Six Sigma Green Belt-Value of Six Sigma
CP & Cpk are capability studies - Capacity Performance
- estimates of the future of the process; -Predicts into the future, process must be stable Mapping voice of process vs. voice of customer
Walter Shewhart
Father of statistical quality control originator of PDCA cycle Plan do check act. introduced the control chart and ramdom variation
Control Phase
Finalized control systems are determined and long-term capabilities are verified for sustainable and long term success. Import to understand how the process will cont. to be measured after implementation. this is performed through statistical process control and control plans that determine exactly who and what should be measured. Standard processes are put in place so that the old way of doing things does not happen again. Improvements are quantified to make sure the goals of the project were actually achieved
Randomization-I
Mix up order of experiment each time, don't do same order, might not be practical because of economics may have to break a machine down and put back together in different order to perform test, it may take two - three hours to do this, not practical/or economical to keep changing back and forth.
Defects are
Mistakes, Flaws, Customer complaints , Warranty returns, liability exposure , all non quality items
Tree diagram
critical to quality diagrams translate customer requirements from a broad perspective to specific, actionable items that are able to be measured / helps show scope of project
ordinal scale
difference bt data values cannot be determined or are meaningless Ex grades-ABCDF Movie ratings PG, X, R ; Horse race, 1st place, 2nd place, 3rd place. 0 doesn't mean anything-zero doesn't mean anything
t-test formula-A
difference in group means/square root of correction for group variance over correction for group size t= (xbar-u) mu/s/sq root n n=sample x bar= the mean of the sample data s is the sample standard deviation u population mean for a t test the mean of the pop the data is drawn from the known
Project Closeout
most important close out with clear reports and actions and best practices should be documented for all the lessons learned. Successful project should end with a celebration
Non complementry
mutually exclusive is non complementary
actionable measures -m
need a list of actionable measures when investigating customers need and requirements. This is part of fixing the process to improve your capabilities and customers perception of your performance compared to competitors
calculating Sigma
need to know 2 things either calculate DPMO defects per million opportunities 3 or 4 features of product defective DPMU defects per million units -all or nothing either completely right or completely defective. use sigma table with 1.5 sigma shift Six sigma is 3.4 defects per million & sigma calculator
Project Plan Living document
need to update as information becomes available . you will look at the projects X's, the Vital few X's that are critical to the management of the project Y. and you are going to respond based on what you see during the monitoring process.
Customers needs and requirements -m
needs are = to a process output requirements identify a specific characteristic of an output a customers needs and requirements don't always match up clarify that need and the requirement actually match up
mean cal-m
the arithmetic average of a distribution, obtained by adding the numbers and then dividing by the amount of numbers you added 4 + 3 +2+3=12 /4=3
Visual Management -Key Lean Tool
-It included the 5S and the 6S systems -It allows you to organize your work place get things cleaned up and have a method of sustainability
Kaizen -Key Lean Tool
-Loosely translated into continuous improvement - good change - idea is to constantly review and make improvements so that you. your products get better ,faster, easier, safer, cheaper
The 5 S's -I
-Sort is elimination of unneeded materials -Set in order is where needed materials are organized -Shine is when work areas are cleaned and inspected -Standardize is writing and adhering to standards -Sustain is where standards are regularly applied in all areas.
SMED-Key Lean Tool
-Stands for Single Minute Exchange of Dies -It is a form of quick change over
Assumptions for t-test -A T- test
1) 2 samples are independent * critical 2) 2 populations are normally distributed 3) 2 populations have equal variances formula works if sample sizes are equal and assumes equal variance sample size less than 30 they are equal Not just for analyze phase answers question Are these two means statistically or reliability different , can use in define phase
Define Phase
1. Project Charter
How many types of data
2 Attribute ( discrete data) & variable data
Monitors of Process Capability
2- Pp and Ppk Pp - how wide the bell curve is compared to the window of opportunity from the customer Ppk- Indicated the centering of the data or the curve
Cost of -Poor Quality
25-30% of annual revenue is lost to it
FMEA Issue Criteria - I
3 main criteria for selecting issues to address are 1-The highest RPN number (risk priority number) 2-A major pain point that operators continue to complain about 3-A main cause of downtime that is reported repeatedly
Measurement Scale Chart
4 types of data; Nominal, Ordinal, internal and ratio
Implementation Planning -I
5 things to take into consideration -People- are they available are they bought into the goals. -Budget -you know what moneys are available . what resources are available -Change management- how bought in are they to the problem and the solution or the improvements? -Timeline- is there an expectation that this needs to be done at a certain time in order to realize the value? -Measure of success- not as straightforward as measuring that end goal, there may be leading indicators to success that you want to measure to ensure you're moving in the right direction
Three Factors/ 3 Levels -I
= 27 trails or runs written 33 (3x3x3)=27 Try to stick to two levels The problem with more than two levels is the cost goes up as the number of levels do
Three factor/ two level experiment
= 8 trails or runs written 23 (2x2x2) =8 Use this for experiments. most cost effective
Tollgates -15
Are clear, logical, and thorough. DMAIC and the tollgates are a systematic, repeatable, reliable process to drive success during the improvement projects that can be applied to any process
Cost of Poor Quality- Appraisal Cost
Audits, testing, Inspections, equipment calibration
Define
Covers the project Charter, customer needs and requirements and process mapping
How much data to collect
DO NOT Guess worst way; consider cost, can be expensive, government agencies and regulatory groups may tell you what your minimum sample size should be. Industry standards will tell you requirements/ formulas are available to help you determine good minimum size.
