The cause and effect diagram was created by Kaoru Ishikawa in the 1960s.

0:08
It's a formal brainstorming tool for identifying possible root causes.
The purpose of a cause and effect diagram is to gather information and ideas from
as many people as possible and to explore all possible causes of a problem.
0:25
It allows you to identify, explore and
display a problem and all the possible causes in a structured way.
0:34
Strangely this tool has three commonly used names.
Ishikawa called it a cause and effect diagram.
Others called it an Ishikawa diagram in his honor.
And because of its appearance, many people call it a fishbone diagram.
0:50
These are all names for the same tool.
0:54
There are a lot of ways to brainstorm ideas.
The fishbone or Ishikawa diagram is a somewhat structured
approach to brainstorming, specifically aimed at identifying potential causes.
1:09
This is the basic layout of a fishbone or Ishikawa or cause and
effect diagram, you see it looks like the skeleton of a fish.
And our analysis is intended to identify major factors,
those are the big bones, that contribute to the problem.
And then further break these down in to more elementary causal factors,
the little bones, that cause these factors to create or prevent problems.
1:39
There many different suggested categories for fishbone diagrams, but
you don't need to be constrained to follow any of them.
The major reason for
these labels is to get participants to think of different types of causes.
1:54
This scheme is probably the first one that was widely used.
And you can make up your own categories, whatever fits your department or
your application best.
Do not confine yourself to a certain number of categories.
The number of major bones is not critical, but it's usually four to six.
2:15
Let's consider a few other examples.
2:19
Oftentimes, customer service problems are caused by an institution's policies.
Legitimate and necessary policies may cause problems for a variety of reasons.
2:32
A policy intended for one purpose may have unintended consequences in another area.
Policy may have become absolute, but remains in place.
2:43
A policy maybe applied incorrectly because it's poorly understood or
ambiguously defined.
A policy maybe causing problems because no one has bothered to identify legal
approaches to living within the policy and still accomplishing what needs to be done.
3:02
Procedures are a frequent source of perceived waste in customer service
problems.
When a customers looking for a service, they're not likely to enjoy being told
that they must follow some lengthy procedure that from their point of view
has little or nothing to do with the service they're expecting.
3:20
Long lists of required signatures, multiple forms requiring
redone the information in inexplicable waiting periods cause people to
feel like the procedures rather than their needs are what is driving the process.
3:36
Service enterprises involve lots of people who interact directly with customers.
As such, people represent our greatest opportunity to delight customers and
our greatest opportunity to disappoint them.
3:50
People must have the knowledge, skills, and
authority to provide the services that our customers are looking for.
3:58
Every time one of our employees has to say to our customer I don't know how to do
that or I tried, but it didn't work or I'm not allowed to do that,
we create at the situation where our people disappointed our customers.
Finally, we have potential problems with our physical plant or facilities.
This can include lack of facilities or lack of access to these facilities,
or old and outdated facilities that do not meet customer requirements.
4:27
Some naming schemes seem to fit different situations better than others.
4:32
It's really not important which scheme you use or whether you create your own,
the important thing is to consider lots of different types of causes.
4:43
To use a fishbone diagram, first draw the fish
bone structure with the problem at the head.
If you've carefully constructed a problem statement, including measures,
it should go there.
Then add your categories, remember, these are just to spur different ideas,
different areas of thought.
Don't agonize over the categories.
Now you're ready to brainstorm with your team.
One approach is to use post-it notes.
Just write the idea on a post-it and stick it next to the appropriate bone.
5:19
Do not let the team get in a debate about which bone is the best fit,
this is a distraction.
Either pick a bone or write the idea twice.
When the diagram is complete, you can take a picture to preserve the work for
subsequent use.
5:35
Here is a customer service example.
A precise problem in the head is a good idea, but it can be a general
statement of the problem, it can be a product or a process or a service.
5:49
And the causes can be thought of more as enablers, rather than sources of problems.
5:56
This tool can be a lot of work to set up and use.
It's small groups that are often simpler tools for identifying root cause.
6:05
But, if a problem is particularly complex or difficult, or
if you want the input of a large number of people,
there's an alternative to the small group brainstorm.
You can create a large fishbone diagram perhaps on four by eight foot foamcore.
Then supply post-it notes and mount the diagram in an area where anyone and
everyone will see it.
Anyone walking by can read the diagram, and contribute by posting a note.
