[racer_7]

Hi. The company for which I work for uses six sigma in each and everything. Atleast they think so. Cannot disclose the name but its a 30bn USD American manufacturing company and their six sigma certification apparently holds a lot of value in the market.

As mentioned by others, you have DMAIC and DFSS. We use DMAIC mainly for sales and marketing process improvements and DFSS for manufacturing productivity improvements. DFSS is more about quantifiable goals and DMAIC (my thoughts) is more of theoretical improvements.

Now, coming to my main point, in our organization to grow, one MUST go through six sigma training, do a six sigma project(atleast to get a green belt certificate given by the company), be a trainer in six sigma team of the company training other employees for a year or so and finally get your black belt. One needs to go through this before they head a particular division.

My thoughts are that six sigma works reasonably well in manufacturing if you have others who also believe in the same. Else, it is just one guy running after others to get hold of data to fill up the tools and finally ending up manipulating the same

[fine69]

Originally Posted by racer_7

We use DMAIC mainly for sales and marketing process improvements and DFSS for manufacturing productivity improvements. DFSS is more about quantifiable goals and DMAIC (my thoughts) is more of theoretical improvements.

DMAIC (Define-Measure-Analyze-Improve-Control) is applied to existing processes whereas DFSS aka DMADV (Define-Measure-Analyze-Design-Verify) is for new product creation.

Objective of both the approaches is to meet the customer requirement hence it would not be right to say that one quantifies the goal where as the other doesn't. In fact, the data available in a DMAIC project is generally found to be diverse as compared to DMADV projects, simply because an existing process is bound to churn more data than the one that is yet to be created.

DMADV might be applied to existing processes also but that is when DMAIC has failed to meet the customer specifications, basically you go in to a re-engineering mode.

I've tried to show the comparison between the different stages of the two approaches. The Define stage in both is all about objectively stating the goal and customer specs. Of course all stages have a lot more to offer but I've tried to put it in a very simple manner.

[Coupé]

Originally Posted by Samurai

This is why I keep away from process or methodology wars. I just use common sense processes that have worked for 24 years of my career.

Couldn't agree more. Common sense is the most useful of all improvement tools. My first lesson in Six Sigma was application of common sense while using statistical methods. It started with a thought provoking statistic: After measuring the number of people with leg amputations and those with leg deformities the average number legs per human came out to be less than 2!

Originally Posted by fine69

To people seeking certification, my advise would be to go get it from anywhere because ultimately knowledge & experience would talk more than the certificate. I've interviewed 100+ Black Belts till date and have seen nincompoops (w.r.t. to Six Sigma knowledge) from the best of organizations & best of institutes while at the same time some very knowledgeable people from roadside institutes. People who've learnt and worked hard on the subject would always stand out, I firmly believe in this.

You said it right. To be the best of breed one's emphasis should always be on practical knowledge gained by solving real-world problems.

Originally Posted by fine69

As a Black Belt consultant let me openly say that 50% (I'm being very liberal here) of the Six Sigma DMAIC projects are actually a farce because that situation isn't the one where both cause and solution to the problem are unknown.

A master black belt worth his salt should disapprove Six Sigma projects where both cause and solution are already well-known. And where the projects are planned to give the appearance of a structured problem solving in favour of a pre-determined action. Someone mentioned earlier that there are no full-time black belts in his organization because it is supposedly everybody's job to partake in improvement initiatives. Every orchestra needs a conductor.

Originally Posted by racer_7

Now, coming to my main point, in our organization to grow, one MUST go through six sigma training, do a six sigma project (atleast to get a green belt certificate given by the company), be a trainer in six sigma team of the company training other employees for a year or so and finally get your black belt. One needs to go through this before they head a particular division.

My thoughts are that six sigma works reasonably well in manufacturing if you have others who also believe in the same. Else, it is just one guy running after others to get hold of data to fill up the tools and finally ending up manipulating the same

This is probably done so that those who learn to manage by facts and data are given the top jobs. Six Sigma does not work reasonably well in manufacturing, it works extremely well. Statistical measurements like MTBF enable companies to indicate how "reliable" their products or parts are and provide guarantees that they will last the distance: for example a car or it's engine or even it's belts. Six Sigma is what helps them ensure that this guarantee can be extended to every car they manufacture. In the services sector, techniques of lean are increasingly being combined with six sigma to reduce waste along with variability with outstanding results and releasing hidden value. These techniques are used by courier companies to deliver packages on time, every time. By banks and financial services organizations for clearing of funds within a defined time frame every time. By airports to ensure that flights take off and land like clockwork every day. The list is endless.