Fishbone Diagram
Excellent tool for helping teams brainstorm causes of a problem or effect. Called a cause and effect diagram
Black Belts
Facilitate or lead large complex improvement projects/Coach mentor Green Belts during their projects/ ensure tollgates are completed correctly before moving forward/manage the improvement plans for entire product lines-short term, long term solutions aiming towards one goal. / dedicated fully to six sigma projects
FMEA -I
Failure Mode and Effects Analysis tool used to select action and prioritize projects FMEA answers these questions What might cause a problem? What is the risk assessment ? What steps are next?
Gage R&R
Gage Repeatability & Reproducibility will check precision but not for accuracy
During Meeting
Greet Sponsor/ Remind them about which project is being discussed/ Review the prior phase or tollgates, mention any recent updates to thing like the charter, scope etc- always notify sponsor of any changes and provide most up to dates documents/Review current info. Goals & tollgates, what u planned to do & what you did, tools used and findings- Keep tight to the point Answer questions- be honest- get back to them Request approval to move on or revisit if need be Ask for resources if needed If approved let sponsor know next tollgate & your goals, ask for their input , guidance and suggestions Thank then for assistance
Process Owner
Have influence over the improvement team, holds team accountable for completing action items/subject matter experts on process/can be a formal (supervisor or manager expert on process) or informal leaders (with comp a long time people respect their opinions & look to them for improvements) or a sponsor
Larger Ranges
Have less consistency
Dr Armond Feigenbaum
He believed that quality improvement required everyone to be fully committed for it to be success. Total Quality Control or TQC. High quality means lower costs, that management must be involved and that quality is everyone's responsibility
In Analyze Phase
Identify root causes of the gap, those root causes need to account for the gap
Tools of variation -m
If I want to know where the highest and lowest points will be STANDARD DEVIATION Best and worst performance with in customer requirements use RANGE Want to know bottom 25% data or upper 25% where middle is QUARTILES How tight or lose data is a round the middle relates to sigma MEAN
Value Add Rule 1
Is critical to six sigma/ value has to change the form or function. hamburger from raw to cooked inedible to edible( transporting beef from farm to restaurant did not add value. factory two parts put together adds value, measuring them doesn't
Discrimination- A
Look for in the lower chart . It's having at least 6 possible units of measure be 0 and the upper control limit. If point out of control measure look at 6M's
Waste
Not adding value it is waste
Milestones -d
Not common to every charter; high level actions to put in place. follows DMAIC process s/b simple 1 page doc. each phase has 3 sub phases 15 tollgates all together, if anything changes modify charter
Two sample t test null and alternative hypotheses A
Null hypothesis (Ho): M1=M2 Alternative hypothesis (Ha) M1 does not = M2 reject the null hypothesis Rejecting the null hypothesis means that there is a signification different bt the two data sets being compared
Inputs
Observation /data collection/ brainstorming/ application of stats
mapping tools-d
Ovals- starting and stopping points Diamonds- decision making Rectangles -process steps Arrows- show direction process is moving
Sponsor Responsibilities
Provide Guidance Insight, suggestions & help /Keep eye on goal and direct team/Provide support in funding and staff/Approve tollgate/stop project
Pull System-Key Lean Tool
Pull sutems signal when it is okay for you to produce (Kanban Circle) Eliminate waste in a lean system by pulling things through to get flow
Median cal -m
Put the numbers in order; the middle score in a distribution; half the scores are above it and half are below it 1234567 4 is the Median if two number in the middle add and divide by 2. Use the Median if you have an outlier in the data set. real estate used one house can be 10x higher than rest mean unfair bias
Variable data
Quantitative in nature, what we measure, continuous data measured over a continuum can divide in 1/2 makes sense
After meeting
Report back to team about out come, follow up on sponsor requests asap
SMED, Single Minute Exchanges of Dies -I
SMED, Single Minute Exchanges of Dies -Theory and set of techniques that make it possible to perform equipment set ups and changeover operation in under ten minutes
Outputs of Market Reasearch -d
Sales forecast, Sensitivity analysis that shows how much a particular feature of a product or service relates to sales or profitable, advertising budget
Competing Critical to Quality characteristics -m
Sales people want fast and easy while management needs a lot of information
Multivoting
Series of votes used to arrive at group decision and assign priorities after brainstorming session, reduces long list of ideas
Pp & Ppk are potential studies
Shows performance past or present
Nominal Voting
Silent version of brainstorming, a method to reach consensus, allow quieter members to contribute with out pressure
SMART goals - D
Specific, Measurable, Attainable, Realistic, Timely
standard deviation for Xbar R and XmR Charts
Standard deviation from am XbarR chart s=Rbar/d2 For Standard deviation from an XmR chart s=mRbar/d2 Same method for the Xbar, R chart; only difference is that it uses the moving range or MRbar. Use Table of Factors for Control chars the look up the d2 value
SMED-Key Lean Tool
Stands for Single Minute Exchange of Dies
Expert Facilitator
Subject matter expert helps establish ground rules, boundaries to operate with in as a team, help work through conflict, trusts the process, uses tension as part of creative process
Six Sigma Culture
Successful six sigma is more than analysis of variance, design of experiment or regression analysis. If our culture is not six sigma ready the competent use of statistics will npt be enough for a projects success
SIPOC -m
Suppliers, Inputs, Process, Outputs, Customers start your search for customers. it will help you find a list of key internal and external customers. The map relates the customer to specific processes your project has targeted customer needs and requirements
Resistance
Termination of a project can lead to resistance from the team and management, so you need to be ready for that. The team might show some resistance due to anxiety- they are not really sure what's next for them. On the other side , management might show some resistances because they do not want to be seen as a failure.