6:36
Used this way, the diagram must be actively managed.
At least once a day you should organize the notes, and
remove any that are not serious suggestions.
You might leave it up for a week.
And you can get suggestions from people on other shifts, even from suppliers or
customers visiting your work space.
6:58
Cause and effect diagrams may be an effective tool for
generating ideas from small or large groups.
This is a brainstorming tool.
It generates a large quantity of ideas about possible causes.
It does not identify root cause.
The team needs to investigate all of the ideas and
use other means to identify probable causes.
Checksheets are simple and flexible tools for collecting data.
When you begin an improvement project,
you may find that you don't have any of the data that you need.
A Checksheet is a simple, quick,
and usually temporary tool for collecting the initial data that you need to get started.
There is no single checksheet form that can be used in all applications.
Therefore, the design of the form must be
customized to the type of data that's to be evaluated.
You can use check sheets to count or measure defects,
complaints, phone calls, or many other things.
There are three main types of checksheets.
Recording checksheets, where you enter counts or measurement data.
Checklists, which are written reminders.
And the Measles chart,
which can be used to identify the physical location of defects or other issues.
An example of a checksheet is shown here.
It's based on the method that customers use to make a complaint.
Complaints were received four different ways: by letter to a post office box,
by calling a toll free 800 phone number,
by e-mail to a special email address,
by walking into the store and complaining to a customer service representative.
Each time one of these events happens,
a corresponding mark is made in the appropriate row.
This table summarizes how
complaints were received during the past week.
The total number of complaints is 25.
We can calculate the percentages as follows: There were three written complaints.
Three divided by 25 equal 0.12.
We can multiply that by 100 to obtain the percentage.
Repeat this for each row.
An analysis of complaint methods like this might help you to
start analyzing staffing needs.
This is another example of a Recording Checksheet.
This one is used to record the measured time to resolve a help desk issue.
If a call lasts less than a minute,
a mark is made in the first row in the calls column.
If it's more than one minute but less than two,
a mark is made in the second row in the calls column and so on.
It's always a good idea to include important information about the process,
the location, area, shift,
and who is involved.
This can be done at the top of the checklist.
If you've ever made a grocery list,
you've made a checklist.
Pilots routinely use checklists to make sure that nothing important is forgotten.
They're also used in areas such as housekeeping at a hotel or in medical procedures.
Whenever there is a complex or delicate procedure,
a checklist will help prevent errors.
A Measles Chart is a physical representation of something of interest.
This can be an accurate drawing or a rough sketch.
The idea is to make a mark on the drawing each time you find a defect,
and to make that mark in the same location on the drawing as the actual defect.
In this case we have a drawing of a bumper with several X marks
to indicate where defects have occurred.
It's also possible with the Measles Chart to record more than one type of defect.
And with every type of checkesheet you can
separate data by shift or by different production lines.
Before you collect any data,
even with something simple like a checksheet,
you should make a plan.
You should know what data you want, how,
and who is going to collect it,
and what you think you will do with it.
The data collection plan does not have to be complicated.
Just create some columns and list the answers to these three questions.
Checksheets are simple way to collect data that is not otherwise available.
They're quick, flexible, and can be customized to suit your needs.
When you don't have any of the data that you need this might be a good place to start.

One of the basic and primary tools for process improvement is a process map,
also called a flowchart.
A process map is often the starting point for improvement efforts.
They can be used to demonstrate actions or procedures.
Before we can understand a process map,we need to understand what a process is.
A process is any activity,
usually a number of steps that receive an input and convert that input to an output.
This could include changing raw material into
a component or finished product in manufacturing.
It could be the steps in the delivery of a service or it could be acting on information.
There are many types and variations of process maps.
Here we see five major types.
We will focus on the most basic.
There are also many different variations of process maps.
We'll discuss some of these in future modules.
The purpose of creating a process map is to gain
an understanding of a process so we can improve it.
Process maps are useful for documentation and for training,
for certification and other uses.
Process maps can help to eliminate
written procedures by presenting them in a visual schematic.
For high level business processes that cross functional boundaries,
there often is no one person who knows the whole process.
This is why we need teams to map processes.
The participants in a process mapping exercise should include
the people who actually use the process on a daily basis.
An example of a business process map might be the process of creating purchase orders.
Process maps are appropriate and necessary in nearly any process improvement effort.
They should be one of the first things that a process improvement team does.
It's important to map the process as it is actually being performed.