Coupé

[Trojan]

Originally Posted by Samurai

Six sigma was created for manufacturing industry where automation is a reality. The quality and quantity of input, the output, and the time taken at each step are all measurable.

Agreed, it started with manufacturing and even today it works 'best' with them. When six sigma later started to be applied to IT & ITES, the closest and best application of the same was first with BPOs, again the key word being 'Measurable'.

IT on the whole was more about standards and capability, as measurable's though available were a big variable by itself and had many external factors involved. Arriving to a clear conclusion becomes all the more difficult and is sometimes extremely tricky

As you have rightly said, this industry is dependent largely on resources and their skills. There is no best way to measure it in an absolute manner.
What is possible is to help improve their current skills or add a few more. But focus should not be on skill-sets alone, other factors need to be taken into account as well. Employee motivation has a huge impact here and having a multidimensional approach has far reaching results.

As I said earlier, continuous improvement is the DNA for all successful organizations.

Coming to user requirement, which is a big variable and the other aspects within being
- Understanding of the requirements
- Incomplete requirements
- Changing requirement

How do we tackle the above
- Who is evaluating the requirement
- Are the requirements complete, what questions do we have on the same to better our understanding
- How do we freeze the requirements? Any changes which come in later, how do we incorporate those

This being the input for work, is extremely critical as mistakes at this level will resurface by all accounts

Customers expectations will always be high, but then they should be willing enough to pay. There is a certain cost for Quality and each customer knows what they are paying for.
In case the customer expects bleeding edge technology, we should draft out our willingness to provide the same (albeit at a cost). Most of the time the customer will retract and stay with the previous, due to the grey area of Support.

We all like to window shop and always prefer to get something 'more' for what we pay. Understanding this aspect of 'more' in our current customers is the key.

More recently we included two of our larger contracts for a zero incident programme. We carried it out and presented the resultant improvements to the customer, they were surprised and yet happy with the same.
Being a new account (just 6 months old at the time of project start), the key was to strengthen the relations. It has made an impact is by getting us more business.

Now I will just give an insight on what I do
We lead a lean transformation, focusing on all areas applicable (Stakeholders, Performance Mgmt, Training & Skills, Employee Mindset, Tools and Automation, Process Redesign, Utilization, Analysis of Defects and their Reduction, Feedback from Customer).
This is keeping in account the central item which is Customer Requirement and Satisfaction
For all of the above factors the slightest issues faced in any of them are listed down and its impact calculated.
Once all the issues and their impacts are available, thorough root cause analysis is carried out and solutions discussed and agreed.

Below are a few of the objectives of a lean wave
- Improved Utilization
- Reduced defects and thus less rework
- Improving employee motivation levels
- Leaner processes resulting in faster processing and improved SLAs (while 90% meets the Service Level Agreement, 95% is better)
- Continuous Training for Employees thus enhancing and sustaining skill-sets
- Robust performance management system
- Improved Customer Satisfaction
- Aligning of goals and objectives with the customers

Implementation of the above not only results in strengthening the relations with our customers, but also has a positive effect on the bottom line.

The entire process is a thorough team effort and requires everybody's involvement. What I have elaborated above is done specifically for System Integration projects and not Managed Services. We have contracts of all types and customized solutions need to be created and achieved.

The idea is not to create more processes, but how can we work on the existing processes in a more efficient manner without increasing any further burden on the resources.

Cheers!
Trojan

[Samurai]

Nice to see all the experts drop the jargon and speak plainly. Finally, I can make sense of what you all are saying.

Originally Posted by fine69

Yes, quality of a painting can be measured, just that your perspective of quality is probably different from mine. Let me make an attempt at explaining something.