Meeting Participants
The Green Belt/the sponsor/ Black Belt mentor/ Team members when appropriate, team -customer experts- keeps small
Mode Cal -m
The number that appears the most in a data set. 234555678 - 5 appears the most. 2,3,4,5,59,9,6,7, in this example 5 and 9 appear most called bimodal distribution of numbers ( 2 central tendency's)
dependent data
The out come of one effects the other
Foundations needed to eliminate waste
The quality improvement side and the speed improvement side
Non-Utilized Talent
Underemployed or over employed
FMEA Helps Prioritize Projects by -I
Understanding the main problems while also understanding the current controls Understanding the severity of the circumstances, along with how often the problems occur and how easy they are to detect it is extremely useful and relatable for everyone
UNISYS, Texas Instrument and ABB
Unisys and texas Instruments were some of the first to inquire about Motorolas mehtods. They went on to be the first to use it out side of Motorola and used it for primarily defects in manufacturing. Mikel Harry and a collegue Richard Schroeder formed a private consulting company which targeted business exc instead of quality prof. and worked with Asea Brown Boveri (ABB) who saw improvements
Key Lean Tools
Value Stream Mapping, Visual Management, SMED -Single Minute Exchange of Dies, Cellular Design, Pull Systems, Total Productive Maintenance, Poka-Yoke, Kaizen, Kaizen Events, Standard Work
alternative hypothesis A
What you are testing for or what you think might happen- court of law testing for guilt -verdict changes Improvements/Change - happen Labeled H1 or Ha
two sample t test-A
a statistical method used to compare the means of 2 groups of subjects. If p <0.05, the null is rejected and the means are different. Alpha level will equal .05 will to accept 5% risk of making a type 1 error. two tailed tests looking to see if the data falls far away from the mean to one side or far away from the mean to the other side. used if asking which two machines is faster? Which two training methods get better results Which two risk assessment formulas works best.- Considered Robust
VOC - Requirements -d
are features or characteristics that customers expect in the products and services, they can change, so products and process will also need to change in order to meet customer expectations Requirements wants fresh salad Requirements 3 terabytes of storage and auto back up Requirement A+ deliveries in sterile conditions with anxiety meds for nervousness
Critical to Quality Characteristics
are lean tools that directly address quality, cost and delivery
Sponsors -d
are more local to the process area that you targeted, and give permission to keep the project going- often sign off on tollgates; ex. shift supervisor
avoid group thinking
bad- one person very loud shares idea everyone follows
Levels -I
can be referred to as settings or conditions
Stem and Leaf Plot -m
clarify the shape of a distribution data is split into the stem. ex of heights 2.3; 2.5, 2.6, 3.2,3.6 data is split stem is to the left and leaf to the right 2/3,5,6 3/2,6 line down he center. whole number on left tenths on the right.; benefits concise representation of data, the ranges, min, max,gaps,clusters, and outliers are easily identified. Extremely large sets of data can be used sp data can be preserved
Interview /focus groups help get VOC -d
comp rep asks prescribed question to current or potential customers. Pros- non verbal cues body lang/ able to clarify questions/ receive fairly unfiltered responses Cons expensive, money and time , can't interview all customers, skilled interviewer, may have to hire, watch customer motives not always answer truthfully
Nominal Scale
consist of categorizes Names, Labels Telephone numbers, street address and social security numbers, can't arrange in any order, no arithmetic operations are performed posses no order or distance origin
CPR
consistent , predictable and repeatable
ANOVA Table-I
construct a one-way Using 3 difference machines collected 4 pcs of data on a particular feature on a part that came from each of these three machines. All recorded on table
Project charter D
contract terms and conditions keeps everyone on same page, states expectations clarifies project. Living document keeps everyone focused and on point with project. - Team lead makes sure everyone understands benefits
Value Add Rule 2
customer must be willing to pay for it- put $3.00 cheeseburger in a gold box, it's now $300.00 easy to transport - changed function but client does not want to pay for it. do not add value
Internal Customers -d
customers within your organization, not buying out put of process, but they'd received the output of your work. They have requirements and specification that need to be met, otherwise it will impair their ability to complete their portion of the process that will be bought by the external customers.
Descriptive Statistics
describe what data looks like
External Customers -d
direct external customers / customers who receive the output of your processes such as a product or service. They are the source of your revenue. Indirect external customers customers who do not receive or pay for the output of your services but have a vested interest in what you are doing. -ex govt. agency
Business case-D
every project charter starts with one, justifies why a project is needed at a certain time. 2-3 sentences with a sense of urgency- attached to money use data to create sense of urgency- short sweet to the point
Hypothesis testing: TYPE 1 and 2 Errors A
examine set of assumptions , understand assumptions and provide a level of precision also call significance level
standard order design -I
experiment design is going to consist of + and - . The standard order is orthogonal, meaning each of these factors are changed an equal and like number of times, it is balanced.
Performing
final stage team work together effectively to implement change ready to develop improvement- Green Belt should keep team focused, recognize improvements, share results with team and other stakeholders
First wave project
first one must be successful, pick one with high impact tied to org. key objectives,/ low performing process better to bring from 60%-80% than 90-95%/ high probability for success not impossible to solve/low impact on resources/potential cost savings/ customer satisfaction
Non rational subgroups -A
have many changes in regards the 6m's ex manufacturing a jet airliner Jets are made singularly and over time a sub process of making a jet maybe ration subgroups such as putting rivets on a jet
Mapping The process- d
helps determine where we are spending out time and what we are doing to create what the customer will label as value add.
Waiting
huge waste-not changing form or function when you wait for something. waiting in line , waiting for mail
Poka-Yoke (Error Proofing)-Key Lean Tool
idea of preventing defects from ever happening
CTQ trees- critical to quality trees
identify critical needs that your product has to meet/ identify quality drivers/ identity performance requirements
Standard work-Key Lean Tool
is the idea of having documented standard practices so that things are completed the same way ever time, no matter who's working on it.