Not the way it was designed or the way the manager thinks it's done.
We call this the current state map.
As we mentioned, there are many levels of processes.
We're going to talk about two.
Business processes and work processes.
Business processes tend to be
high level processes like accounting, customer service, etc.
They often cut across functional departments.
They also contain work processes.
Most processes are also interrelated with
outputs from business processes and work processes,
being inputs for other processes.
Because business processes or high level processes,
they contain multiple work processes.
For example, accounting may include billing and tracking accounts receivable.
Each step in a business process may contain
one or more work processes which can be
mapped separately and will provide much more detail.
There are a number of different shapes that can be
used to create process maps and this may vary by industry.
These are some examples for manufacturing.
For our purposes we will use only a few.
The level of information in a process map can vary.
But we're focused on the simplest form.
There are many different symbols that can be used and we're going to show
you how to map a process with just three symbols.
The symbol for the start or end of
the process is an oval or a square with rounded corners.
Inside this symbol, it may say start or end.
A square or a rectangle is the symbol for a process step.
Inside this symbol you will name the step.
Decision points are of diamond shape.
They can only have two possible answers like yes/no, or pass/fail.
Usually one answer will lead to a continuation of
the process and the other answer will either end the process or redirect it.
If there are more than two possible outcomes,
you might need more decision points.
One easy way to map processes with the group is to use sticky notes.
As the steps develop you can rearrange as needed to get the right sequence.
A standard sticky note can be turned 45 degrees to use for a decision point.
When the team is done with the sticky note map,
it can be preserved by taking a picture.
For a more permanent and professional looking map,
the information can be easily transferred to
one of many computer programs that have drawing tools.
This could be something like Microsoft Word or Excel using the drawing tools function.
There's also free software that will do this such as
Google Docs or Sheets or Open Office.
Vizio is software that's designed specifically for this type of application.
This is an example of a simple process map.
In this case, it's someone's morning routine.
Yours is probably different,
but it includes begin and end points,
a decision point, and multiple steps.
One of the biggest challenges for a process mapping team is
determining the appropriate level of detail and using it consistently.
There is no rule for this.
It's a judgement call by the mapping team.

PDCA is a simple improvement model that has evolved over decades with many contributors.
It probably began with Walter Shewhart,
although W. Edwards Deming often gets credit for it.
PDCA stands for Plan Do Check Act.
Another variation of this model comes from Deming in its PDSA or,
Plan Do Study Act.
In either case the simple four step circular model can be easily
understood and it can be mapped against
nearly any other improvement model such as the DMAIC model.
The PDCA model is a simple approach that can be
applied to process improvement in manufacturing,
service, education, healthcare et cetera.
It provides the team with a disciplined approach and helps them manage the improvement.
Without the discipline of an improvement model,
teams have a tendency to just do it without
the necessary discussion, documentation et cetera.
The first step, of course, is Plan.
In this step the team will define the problem and discuss possible causes and solutions.
The problem definition is critical.
It helps limit the scope of the project and acts
as a touchstone if the team begins to wander.
Ishikawa added determining goals and targets to the step.
The team might also develop predictions and plans in this step.
In the Do step the team implements their solution.
Ishikawa added education and training to this step.
This is where plans from the plan step are carried out.
New problems or observations should be documented and data analysis can begin.
It should also be noted that if possible a pilot
or a simulation is a good way to minimize risks.
In the Check step,
the team evaluates the results of their solution and looks for other effects.
It's not uncommon for a solution to be partially successful.
That is why the model is presented as a circle.
The data analysis is completed in this step.
The predictions are compared with results and lessons learned or documented.
Successful parts of the implementation should be
documented and standardized in the Act step.
For further improvements, return to the Plan step and begin the process again.
Demming considered his model more of a plan for management.
Instead of Check, he identified the third step as Study.
In the Study phase the results of the improvement would be compared with
predictions that were made in the Plan step to ensure learning.

These are two common Six Sigma measures,

defects per unit and defects per million opportunities.
DPU or defects per unit,
is simply a ratio of the number of defects to the number of units produced.
Or the total number of defects found in a sample divided by sample size.
To calculate, you simply divide the total number of defects by the number of units.
Let's say, for example, that you're processing mortgage applications.
0:34
This month, your organization processed 75 applications.
In these applications, 12 errors or defects were found.
We would divide the number of defects, 12, by the number of units, 75.