Firstly, there is always a customer to a business. The customer decides what he wants, so lets say I'm the customer to this painting. Now I like a Monet better just because I like the sound of the artist's name. I don't care what he made till the time its a genuine one. I'm ready to pay top dollars for it so does it really matter to you (you're the painting seller) whether Picasso is better or not. The customer decides the quality in a business. It doesn't matter what the art expert says, you as the supplier would want to give the customer what he demands.

Hmm, in product development, there are lots of areas to which the customer has no visibility. For example, 90% of what we do runs on the servers. The customer only gets to see web reports, and that is where they tend to focus. The remaining 90% is technical gobbledygook, and is ignored by customers as long as it works within nominal parameters. So it is up to me to judge the quality of the 90% using internal measurement process, which relies on subjective judgment. Actually, it is a mix. Recall my 50 criteria scoring sheet, the score for each criteria is subjective, but the grand total gives me a meaningful number, so it becomes measurable.

Originally Posted by Trojan

Once all the issues and their impacts are available, thorough root cause analysis is carried out and solutions discussed and agreed.

Below are a few of the objectives of a lean wave
- Improved Utilization
- Reduced defects and thus less rework
- Improving employee motivation levels
- Leaner processes resulting in faster processing and improved SLAs (while 90% meets the Service Level Agreement, 95% is better)
- Continuous Training for Employees thus enhancing and sustaining skill-sets
- Robust performance management system
- Improved Customer Satisfaction
- Aligning of goals and objectives with the customers

Damn! Now we are talking the same language. We do the same things, except we have not formalized the process to this extent. People are trained to do these things by instinct and habit. We try to instill common sense instead of forcing people to follow processes like a drone. We use simple language and avoid unnecessary jargon. But I often wonder whether this can be scaled to a larger team, probably not. That is where you guys come in I suppose.

[fine69]

Originally Posted by Samurai

So it is up to me to judge the quality of the 90% using internal measurement process, which relies on subjective judgment. Actually, it is a mix. Recall my 50 criteria scoring sheet, the score for each criteria is subjective, but the grand total gives me a meaningful number, so it becomes measurable.

Which is why I would urge you to go through this link http://www.isixsigma.com/tools-templ...e-development/

QFD as I said, is needed for scenarios similar to the ones you've mentioned. Customer doesn't care how we achieve the product that he desires and the various Houses of Quality not only help us establish & measure the technical part (which is internal), it helps us analyze our capability against the benchmark along with various other things.

It is a very very powerful tool and no Six Sigma new product creation project would be complete without this.

Originally Posted by Samurai

We use simple language and avoid unnecessary jargon. But I often wonder whether this can be scaled to a larger team, probably not. That is where you guys come in I suppose.

Calibration becomes an issue as the operating base gets larger. If all employees had the same level of motivation as the business owner, a lot of non-statistical improvement tools (lean tools) wouldn't even exist. Lean principles are very powerful if applied properly to the service industry.

[Trojan]

Originally Posted by Samurai

Nice to see all the experts drop the jargon and speak plainly. Finally, I can make sense of what you all are saying.

Damn! Now we are talking the same language. We do the same things, except we have not formalized the process to this extent. People are trained to do these things by instinct and habit. We try to instill common sense instead of forcing people to follow processes like a drone. We use simple language and avoid unnecessary jargon. But I often wonder whether this can be scaled to a larger team, probably not. That is where you guys come in I suppose.

Great!!! I am not surprised to hear this, as I said improvement is an implicit part of every organization. While carrying out lean transformations we hear this more than often.
Since, its already a habit in your organization, what can be done is to start putting a structure around it.
Having a top down approach helps, but having a leader who understands why we are carrying out a particular improvement exercise and be passionate about it as well is the differentiating factor.

While improvement is being carried out, it is also necessary to quantify the same.
Once quantified, solutions need to be segregated into Short and Long wins, and tracked to closure. Post closure, sustainability needs to be checked, to prevent a flash in the pan type of project.

There are many ways to achieve this, lets look at a few examples

1. External Consultant
You could rake in the services of renowned consultants, who can help shape the lean journey in your organization. They would also help identify a team of such resources who can then continue working and carrying out the lean journey.