Destructive vs. Non-Destructive production of data point
is whatever i'm measuring being destroyed in the process or not cost concern addressed
Scientific Method of experimentation-I
is where you manipulate one factor at a time to see it's effect on some desired result. That might be a larger is better characteristic, a smaller-is - better characteristic or maybe some nominal value you are shooting for. Problems -It is not balanced -No way to see the interactions between time and temperature, for instance Is their possibly an interaction bt time and temp? Probably, but here is not way to access that. -the assumption of holding these things constant is questionable bc in this case the temperature may change with normal variation over time
Restraining forces
lack of time , alot of travel, changing priorities
Critical to Quality Tree Example- CTQ -d
lookin for VOC- 1.waiting time too long in ER 2.what's the important issue- needs to be reduced 3. how do we deal with faster service(triage- initial screening critical patients go one way not critical another 4. CTQC - critical to quality characteristics how are we going to measure our success- hospitals per capita per person
Response (part of design experiment)-I
manipulation of factors at different levels to see their effect on some desired result
The Kano Model -d
measures customers needs. It identifies where we are meeting or not meeting customer needs. The data is categorized into three sections; the dissatisfiers, the satisfiers and the delighters. Dissatifsiers- basic requirements customer must have. Satisfiers-not static, variable requirements, customer will notice when you meet them and will need a certain number to keep them happy. Delighters- are the latent requirements not expected customer didn't even know they wanted or desired. not expected but you meet them and the customer is delighted. When creating a model take care of customers dissatifiers first then, satisfiers and lastly delighters, delighters wont matter with out the other two.
Active listener is not
merely eye contact or head nodding.. convey listening but may be superficial
Driving forces
more one on one meetings & increase management involvement with team
Levels of Measurement
nominal, ordinal, interval, ratio
Defects per unit (DPU)
number of defects discovered / number of units processed 4/2=2 or 4,5,8,2,1 (defects found) 20 divided by number of units processed 5 DPU= 4 total defects over total units produced; simple metric for assessing the cost of poor quality; doesn't account for varying complexity.
Quantitative data -m
numerical data divisions of actual values; Name/Order Interval Origin
Techniques for balancing participation
open ended questions toward quieter individuals may pull out some surprising information/break up into small groups/ make a quieter team member team lead & present findings/write down ideas, set quota for each person to submit a certain number of ideas.
t distribution or student t A
probability distribution used to estimate the population mean or test a claim about the population mean based on the fact that the population distribution is approximately normal and the population variance or standard deviation with one is unknown with small samples (<30) and unknown population variance. It is very common to have sample sizes less that 30 bc it is expensive to collect data Degrees of freedom = # of Observations or sample size - 1
Define Phase
problem is initiated, defined, goals and scope identified as well as customer, project charter is developed based on real life problem effects customer, business case is defined, benefits are determined by asking basic questions, Project management is determined including the plan, stakeholder, management support and teams. need final approval to move forward and ensure success (sponsor)
measures of dispersion (variability)-m
range, interquartile range, variation, central tendency, standard deviation
White Belt
received in one day or less formal SS training , awareness of methodologies and voc, for team members sponsor or champions/ entry level employees
Mitigation -I
reduces the seriousness of an issue from arising but does not eliminate it from happening
beta error A
rejecting a good hypothesis (guilty guy walks)- Type 2 error- failing to reject the null hypothesis, made significance area too small- can occur by chance alone or bad luck, but most often be of failing to reject the null hypothesis-make test stat smaller a beta error
Measure Charts -m
run charts- running record illustrate shifts & trends, like motion pictures, better than snap shot Histograms, frequency distributions , stem and left charts and box plots- like snap shot. However with out these control limits of variation it is impossible to know when to react and when to leave the process alone except for the shifts and trends. so that's why control charts are more associated with six sigma than run charts
Short Term fixes-A
short term fixes include Guards put in place in manufacturing lines . buckets put in place to eliminate waste. The problem still occurs with short term fixes bc measures are put in place to try and eliminate waste but prevention of the problem does not occur
customer requirements
should be clearly defined on the upper & lower specification limits VOC should span wider than to VOP voice of the process
Dashboard or monitoring system
should be established to allow process owners to measure and control the critical leading X's and lagging Y's on a continuous basis.
Example of Variable data-continuous data
speed, pressure, time, feed rates
Finding the test statistic -A
students take same exam before and after the course. What to see if test scores go up, get higher. improve the alternative hypothesis is test scores improved. (right tailed test, positive values) The null is that the test scores stayed the same
Ownership transfer
the control plan makes it easier to transfer ownership of the project and its deliverable.
Continual Improvement-Control phase
the relentless pursuit of perfection very important in six sigma, meaning continual improvement. Do not stop improving the process bc you closed the gap.. during DMAIC u identified a lot of potential issues and a lot of potential solutions, don't just let them go. quality is all about continual improvement. Now that you have closed the gap keep going. Make the process as good as you can make it.. Celebrate!!!!!