There're 0.16 or 16% defects per unit.
We multiply the decimal by 100 to find the percentage.
1:04
So when we multiply 0.16 by 100 we find that about 16% of
the mortgages had an error.
But, we need to remember that a single mortgage application
could have multiple errors.
1:18
Let's say we're making automobiles.
In this month, we've produced 1000 new cars.
1:24
Automobiles are complex machines and many things can go wrong.
We found 9000 defects in our automobiles.
When we divide the number of units produced,
1:36
1,000 into the number of defects, 9,000.
We find that we have 9 defects per unit.
1:48
Some units may have more defects and some may have none at all, but
the average number is 9.
1:58
Defects Per Million Opportunities, for this measure, we want to convert our
defects, or errors, to see how many there would be out of a million opportunities.
First, we should define opportunities.
Every possible defect in a product or
error in a service represents an opportunity.
So from our mortgage application example, let's say there are 30 different fields
that need to be properly filled out in the application.
We multiply the number of fields, or places where an error could occur,
by the number of mortgages processed.
30*75 = 2250 opportunities for
error in our 75 units.
To calculate DPMO we first divide the defects by the opportunities.
2:54
Then we multiply this number by 1 million.
2:57
In our example, there were 12 errors.
And we just calculated opportunities, and
found that the 30 fields on 75 applications
represented 2,250 opportunities for error.
Then we divided 12 by 2,250 and
we get 0.00533,
multiplying this number by 1 million,
gives us our defects per million opportunities of 5,333.
If you´ve already calculated DPU, or defects per unit,
you can use a little simpler formula.
You can divide DPU, defects per unit, by the number of opportunities per unit.
3:48
That and multiply that by 1,000,000.
As you can see, you'll get the same result.
This measure gives us a different perspective regarding errors and defects.
4:01
Let's turn to our automobile example again.
We found that we had an average of nine defects per vehicle or
9,000 defects across 1,000 vehicles.
Consider that automobiles are complex machines.
There're a lot of opportunities for error or defects.
Let's say there're 500 opportunities for a defect in each vehicle.
Actually, there're probably many more.
If we multiply the number of opportunities per vehicle by the number of vehicles,
we find that we have 500,000 opportunities for defects.
Remember, to get DPMO we first divide defects by opportunities.
We had 9,000 defects, so we divide 9,000 by 500,000.
This equals 0.018.
We then multiply this by one million.
Again we could use the shortcut but since we have already calculated DPU and
the result is the same.
5:03
Remember, defects per unit is simply the number of defects found in a sample
divided by the number of units in that sample.
5:12
To calculate defects per million opportunities,
we first have to calculate the number of opportunities.
5:20
We get this by multiplying the number of opportunities for
error in a single unit by the number of units in our sample.
This gives us the total number of opportunities.
Then, we divide the number of defects by total opportunities, and
multiply by one million.
5:39
Or, we can use the shortcut, and
divide DPU by the number of opportunities for error per unit.
And multiply that result by one million.
In either case, the result is our DPMO, or defective parts per million opportunities.

Our next measure is called Roll Throughput Yield.

Roll Throughput Yield is the probability that a process with more than one
step will produce an error or a defect-free unit.
0:17
This also applies when multiple components are used to create a single product.
0:24
Before we can calculate Roll Throughput Yield,
we need to calculate First Pass Yield for each operation or each component.
First Pass Yield is actually just a complement
of our defects per unit calculation.
Remember to get defects per unit,
we divided the number of defects by the number of units.
Let's take an example.
Say we are producing 50 units and we find 3 defects.
0:52
We divide 3 by 50 and we get a DPU of 0.06.
We can get the complement of our DPU calculation by subtracting it from 1.
In this case, the answer is 0.94.
We can interpret this to mean that we have 94% good quality.
It also means that, for any individual part,
there is a 0.94 probability that it will be defect-free.
This is our First Pass Yield.
1:24
It's also possible to account for rework in this calculation, but for
now we will keep it simple.
1:33
Now let's add some more components to our assembly.
1:38
We now have four components in our assembly.
The First Pass Yield for
these components is 0.94, 0.96, 0.98, and 0.96.
Again, that is also the probability that a component will be defect-free.
It would seem that these numbers do not look too bad, but
we want to know how many of our assemblies will be defect-free.
2:08
Our assembly could have a defect occurring from any one of these components.