2. Idea generation programs
You can ask your team to come up with ideas which can benefit the organization/well being. Its interesting to see what the team can come up with. But yes, we really need to be prepared to see some of the ideas which come up there
As humans we tend to follow what is the norm and after a point of time even stop questioning.
In 2003, when working for a US client, I found myself and many of my colleagues struggling to answer the questions of the customers on a simple thing as 'What is the time right now?'
Of course the customer was expecting the time at his place as the answer, and we would go crazy trying to calculate.
Solution - Have clocks with all 4 time zones placed at different locations on the shop floor.
This was my first idea - addresses a very small issue but something that everyone started living with

3. Incentive Schemes
While the Business Owner has a very simple reason why he would want to improve, the same thing gets difficult in the way of getting the same implemented and expecting the team to follow.
Reason - employees are paid salaries, it is a result of services for the organization for which hours are clocked and people are expected to provide a certain quality of service (scale of 1-5).

Majority of the crowd is more than happy to achieve a score of 3-4, people usually have the perception of a rating 5 (top) being there for people who are close to the management and doing nothing exceptional.

What does an employee really gain by providing exceptional service?
The client is 'ok' with having a few defects in the delivery, and that does not hamper the contract as well, then why take the trouble of ensuring 6 sigma and/or zero defects.

Now with the same situation, if you put a clause of
- any employee who can provide a delivery with zero defects and/or within a stipulated time frame will get
a. Monetary Benefit
b. 4 Day working week for a month
c. Additional leaves added to the kitty
d. Choose any of the above

Moreover, all employees could avail of this incentive scheme for not more than 3 times in a financial year. So while the incentive of scoring it again is available for the employee. His/Her peers would not feel left out as they have an equal opportunity to achieve the same as well.

The results will be radical, for the same set of employees now they have a dual cause for working towards.
People like to have targets, small time goals, tangible ones which they can relate to.

And the results of such incentives have been gained by organizations

Any or all of the above helps in achieving overall goals of the organization. It also entrenches the idea to exceed beyond expectations and a team which strives towards a more efficient and yet a thorough focus on quality of the deliverable's and the voice of the customer.

Cheers!
Trojan

[noopster]

I don't have a dog in this race but have studied statistical tools to improve efficiency in Production Management back in B-school and been exposed to various techniques like kaizen, TQM, Malcolm Baldrige Framework, Agile etc. in my 15 yr+ career, so let me just say this: most of these techniques are just commonsense packaged as wisdom. This is why Samurai identifies with a lot of what is being said, once it is stripped of jargon.

@Samurai- may I just say one thing. Not everyone has an intuitive bent for problem-solving the way you or I may have. Tools and techniques in the Management world necessarily stem from creating order out of chaos (or at least disorganization). Every model we studied in B-school has the sole objective of structuring your thoughts and extracting the data in as accurate a manner possible, so that is then becomes easier to analyse and process. There is nothing wrong in this. The problem arises when some people make the process the end objective. To adhere to process they start playing around with data and then you are left with a GIGO situation. Garbage In, Garbage Out. At this point a lot of employees become cynical- if the system is anyway going to give garbage as a result, why bother feeding in actual data- so it becomes a vicious circle. Managers also lose track of what the end objective is, or in some cases genuinely do not know what it is. Why do you think a new efficiency theory becomes de rigeur every few years?

The solution is not to trash the theory itself as useless but to recognize that every tool and technique has limitations and use it suitably. I also believe these tools need to be used infrequently rather than regularly- for example, using de Bono's Seven Hats technique while brainstorming a new product concept can produce excellent results but if you are going to do it every week, it will stop being as effective. I know this has nothing to do with Six Sigma per se but all management techniques are grounded in more or less the same principles.

Edit: SIX Hats, apparently.

[Samurai]

Originally Posted by noopster

I don't have a dog in this race but have studied statistical tools to improve efficiency in Production Management back in B-school and been exposed to various techniques like kaizen, TQM, Malcolm Baldrige Framework, Agile etc. in my 15 yr+ career, so let me just say this: most of these techniques are just commonsense packaged as wisdom. This is why Samurai identifies with a lot of what is being said, once it is stripped of jargon.

Even I studied many of these stuff in CQA training and later in MBA. But as we all have witnessed, most managements end up practicing form over function. It soon becomes dogma, and one has follow it without question.