Variance
the square of the standard deviation
Determining bins for Histogram -m
too many data spread out too small ; too few data will be will stack up into one place. Make to see if shape is normal or not; start by point of data If less than or = to 50 use 5-7 51-99 use 6-10 100-250 use 7-12 more than 250 use 12-20 or use formula 2 k=1 raised k=number of bins n= sample size if n=40 how many times do we need to multiply 2 x power to get a number higher than 40? 6 times - 2x 1=2 ; 2x2=4;2x4=8;2x8=16;2x16=32;2x32=64 Add 1 to k to determine how many bins; 6+1=7 bins How large should each bin be; Largest point of data - the smallest point of data largest point of data 50 smallest 20, we have 6 points of data (bins) 30/6 = 5 if 4.2 round up to 5 can add a bin if needed but can't take away, not a science some trial and error needed
Bell curve too wide
too much dispersion in the process
Tally -m
tool to keep track of data -snap shot
Discrete data
u need about 100 points of data- 3x as much
latent needs -m
unspoken needs; customer may not know their needs; presumptions of your product or service may not match what the customer needs/requirements are
leading indicators -m
upstream in the process- example diet, calories, fat grams, exercise- leading indicator outputs that occur up stream in the process ( identifying these will help us identify the lagging indicators and get us started through the measurement phase)
future performance
use Cp and Cpk to determine process capability for future performance CP & Cpk are capability studies- estimates of the future of the process. With future performance, teams need to gather data. Four numbers are needed; 1. The average of the data (x bar) 2. An estimate of standard deviation (s^) ^ is over s; basically the same thing as standard deviation (s) 3. Upper specification limit (USL) 4. Lower specification limit(LSL) Important questions... Is data normally distributed? Is data consistent, predicable and repeatable (CPR) -Predictability is the most important
DPMO (defects per million opportunities)
use for more complex products or services formula DPMO=Total defects in sample/(total opportunities x 1,000,000) provides numbers that are easy to discuss and make it easier to compare different products or industries - Accounts for the opportunities for failure; allows an organization to compare products or services of varying complexity across the same process steps
Taguchi Signal to Noise Ratio-A
vital few vs the trivial many the vital few are the signal they are things that cause an output - X's or independent variables are the vital few The trivial many are the things that are happening that do not cause an impact. relates to Lean, get ride of the waste(Noise), trivial many and focus on the value added activities, the vital few Principal of Kaizen is small incremental, continuous improvement , everyday good change, what to focus on the on the signal, improvements and not on that noise around them. critical few is X's Y depends on X and X is the root causes Y is the out come(critical few)
Risk Priority Number (RPN)- I
(a rating used in FMEA to indicate the combined probability, severity, and undetectability of a failure mode)- multiply severity x occurrence x detection ex 5x3x3=45
null hypothesis A
(a statement or idea that can be falsified, or proved wrong) what you would expect to happen due to chance alone if nothing really influenced the process out of the ordinary. Includes all things not in the alternative hypothesis. When testing if you reject the null hypotheses you are forced to accept the alternative hypothesis. court of law presumed innocent is the null only reject innocents if evidence is provided that leads to accept his or her guilt No improvements, no change Labeled H0 Alpha by chance alone
Key Metrics Used to Monitor a Process
-Define effectiveness or success-this metric eliminate the quality issue in the beginning -Identify records that need to be monitored- The type of data collected needs to be capable of detecting the quality issue should it reoccur _ Identify the minimal sample size and determine the length of the observation period -verify that changes did not introduce new concerns. This metric has a direct relationship t the measure phase
Steps to create an improvement Plan
-Determine the True North -Break down plan into short term, medium term, and long-term solutions. -Determine leading and lagging indicators to ensure improvements you're making are getting desired results -Decide on a cadence of progress reviews with leadership, meaning choosing daily, weekly, or monthly reviews.
Too Much Methodology-scenario
-Do not become trapped in too much methodology -For long complex projects that aren't delivering results, do not sue six sigma tools just to use them -Only use the tools necessary to solve your problem. -Stay flexible -Spend the right amount of time on each tollgate Maintain the DMAIC structure Follow the steps Go through the 15 tollgates as intensely as appropriate while being flexible.
Signs that defects/out of control condition is occurring in your process
-There are out of control points on your control chart -Nonrandom behaviors are present There are failures with in the check sheet; If any of these are found the response plan gives you a way to handle the defects and track them so you can find where they actually started to happen in the process, even thou you many not fin them until the very end of the process
What is Included in a response plan/
-Who will be taking action if a defect found -What kind of action(s) will be taken -Be very specific bc there could be some ambiguity in some of the terms. -create operations definition for any terms that may have different interprtations -A schedule that explains the timeline for investigation and implementation of the corrective action -You want to investigate to make sure the right corrective action is chosen and implemented -Any troubleshooting procedures and best practice approaches to correcting the issue
The Purpose of Lean
-focus on eliminating waste and creating flow (root cause is waste) -In order to increase flow , you should decrease cost bc time is money -The longer an item stays with you, the more it costs
Monitoring Phase
-purpose is to perform checks in order to ensure that a new process is preformed as designed -Metrics will define the effectiveness of the change to verify that the implemented changes resolve the issue
The Control Plan
-written summary -describes the systems used to monitor and control our, Processes, products, Service quality -formal record of approaches used to meet project goals. -defines everything that needs to happen. -keeps everyone informed as to what is going on in the project. -This plan starts with the define phase and flows throughout the DMAIC process
2 Rules for Control Charts-A
1) Never put specification limits on a control chart; If a point is above the control limit bu below the spec limit then the operator is likely to ignore it bc it's in spec. 2) Never construct control charts from inspection records; Inspection records are what happened in the past they are not thing that you can change now
Control Charts -A
1) gather and record data in order of production 2) calculate subgroup averages and ranges, collect data sets over a period of time ( any time manufacturing over 1 hour, how many?) subtract the nighest and lowest data value in each data set to get a range 3) Calculate control limits , averages are call X-bars, averages of averages called x double bar If your average is 275.8 and you have 12 subgroups, then divide 275.8 by 12 to get an X bar of 23 The average ranges is called the R-bar If your ranges total 46 and you have 12 subgroups, divide 46 by 12= 3.83 Upper control limit - multi the A2 x R bar then add Xbar To find the Lower Control Limit multi A2 x Rbar then subtract Xbar 4) plot the control limits - two parts to plot on control chart, upper part monitors centering ; lower part monitor dispersion 5) Plot the connect the points for the upper chart. where there is an X place a point until there are a string of dots representing time from left to right Plot and connect the points for the lower chart 6) Act on what the chart tells you.. trends and shifts don't happen by chance but bc something has changed with the process. A Trend is seven points in a row in an upward or downward direction A Shift if seven points above or below the average Look for Discrimination in the lower chart . Discrimination is having at least 6 possible units of measure be 0 and the upper control limit. If point out of control measure look at 6M's Act Now!!!