2:13
According to the multiplication rule of probability,
we can find the probability of multiple events
happening at the same time by multiplying the individual probabilities.
2:26
So if we multiply the First Pass Yield or
probability of good parts for each component together,
2:35
we can get the probability that the assembly will be defect-free.
This is RolledThroughput Yield.
The answer rounded off is 0.85.
That is, there's only a 0.85 probability that
these four components will produce a good assembly.
Imagine how this works in complex products, like a computer or
an automobile transmission.
The more components or steps in the process, the more opportunities for
failure.
3:08
This time you try it.
Remember, to get DPU, divide defects by units.
To get First Pass Yield, subtract DPU from 1.
First Pass Yield is the probability that a component will be defect-free.
3:25
Now multiply the First Pass Yields for
each component together to get the Rolled Throughput Yield for the assembly.
Now pause this video, and try the calculations.
3:43
For each component, we divided the number of defects by the number of units to get
the defects per unit.
3:50
Then we subtracted that result from 1 to get the First Pass Yield.
This is the probability that a component will be defect-free.
Then we multiply those probabilities together
to get the Rolled Throughput Yield, which was 0.95.
So there is a .95 probability that an assembly made up of these
components will be defect-free.
4:15
Did you get it right?
If not, review the video and try again.

Affinity diagrams are another silent technique.

It's particularly useful when the number of ideas is very large, or
the problem or causes being brainstormed are not well-defined.
Allow five to ten minutes for participants to write ideas down on Post-it notes,
0:20
one idea per Post-it note.
When everyone is done, put them on all or whiteboard.
0:27
All the members of the team then go to the board and
silently rearrange the ideas into logical groups.
No talking while this is done.
If an idea appears to be closely related to two groups, then write another note and
put it in both.
0:45
This method can take 50 to 75 ideas and reduce them to 6 or 8 groups.
Once the logical groups have been developed, the team can talk, and
discuss a name or label for each group.
Activity network diagrams are excellent for
organizing tasks resulting from the brainstorming session.
These task items need to be sequenced,
meaning, which task can be performed before the other.
You may discover that some tasks are not related,
and can be done at the same time as other tasks.
While these tasks are being refined or perhaps broken down into further tasks,
the duration of each task can be estimated.
The duration can be expressed as: the Earliest Start Time or ES,
the Earliest Finish Time or EF,
the Latest Start Time LS,
and the Latest Finish Time or LF.
There are a number of software packages capable of
organizing and tracking these tests such as Microsoft Project.

Another decision making tool is multivoting.

0:04
In multivoting each team member gets multiple votes, usually five, but
if there are a large number of ideas, there might be more.
Again, start with a reduced list.
If you have 12 ideas, you might allot five votes to each participant.
Participants can apply these votes anyway they want.
They could vote once each for their top five ideas, or
if they feel really strong about one idea, they can put all five votes on that one.
Or they can do anything in between.
Record and add up all the votes.
Discuss the reasons for voting.
If the decision is not clear, you may as a group remove items that got no votes or
only a few votes and then vote again.
Nominal group technique is similar to the slip
method of brainstorming that we discussed earlier.
The topic or problem should be defined to start with.
Participants then get five to ten minutes to quietly write their ideas on
a piece of paper.
The facilitator then asks for one idea from each participant in sequence.
0:21
The facilitator will write these on a flip chart.
Other brainstorming methods, no discussion or evaluation is allowed at this point.
When all ideas are accurately recorded discussion begins to clarify and
evaluate ideas.
0:37
Finally, the group will select one or more ideas through the use of the team decision
making tools we will talk about in a few minutes.
Force Field Analysis is a team tool with lots of applications including root cause.
This is another structured brainstorming tool.
A Force Field Analysis seeks to identify forces that will
help you accomplish what you want and those that might work against you.
If you can identify the driving forces that will help your team accomplish their goal,
you might be able to leverage those forces.
If you can identify forces working against what you want to accomplish,
you may be able to minimize or mitigate them.
At the very least,
this information will help you to understand the problem better.
A common way to do this is to use a flip chart or a white board,
draw a large T-shape and label the two sides as shown here.
It's good to pair driving and restraining forces but it's not always possible.
Sometimes, your team might also identify
forces that seem to belong on both sides of the chart.
Let's say that you're leading a team that has the task of implementing
a process change or even a process improvement initiative.
You might brainstorm the forces that will help you succeed,
and those that might work against you.