Originally Posted by noopster

The solution is not to trash the theory itself as useless but to recognize that every tool and technique has limitations and use it suitably. I also believe these tools need to be used infrequently rather than regularly-

I don't question the theory, only the practice. The wide spread misuse of processes and the tendency to shelter behind jargon leads to perceptions like this:

https://www.youtube.com/watch?v=NcoDV0dhWPA

Originally Posted by noopster

The problem arises when some people make the process the end objective. To adhere to process they start playing around with data and then you are left with a GIGO situation. Garbage In, Garbage Out. At this point a lot of employees become cynical- if the system is anyway going to give garbage as a result, why bother feeding in actual data- so it becomes a vicious circle. Managers also lose track of what the end objective is, or in some cases genuinely do not know what it is.

So what is the solution for this? What kind of self-correcting mechanism can be employed against this?

Originally Posted by noopster

for example, using de Bono's Seven Hats technique while brainstorming a new product concept can produce excellent results but if you are going to do it every week, it will stop being as effective.

Is the 7th hat yours?

[tsk1979]

Originally Posted by Samurai

So what is the solution for this? What kind of self-correcting mechanism can be employed against this?

We have to understand, that in the end we are all particles.
So everything is determined. Nothing can change.
So no matter what you do the outcome in certain.
There is only one fixed route.
As laplace said, if you knew about all the particles of the universe, you would know the entire future of universe. There is no free will.

When you talk about metrics(objective) think of it as laplace's universe.
What will happen will happen. you have no role to play whatsoever.

However, there is a universe which is real and subjective. Which is quantum.
So at any given point, you only know the probability of a particle being at a certain place, but never its exact location till you measure it.

So there is free will.
There is randomness.

Companies do not like free will and randomness.
So they make this non quantum universe.

The reason for this is very important. Imagine a large company making profit today. They want to make sure that no matter what any manager does, it should have no consequences what-soever on the outcome.

So actually, the manager has been replaced by a deterministic script. There is no spoon, and there is no manager.
Every that has happened would have happened no matter what, and whats going to happen, will happen no matter what.

So the answer to your question lies in deterministic or quantum universe model. What sort of a particle you are. are you a Einsteins quantum particle, or are you 18th century deterministic particle.
do you syndicate impactful experiences, or envisioneer deterministic networks.

A quantum particle can exploit extensible models and disintermediate mission-critical initiatives.
However a deterministic one can only scale vertical metrics.

You being a product development firm can easily forget all the deterministic stuff and simply synthesize intuitive niches. You can grow magnetic infrastructures which will pull customers, but you will not know when and how. Its quantum.

So fear not
To roam the quest is to become one with it.

[noopster]

Originally Posted by Samurai

So what is the solution for this? What kind of self-correcting mechanism can be employed against this?

I wish I knew. If it were up to me I'd stop using that particular technique completely. But like you said, corporate entities often treat these buzzwords as dogma. You are lucky you run your own company but corporate lackeys like myself have no such choice!

An example comes to mind: at the very start of my career I was in FMCG Sales with a food & personal care products company. We got a new Sales Head, one with no background in retail whatsoever- in fact he used to be the HR head and was on rotation . Anyway he had this bright idea that "number of outlets serviced" should be a metric. We were all given targets for "Outlet Coverage" and these had ambitious growth goals like 5% a quarter or so.

You can imagine what happened. The smart managers underreported their outlet numbers at the start of the year, and gradually caught up with the actual number. The honest ones gave an accurate number upfront and then struggled to meet the targets. I protested saying there was no objective way to measure this let alone track and "improve" on it. My then boss, a very wise and gentle fellow, took me aside and explained how one needs to pick one's battles in corporate life and not necessarily be the one to bell the cat every time . I took the hint, and started enthusiastically reporting numbers that my team brought in.

After about 4 quarters of this, the management decided that an HR head in Sales wasn't working out and replaced the guy. The new head took one look at the numbers and did a double-take: apparently we were "covering" more outlets than giants like Lever and ITC were (*not* possible!). Funny thing is that even with all this fantastic new business we were supposedly getting, the numbers didn't go up proportionally. We justified it saying that these were remote/low-value outlets anyway since they were being ignored or missed out earlier. They called our bluff. That was the end of that experiment.