Key concepts of Lean -I
1)Understanding the VOC -Define value from customers perspective -Producing goods and services at a rate that reflects customer demand is called takt time -Create flow production where products flow in a continuous, uninterrupted movement -reduce strategic inventory 2) Enabling pull material systems -Meet the flow of production by replacing only what has been used -Reduce waste associated with excess inventory 3) Continually eliminating waste -Reduce waste activities that do not change the form, fit or function of the product. -Focus on value adding activities -Identify and remove waste 4)Value Stream Mapping -Paper and paper representation of all activities, both value-add and non value added, required to deliver a product or service to the customer -Benefits include; visualizing more that a single process level -Visualizing the flow of material and information -Enabling everyone to see the waste and sources of waste -Providing a system level view of the process from the customers perspective -Establishing the foundation for improvement activities that yields bottom line results -Allowing prioritization of improvement ideas with highest impact 5) Organizing the 5 S's -Sort is elimination of unneeded materials -Set in order is where needed materials are organized -Shine is when work areas are cleaned and inspected -Standardize is writing and adhering to standards -Sustain is where standards are regularly applied in all areas. 6) Utilizing SMED, Single Minute Exchanges of Dies -Theory and set of techniques that make it possible to perform equipment set ups and changeover operation in under ten minutes
Tollgates of Improve
1- Generate possible solutions -Generate a large number of possible solutions 2. Select Solutions - Utilize various tools to select one or more ; a) Team based or statistical tools b) Design an experiment to select the optimal solution from a few similar ones 3.-Implement Solutions -This phase forces you to choose from options to get the best solution -Evaluate solutions based on criteria and data -Verify that solutions will work with a pilot test -Develop a plan to execute solutions- close gap
Fishbone diagram steps
1. Determine the effect or problem. 2. Determine major categories of root cause. The 6M's / 8 forms of waste, DOWNTIME, The steps in a process. 3. Conduct brainstorming event. 4 Prioritize root causes
Calculating DPMO cal
1. Determine the number of defects (D) D=150 Gold Wedding bands 2. Determine the number of units processed or services offered (U) U=10,000 3. Determine opportunities per unit (OP) OP=3 (size, purity, weight) 4. Calculate the total opportunities (TOP) TOP= 3x10,000=30,000 5. Divide the total defects by the total opportunities (DPO) DPO= 150/30,000= 0.005 6. Multiply that answer by one million DPMO= 0.005 x 1,000,000= 5,000 defects per million
Improve Tollgate-GSI
1. Generate possible solutions, review all possible solutions to each of the critical root causes, need option in order to choose best solution. 2. Select solutions- select the best solution, based on criteria important to the customer, company project, decision should be data driven. 3. Implement the solution- solutions are put in place, implement plan drives the execution.
Green Belt common mistakes
1. Having preconceived assumption of solution -be open minded and go through the steps use tools to determine best solution 2. Having poor definition of problem statement -Use smart acronym( specific, measure, attainable, relevant and time bound.) 3. Lack of sponsor or champion buy in -if sponsor doesn't by in the team won't move -forward 4. excluding key function from a team. -teams should be cross functional include front line employees who work with process everyday
CTQC- critical to quality characteristics -d
1. Measure the number of corrective and preventive actions that we put in place- quantitative method 2. For preventative actions, details which are proactive and reactive. reactive - result of finding a problem and fixing an issue. Proactive- potential for an issue or a defect or a problem where you put in place a preventive action before problem occurs
Control Tollgates MRC
1. Monitor and control plan- cont. to measure performance, ensure the new process and procedures are being done correctly, avoid slipping back to doing things the old way. 2. Response plan-create plans to address potential problems, remember if it can go wrong, it will go wrong 3. Continual Improvement plan- CI is the core concept of quality, develop plans to continue improving the process over time, do not stop if u do competition will pass u
Measure Phase Three tollgates -m
1. Performance indicator identification- what do we need to collect data on . 2. Data collection Plan- plan to gather meaningful data/what is meaningful as determined by performance indicators/ 3. Base line performance measurements- before pic/ allows you to show management progress
Analyze Tollgate-PNC
1. Potential root cause-identify anything that may cause the problem, don't rule anything out.; generate ideas keep in case u need to revisit 2. Narrow root causes- identify the root causes that are likely the largest contributors to the problem. Can't fix everything. ; use ream decision tools or graphs ex Pareto chart, reduce possible causes, keep in case u need them 3. Critical root cause-verify which root causes will make a difference when reduced or eliminated, these we will focus on;identify significant few causes and verify to be the source of problem Not just symptoms but actual causes, use stats. addressing these will close the gap - at end of this phase you can still expand the scope revisit measure phase if data is wrong, still living document , don't have enough data to close the gap. Can continue to go back and update as you discover new things
Define -Tollgates-PCM
1. Project Charter-Provides business case, problem statement, scope of project, critical foundation step of whole process- why should management support project 2. Customer needs and requirements- define exactly what customer wants; based on VOC and use specifics and detailed terms 3. Map the process- specify steps for beginning middle and end
Two types of replication-of experiments-I
1. Repeat when you run an experiment once and then you get an 8 and you run the exact same experiment again right away and you get a different result. 2. Replicate - BETTER- is where you go through all of the eight experimental runs or experimental trains first and then you go do the same thing again and you get a different result
Process Steps for Design of Experiments-I
1. Select a process for improvement 2, Select factors - be careful pick two few and you may miss the key factors contributing to the problem. pick too many and the cost goes up quickly. Recommended is the run at just two levels (3 Factors 2x2x2=8) not more than 4 factors (2x2x2x2=16) 3. Choose the levels If choose a level too close to current setting not good. If too far apart from the current setting not good either.( n-1 talking about degrees of freedom)