The team will probably come up with many more than are shown here,
but as an example,
there may be a strong desire to improve
customer satisfaction but you may have employees or managers who are resistant.
Perhaps you have a particular manager who is supportive and wants the team to succeed,
but there may be others who feel differently.
Some employees may be eager to be involved and to help make things better,
but they also face the pressure of the production schedule.
Employees may have lots of good ideas,
but do they have the resources to implement them?
When you've identified these for and against forces,
you may be able to come up with ways to use the strong driving forces to help you,
and you may be able to head off or minimize some of the restraining forces.
You might use Force Field Analysis to identify
root causes of something like late customer orders.
The idea is that for large problems like this,
there are a number of forces that combine to produce the current situation.
If we can identify them we might be able to address the causes.
What if we have this problem with late customer orders and we want to fix that?
There are a lot of things that could cause this problem.
These are just a few.
Management's desire for getting rid of late customer orders is certainly a driving force.
Lack of capacity to do all of the work that
the organization has committed to may be a restraining force.
Employees usually want to do a good job and may be willing to do what is
necessary but perhaps management is not authorizing overtime.
Maybe poor quality is working against us as we have to
rework product or rerun some orders causing us to be late.
But we have also just begun process improvements,
so maybe we can leverage that effort to help solve the problem.
Another restraining force may be a bottleneck in the process.
Perhaps we can alleviate this bottleneck if
our employees are flexible and willing to stagger breaks and lunch.
These were just a couple of examples of the use of the Force Field Analysis.
It's usually done by a team to help them to understand a large project or problem.
Part of the attraction of this tool is it is so simple and easy to use.
All you need is a marker and a flip chart or a white board,
and a team with ideas.
When brainstorming root cause with a cause and effect diagram for
instance, the team may come up with a large number of possible causes.
You may not have the time and resources to investigate all of them so
you need to prioritize.
0:16
You can also use this tool for
other team decisions like selecting a process improvement or a vendor.
This tool might be used in the analyze phase, when determining root cause,
or it might be used in the improve phase when selecting an improvement.
0:34
It's easy to do a prioritization matrix in Excel.
There's a little math involved so Excel simplifies this as well.
0:45
In this first example we'll use the tool to select a vendor from
three possibilities.
0:52
Across the top row the team will list their decision criteria.
0:57
These are not predetermined,
this is one of several team decisions that must be agreed upon as you use the tool.
1:05
After discussion the team decided that the most important criteria for
this selection were cost, quality, flexibility and delivery.
But cost, quality, flexibility and delivery are not of equal importance.
So the team must decide how to weight these factors.
We use decimals, but you can also use percentages.
The team decided that quality was the most important factor and
gave it a weight of 0.4 or 40%.
Cost was 0.3 or 30%.
Delivery was the next most important criteria and at 0.2 or
20% and flexibility was weighted at 0.1 or 10%.
Regardless of how many criteria you choose,
the weights must add up to 1 or 100%.
The next step is to rank order the options with
the largest number being the most desirable.
The team determined that Vendor C had the best price,
followed by Vendor A, and Vendor B was the most expensive.
2:16
So since B was most expensive it gets the lowest ranking, 1.
Vendor C was the most desirable price so it is 3.
This was repeated for each criterion.
2:29
Each time you assign a ranking to the options based on a criterion,
the team will have a conversation and reach consensus.
This tool takes some time.
2:40
Once all the option have been ranked against all the criteria,
you'll multiply each ranking by the weight assigned for that criteria.
For example, under Cost, Vendor A was ranked number two, so
we multiply 2 times the relative importance of cost, which is .30.
The result is .60.
Vendor B was the most expensive, so
it has a ranking of 1, and its score is .30.
Finally, Vendor C had the most desirable price.
So it's score will be the weight or
relative importance times its ranking of 3.
So its score is .90 on cost criteria.
Repeat this for every option in each criterion.
3:34
The multiplication and addition can be done more quickly and
more reliably if you use Excel.
3:40
Once you've created a score for each option against each criteria,
you can add the scores for each option going across.
The scores here are shown in parenthesis.
So for vendor A it's .60 plus .40 plus
.30 plus .60, for a total of 1.90.
Repeat this for each option, adding across.
The highest total score is probably the most desirable choice.
4:13
A caution about the Prioritization Matrix.
As we go through this process, we're assigning weights and
ranks and multiplying and
adding numbers to these fields like it's quantitative but it is not.
