issues or accuracy issues)
Now the biggest question is which one to get priority over the other ?! The answer is “stability”. First if the process is not stable, the corrective measure will have to be taken through Root cause analysis or Six sigma data analysis. This is the time to handle special causes. Rigorous statistical data analysis comes with solution which may 1. shift mean and 2. reduce the variation. As this stage, the process improvement goal is to reduce considerably standard deviation of the process with small or zero improvements in process sigma(capability), as the process is not delivering at all within spec limits. Here, you may use DoE or Taguchi’s orthogonal parameter design to focus on outliers and its root causes.
2. Stege 2 is after successfully achieving stability, you have a process at hand which listens to your instruction, it only varies with random factors and has become consistent. This is the best time to push the process towards spec limits and make the process capable. Here you may look for technology change, e.g. One web portal was failing to meet customer performance CTQs, you had already undergone exercise to make the page load time consistent by reducing number of database calls, using less overhead on server more on client and fine tuning the querries. Though, after having these improvement solutions, mean response time is 10 sec and very consistent around 10 secs. Where your spec limit is <= 5 sec. One of the innovative solution using TRIZ (theory of inventive problem solving) is, shift the architecture of the portal from ASP to JSP with EJB(from microsoft to J2EE using enterprise Java beans). As a result, the mean response time came down to 4 sec. Now you can start drawing control chart to check the "capable" process is stable or not.
Note : Very ofthen this happen that, a mean shift brings in “free”(unwanted) inconsistency in the process. So, we need to be aware.
Hope this helps.
Regards.
Pallab, TCS MBB0March 4, 2004 at 3:19 pm #96403Andy,
yes I have implemented this method already. Can you please elaborate your question a bit further ?
Regards.
Pallab.
0March 4, 2004 at 3:15 pm #96401March 4, 2004 at 2:23 pm #96391Mannu,
I am late to reply. I understand what you are saying, but be more specific to my case. I have selected MTBF as Y measure. You can not break this any further. For such kind of a measure we have to wait for atleast one cycle(6 month) to prove the improvement. What you say ?0February 15, 2004 at 12:16 pm #95518Hi,
Just trying to elaborate a bit more. We really cant say that with only DPU at hand we cant calculate process Sigma(Z). Yes we can if and only if both of the following conditions hold true :
1. There are a considerable number of defect opportunities per unit.
2. Defect opportunities are same for all Units in the sample. All your units are of same category, as in your example. If you are only looking at Bills of same category having a constant number of Defect opportunities.
If both of them are true, you do not need to have DPO for Z calculation. HOW, here is how :
The defect distrubution may be approximated as a Posson distribution with a average defect probability which is Mean of the distribution which is nothing but = Lamda=DPU.
Now, P(X~ B)= exp(Lamda) * (Lamda) ** X / factorial(X).
Where, X is the defect distribution of the project Y.
and P(X) is the probability of X defects.
Hence, if we say, Yield= Probability of Zero defects=P(0)
P(0)= exp(Lamda) * (Lamda) ** 0 / Fact(0) = exp(Lamda)
=exp(DPU).
Now, from Z table, we find process sigma value.
But the application of the above case is rare, generally this happens in manufacturing process(assembly lines) like Pumps, Engines etc where Opportunities per unit is very very large.
Hope this helps to establish the importance of DPU.
Regards.
Pallab B.
MBB, TCS.
0February 15, 2004 at 12:16 pm #95517Hi,
Just trying to elaborate a bit more. We really cant say that with only DPU at hand we cant calculate process Sigma(Z). Yes we can if and only if both of the following conditions hold true :
1. There are a considerable number of defect opportunities per unit.
2. Defect opportunities are same for all Units in the sample. All your units are of same category, as in your example. If you are only looking at Bills of same category having a constant number of Defect opportunities.
If both of them are true, you do not need to have DPO for Z calculation. HOW, here is how :
The defect distrubution may be approximated as a Posson distribution with a average defect probability which is Mean of the distribution which is nothing but = Lamda=DPU.
Now, P(X~ B)= exp(Lamda) * (Lamda) ** X / factorial(X).
Where, X is the defect distribution of the project Y.
and P(X) is the probability of X defects.
Hence, if we say, Yield= Probability of Zero defects=P(0)
P(0)= exp(Lamda) * (Lamda) ** 0 / Fact(0) = exp(Lamda)
=exp(DPU).
Now, from Z table, we find process sigma value.
But the application of the above case is rare, generally this happens in manufacturing process(assembly lines) like Pumps, Engines etc where Opportunities per unit is very very large.
Hope this helps to establish the importance of DPU.
Regards.
Pallab B.
MBB, TCS.
0December 27, 2003 at 9:18 am #93848Sathu,
I do not think you got my question. We need to improve the current application by reducing the frequency of outages. Generally for various reasons it goes down atleast once each month and this has been happening since last one year.
This improvement opportunity has been taken up as a GB project with DMAIC rigor.
Thx.
Pallab.0September 8, 2003 at 5:04 am #89648PCB : Process Capability Baselining.
0 
AuthorPosts