Measure Phase 3 Tollgates
1. Selected process performance Indicators a .know what you need to measure b.may process improvements struggle or fail bc they aren't measuring the right thing. 2, Create a Plan to Collect Data 1. you have documentation 2. Considered many factors about data including Type of data, amount need, how you'll collect it You've created operational definitions ensure everyone on the same page as to what is being measured and how 3. You got baseline performance measures critical step of the measure phase you have a measure, a target and a gap- know where process stands May have to calculate sigma
FMEA (Failure Mode and Effects Analysis) sequential steps used to make improvements. -I
1. define potential failure modes 2. Define the potential mechanisms of failure 3. Define potential effects of the failure 4. Define the severity of the failure 5. Define the probability of occurrence of failure 6. Define the ease of detecting the failure 7. Multiply the severity , occurrence, and detection to calculate a risk priority number (RPN) 8. Define the recommended action to take
Data Collection Tollgate -measure phase
1.Need a plan that tells the sponsor exactly what data you are going to collect, how you're going to collect it and how much you need. 2. An estimate of cost and time that it will take to go get the data. 3. Give data collection plan to the sponsor a head of time. Be ready to clearly explain in detail 4. be prepared with a back up plan in case the sponsor thinks this one is too expensive.or time consuming.
Measure Tollgates PDB
1.Performance indicator identification- exactly what needs to be measured, focus on leading measures and upstream in process, get right the first time. 2.Data collection plan- clarify how data will be collected, clear definitions and procedures 3. Baseline performance measurement, evaluate how the company is doing currently, determine the importance of the project.
Three ways to test Solutions -I
1.Piloting- Implementing a live solution in controlled environment 2. Concept testing- gaining consumer response to the solution 3. Working with models/pro types -,implementing the exact solution on a smaller scale
How many DMAIC tollgates?
15 three in each Stage/phase of DMAIC
Vilfredo Pareto
1800's Italian economist who developed the Pareto chart from the Pareto principal . it identifies the separation of the significant few, the 20%, from the trivial many, 80%.Excellent way to show large amounts of data through graphs. And shows imbalance of a process Used for economic analysis to measure wealth distribution but was adopted for fields such as sales, technology and project management
General Electric
1995 CEO Jack Welch committed to Six Sigma, his implementation of six sigma surpassed it's original usage as that of a defect removal system. They formalized Six Sigma as an expectation of every employee. Above all else they created a Six Sigma Culture. It became a condition of promotion. The goals was to become a team lead, referred to as a black belt after serving as a black belt for 2-3 years, you could then be promoted into a leader position. GE built on Mototolas methodology by adding a "D" for Define. DMAIC. they focused on team dynamics and managing change. This was the biggest contribution GE provided to the cultural and business phenomenon of Six Sigma
Yellow Belt
2-3 days of training help with basic tools of brainstorming process mapping, cannot manage a project on their own for team members sponsors or co facilitators, entry level employees
Closing Out a Project
3 -termination options -Extinction - When the project mission is either a success or a failure, it goes away because it is done. I has either accomplished it's mission or it hasn't. -Inclusion- The team and all of it's resources are absorbed into anew operational unit -Integration- All of the deliverables are integrated into the carious units of the parent organization
DMAIC 15 Tollgates DMAIC Tollgates
5 phases in DMAIC each effects the next as do tollgates. Each feeds the next and they build on each other Therefore you can never skip a tollgate (step). LLS recommends 15 tollgates 3 for each phase. Each needs to be completed before moving on to the next one. * Define -1. Project Charter-Provides business case, problem statement, scope of project, critical foundation step of whole process- why should management support project 2. Customer needs and requirements- define exactly what customer wants; based on VOC and use specifics and detailed terms 3. Map the process- specify steps for beginning middle and end.. *Measure 1.Performance indicator identification- exactly what needs to be measured, focus on leading measures and upstream in process, get right the first time. 2.Data collection plan- clarify how data will be collected, clear definitions and procedures 3. Baseline performance measurement, evaluate how the company is doing currently, determine the importance of the project. * Analyze 1. Potential root cause-identify anything that may cause the problem, don't rule anything out.; generate ideas keep in case u need to revisit 2. Narrow root causes- identify the root causes that are likely the largest contributors to the problem. Can't fix everything. ; use ream decision tools or graphs ex Pareto chart, reduce possible causes, keep in case u need them 3. Critical root cause-verify which root causes will make a difference when reduced or eliminated, these we will focus on;identify significant few causes and verify to be the source of problem Not just symptoms but actual causes, use stats. addressing these will close the gap - at end of this phase you can still expand the scope revisit measure phase if data is wrong, still living document , don't have enough data to close the gap. Can continue to go back and update as you discover new things. *Improve 1. Generate possible solutions, review all possible solutions to each of the critical root causes, need option in order to choose best solution. 2. Select solutions- select the best solution, based on criteria important to the customer, company project, decision should be data driven. 3. Implement the solution- solutions are put in place, implement plan drives the execution. *Control -1. Monitor and control plan- cont. to measure performance, ensure the new process and procedures are being done correctly, avoid slipping back to doing things the old way. 2. Response plan-create plans to address potential problems, remember if it can go wrong, it will go wrong 3. Continual Improvement plan- CI is the core concept of quality, develop plans to continue improving the process over time, do not stop if u do competition will pass u
degrees of freedom-I
=(Number of Rows-1)x(number of Columns -1)
KPI (Key Performance Indicator) -m
A financial or non-financial measure that is used to define and assess progress toward specific organizational goals and typically is tied to an organization's strategy and business stakeholders. A KPI should not be contradictory to other departmental or strategic business unit performance measures.