The team decides what those weights and ranks will be so it's qualitative.
The real value is in all of the team discussions.
The team has to discuss and decide what the criteria would be.
The team will have another discussion to decide what the weights will be,
then team will yet have another discussion to decide on how to rank the options.
The real value in this tool is in these discussions.
4:58
Through this process,
the team gets a deeper understanding of the decision it's making.
5:05
Do not let the numbers make the decision.
Sometimes a team after all this discussion will decide
on the option with the second highest score.
Because after all of the discussion they may decide that the weighting or
ranking were not quite right.
5:23
Let's look at another example.
This time we're selecting which improvement option to give priority too.
In this case the team had chosen the following criteria and weights.
The cost of the implementation is first with a weight of 30% or 0.3.
Speed of implementation is the most important
because we want to be able to demonstrate some quick results to gain buy-in.
So it's weighted at 40%.
Complexity is probably closely related to speed but
the team decided to include it and
its weight is 10% and finally the benefits of improvement are weighted at 20%.
When we ranked the options the teen felt that improvement two and
improvement three We're the same on the speed criteria.
If there is a tie like this, use the value half-way between.
So the team could not decide which was one and which was two so
they ranked them both 1.5.
Now we multiply the relative importance or weight times
the ranking just as we did before and add up the scores for each improvement option.
Improvement 1 had the highest score since the score
between Improvement 1 and Improvement 2 is fairly close
the team might have some additional discussion about this.
However, since Improvement 1 had the highest score,
in the most important criteria and was no lower than second in any other,
it seems likely that the team will choose Improvement 1.
Remember the team makes the decision not the tool.
A SIPOC is a high level process map that includes suppliers,
inputs, processes, outputs and customers.
SIPOCs serve to bind the processes together,
and like other process maps,
serve to promote effective communication.
The SIPOC diagram will help you understand the process,
know what your customers consider to be the most important,
and where to start moving the project forward.
The SIPOC is used to help see the business in a process perspective.
It also shows the cross-functional nature of a process.
Additionally, it shows the inter-relationship
between the customer and the supplier in each interaction.
This is a key difference from a normal process map,
flow chart, or value stream map.
When completing a SIPOC,
here are some questions that will help.
Note that these are all closely aligned with the types
of questions we ask regarding processes.
Here is an example of a SIPOC.
Note that inputs and outputs can be tangible and intangible.
The scope of the suppliers and customers are also quite broad.
In addition to the apparent benefits regarding inputs, outputs,
customers, and suppliers, the SIPOC helps us establish project boundaries.
Steps for creating a SIPOC include identification of process boundaries and activities.
We must define our outputs and associated customers.
Next, our inputs and associated suppliers.
Critical to satisfaction measures are specified in relation to inputs.
Process steps and outputs are then added.
Make sure to be specific.
SIPOCs are bound to need further adjustment.
Here is an example for taking a business trip.
Where multiple processes are in play we can
combine into one overarching flow that defines all steps in the process.
We often uncover more inputs,
outputs, and process steps when creating SIPOCs.
Additionally, we may discover suppliers and customers we never knew we had.
Here is another example for washing clothes.
Processes can be bound over simple tasks or complex tasks,
even cyclic tasks where you are both the supplier and the customer, can be explored.

Tree diagrams are popular, and looks like an organization chart.

This can also be accomplished during a brainstorming session.
The problem is represented at the top, while the affinities or
categories are branched out along.
With the possible problem areas to explore.
The tree diagram can be used to develop elements for
a new product, it can be used to show relationships of a production process.
It can be used to create new ideas in problem solving,
it can also be used to just outline steps needed to implement a project.
0:36
In this example, we see that the paint inspection failed.
0:42
Possible areas of concern could be the paint booth,
it could be the adhesive assembly, or it could be simply the design.
0:50
So, we break that down even further to say what is it we're going to look at within
these areas.
So you see for instance, the paint booth we see that the sprayer could be broken.
It could be an over-spray or under-spray.
It could be the paint booth, if it's on an assembly line,
it could have a line speed issue.
Or you could have the wrong paint.
1:11
So, these areas are explored by the team.
We control things to consistently meet standards.
In a control process we continually observe
the performance of a process to maintain gains,
analyze the variation, and move closer to a state of self-control.
The PDCA method is a fundamental yet effective means of driving change.