Again- not six sigma, but smart management can see through most attempts at data-fudging using experience and a bit of commonsense. That much-misused term "empowerment" needs to be in place, or else managers become as sheep-like in their behavior as the employees they complain are.

Is the 7th hat yours?

Yes. Apparently the one I was talking through! Corrected now...

[tsk1979]

To continue my enlightened discourse.
In a deterministic universe there is no or very little information.
For example
101110110010101 is information
1111111111111111111 is just repeat(1) = less information.

Information comes from entropy.
In a deterministic company, people often worry about this

What if en engineer starts gaming the metrics without doing any real work, just meeting metrics?
The answer is simple. Since there is no information in the deterministic universe how does it matter if everybody does garbage. There is actually no information because there is no randomness.

So why do I need metrics, why don't I do away with them?
Nooooo. Do not do this. If you remove the deterministic model, you introduce information and knowledge, which will cause your universe to collapse, as the quantum model cannot fit into the deterministic model.

Should I go from Quantum to deterministic
Depends. I mean, if you are tired of the free will, and just want your company to run like an equation and you want to retire while pretending to work, do go deterministic, but remember, your customers also have to be deterministic, or your equation will collapse and not work at all. For example, newtons laws of gravity cannot be applied to electrons and sub-atomic particles, right.
Other way round, deterministic to quantum
when people said that earth is not the center of the universe. They were hanged. So there is a risk. Tread carefully

[fine69]

Let me just add that Six Sigma isn't as common-sensical as its made out to be.

For most of the Lean tools one may say that its pure common-sense, but for most of the Six Sigma tools it cannot be said.

While its an improvement methodology like many others, it give you statistical rationale behind most of its conclusions.

Example - if I toss a coin 100 times and I get Heads 51 times and Tails 49, would you say (statistically) the outcome is biased?
What if I get Heads 55 times and Tails 45 or Heads 60 and Tails 40?

This is where we go a little above common-sense. This is where a Six Sigma tool would help immensely.

The movie Moneyball is based on similar situation when data was relied upon more than anything else and it made history. When the data inputs are utilized in the right manner, Six Sigma starts playing God.

That being said, in order to do an analysis on a particular topic across all social media platforms and that too across the world, we'd have to go beyond Six Sigma tools and start playing with R & Hadoop (Big Data tools).

[Samurai]

Originally Posted by fine69

Let me just add that Six Sigma isn't as common-sensical as its made out to be.

For most of the Lean tools one may say that its pure common-sense, but for most of the Six Sigma tools it cannot be said.

That is very true. But people who use these tools should understand statistics truly well.

See what I have said here & here. In our country, we have MSc Statistics graduates who don't know how/when to take average. First hand experience while interviewing statisticians.

Originally Posted by fine69

The movie Moneyball is based on similar situation when data was relied upon more than anything else and it made history.

Just last week I mentioned this movie to somebody for the same reason.

Originally Posted by fine69

That being said, in order to do an analysis on a particular topic across all social media platforms and that too across the world, we'd have to go beyond Six Sigma tools and start playing with R & Hadoop (Big Data tools).

This is the playground of professional statisticians.

[mallumowgli]

That was a very interesting discussion

Originally Posted by Samurai

Hmm, in product development, there are lots of areas to which the customer has no visibility. For example, 90% of what we do runs on the servers. The customer only gets to see web reports, and that is where they tend to focus. The remaining 90% is technical gobbledygook, and is ignored by customers as long as it works within nominal parameters. So it is up to me to judge the quality of the 90% using internal measurement process, which relies on subjective judgment. Actually, it is a mix. Recall my 50 criteria scoring sheet, the score for each criteria is subjective, but the grand total gives me a meaningful number, so it becomes measurable.

A small point : There is a customer for every process. The person next in line receiving the product is the customer. So you can think in that way. Instead of you as the head judging the quality of work, the receiver of the input of a particular process could be the judge

(PS : Having dropped out of the corporate life few years back, I've no use for six sigma or for that matter alpha, beta or delta. But the discussion was interesting...)