Before beginning we should understand what we are trying to accomplish,
what we can change to achieve the improvement,
and how we will measure improvement.
We can also explore economies in our process by standardizing as much as possible.
We cannot solve everything at once.
So starting with the most important root cause can give us the most substantial gain.
A control plan is the manner in which we sustain
the gain and implement the proposed change effectively and robustly.
As we deploy policy changes we must ask ourselves,
which corporate policy should be changed as a result of the project?
Have some policies been rendered obsolete?
Are new policies needed?
Did the project bring the organization into compliance with a standard?
Based on the outcomes of these questions we must be prepared
to modify procedures for the new process,
modify quality appraisal and audit criteria,
update prices and contract bid models,
change information systems, revise budgets,
revise forecasts, or modify training.
One useful approach in controlling a process is through dynamic control planning.
Dynamic control planning is a collection of
essential information that pertains to a process.
This information can include SOPs, control plans,
FMEAs, gauge control plans,
and quality planning sheets.
Document control plans are living documents that outline a larger scope of documents.
As such, when changes are needed,
operators have the responsibility to make these changes known so that
an update can be made and corresponding training can take place.
A gauge control plan focuses specifically on the instruments used
to monitor and assess the proper function of the process.
It is not meant to be a substitute for actual calibration,
but more a guide of what, how,
and when certain tasks shall be completed as pertains to the proper function,
performance, and upkeep of the instrument.
It also outlines contingency planning,
should certain conditions fail to be met.
Control charts show process variation while work is underway.
It provides a means for monitoring the state of
the process in real time and detecting issues.
SPC is a graphical tool used to monitor ongoing performance.
Control charts can trace their origins back to Shewhart at Western Electric in the 1920s.
Control charts established performance boundaries.
They also help us uncover assignable causes of
variation and distinguish these from chance causes.
There are over 30 types of control chart's in use today.
So control charting can be applied to many types of data.
They also assist in creating a foundation from which
professionals can discuss and address process performance issues.
Control charts are applicable for any scenario that varies over time.
We particularly are interested in using these types of tools to
assess process performance after an improvement has been implemented.
Remember, control charting is not just for monitoring.
It is also for detection.
Picking up where we left off,
the components of the control chart include a center line,
an upper and control control limits.
Quality characteristic values are plotted along the vertical axis.
There are two types of control charts, attribute and variable.
With control charting we can know when to take corrective action,
the type of remedial action to take,
when to lead the process along,
the process capability, and the possible avenues for quality improvement.
Before we can eliminate variation,
we must recognize that there are really two types of variation at play.
Special cause is assignable variation,
something that is not inherent in the process.
Common cause is due to chance.
Variability due to common or chance causes is something inherent in the process.
Demming believed that 15 percent of all problems are due to special causes.
Action on the part of management and workers can reduce special causes.
Demming also believed that about 85 percent of all problems are due to common causes.
These can only be solved by management since you must change the system.
The center line is where the average or target value will reside.
Typical processes can be deemed out of control
when points are outside the control limits,
there are nonrandom patterns,
or if target values are on only one side of the center line.
If the process is under statistical control then we can estimate the process parameters,
the mean, the standard deviation,
and the process capability.
Per the central limit there,
as we increase the sample size,
the control limits will be drawn closer together.
This is due to a decrease in the variance.
Control limits are usually placed at plus or minus three sigma.
This captures 99.74 percent of the behavior of the sample statistic.
This is our indicator that the process is in control and
our process consistently performs within these limits.
In most cases sampling is done in time order or by process.
We call this type of sampling the instant of time method.
Observations are selected at
approximately the same time for the population under consideration.
If we want to detect a small shift,
a large sample size is needed.
Choosing large samples frequently provides
the most information but is not always feasible.
We must consider whether we have destructive testing and the cost of sampling.
Once you have localized and eliminated assignable causes we should
remove out of control points and revise the center line and control limits.
There is no need to maintain control charts if
the process demonstrates consistent capability.
Instead, focus your attention and resources on other areas.
Be aware that the control limits on
a control chart are influenced by the variability in the process.
As such, these control limits will change.
Today's technology makes this a very manageable behavior.
Control charts are easy to set but very difficult to maintain.
Often our failure to maintain comes down to our inattention,
in other words, failing to make adjustments when the control chart detects an issue.
Prioritization of productivity also overshadows control charting.
Lack of training can also derail a control charting program