Master Black Belt
A full-time Six Sigma expert who is "responsible for Six Sigma strategy, training, mentoring, deployment and results." for an entire business working with organization to create a five year strategic plan. Uses tools to help blk and green belts to deploy plan, coach and mentor blk belts through cert process
Control Tollgates
A lot of times teams make great improvements and then they start to slip, tollgate of control helps prevent that; 1. Monitor and control plan 2.Response plan 3. Continual improvement plan
Six Sigma
A measure of quality that strives for near perfection. Improving Processes- engaging in war on defects and slow- It's a discipline, data driven approach and methodology for eliminating defects. Reducing companies number of defects to 3.4 defects for every million opportunities began 1985
z distribution A
A normal distribution with a mean of 0 and a standard deviation of 1.
Natural forces against balance participation
A personality - very vocal , competitive and want to speak- B personality are more analytical, quieter and less likely to speak up
hypothesis testing A
A scientific process for examining assumptions provides framework against which you will test those assumptions Provides the level of precision you will need in your results to draw certain conclusions Includes hypothesis statement , the assumptions you begin with and what you r testing for
Performance Indicators PI or KPI -m
A set of measures that helps a company determine if it is reaching its performance and operational objectives- 3 ways in which a project team can gather needs and requirements . Direct, Indirect and third-party
Eliminating Potential Root Causes_A
Allow team to rule out obvious potential root causes. Consider using some graphical tools to eliminate more root causes Use a priority matrix on the cause that are left Set up some criteria Define an impact or importance rating Score each possible cause Ask team what will eliminate or elevate the possible cause Elimination technique will vary for each Select the tool(s) that will eliminate or elevate the possible cause There is no one way to eliminate potential root causes, no rules. many have more than one where each contributes an equal amount to the problem.
Paired t-test =A
A test designed to determine the statistical difference between two groups' means where the participants in each group are either the same or matched pairs. It's important to find the critical paired t value when doing this test
ratio scale data
A true zero origin exists ranked and serve as that which all arithmetic operations including division can be performed arranged by order, distance and origin. Example: money spent, miles traveled, number of kids in household start with 0 then go from there with 0 have none of the traits. speed of a jet at 0 not fling, crime rates distance bt rates and production rates
Lean - I
About creating value for the customer while utilizing fewer resources . customers value is critical and minimizing waste is important
Cost of Poor Quality- Internal Failures
Accidents, work arounds, quick fixes, scrap, rework, defects, mistakes employee turnover, equipment downtime, sorting
7M tools
Affinity diagram/tree diagram/Interrelations hip digraphs/PDPC networks/Cause & effect matrices/Prioritization matrices/Activity network diagram
Narrow Root causes-A
After gather potential root cause need to narrow down the critical few, or one, Make sure you keep a list of all potential root causes for later in case you need to go back, so you don't have to start over.
Improve Phase
All potential solutions are evaluated through data-driven exercises so that assumptions are mitigated, error proofing exercises are put into place to eliminate accidents , risks are assessed, failure modes and effects analysis is performed, pilot plans are conducted to ensure proper roll out before implementation occurs, implementation occurs, validation of stat methods takes place-generation of solutions occurs-best solutions selected based on critrea of business feasibility
Strength of six sigma
Clarity Purposeful Systematic Logical Ensures that you have what you need to be successful
Design of Experiments -I
Another way to do experiments other than one factor at a time. It is much more powerful than just waiting for a change to happen. two things account for 60% of problem, that's how powerful. -like fishing wait for fish to bite( scientific method )or throw a stick of dynamite in the water, design of experiments Testing three things for sleeping, 3 factors, ABC; A is medication 2 types, B gender Male & Female, C age 30 and over 30. set up experiment to see when to see combinations that are best for these factors. there are 3 factors with 2 levels each; Then you manipulate the factors at different levels to see their effect on some desired response. Then run 8 experiments or trails or experiment conditions. Use a standard order design so it is going to consist of + and - . The standard order is orthogonal, meaning each of these factors are changed an equal and like number of times, it is balanced. Look for patterns ex high low high low. +-+-+- also look at the different types of interaction, AB interaction, AC interaction BC and even ABC. We got different results for all the columns-each one is unique this is called full factorial designed experiment
Problem Statement- D
Answers the question what is the primary issue my team will focus on. Identity the primary project Y, Identify the process output that is not living up to expectation identity the secondary project Y's. secondary Ys are fixed by default when you fix the primary Ys. Y's are costs, gaps or something that brings pain. deliver the monetary pain you will get the by in; 2/3 sentences state problem and what current process is.
Performance Indicators PI - m
Ask : are those things measurable?/ effects on customer?/what are customers requirements?/are there leading indicators?
Data Collection Plan -m
Ask: what type of measure r we looking for/what type of data/ Operational definitions anything the team can identify and relate to things to measure- starting point stopping point and target etc , create plan and gather only data that is needed
Control Phase Results
At end of control phase you should have the following; -A plan for monitoring the process to make sure it works well (continue to measure) -Some new procedures and processes in place to it running well -A list of things you did n[t address in the analyze and improve phase. -A continual improvement plan to make sure that the process continues to improve in performance The critical part of the control phase it that you are watching the process and are not going to let it get out of control again. However if it does you already thought of what you need to do to fix it.
6 M's of Fishbone Diagram
Bad flight Man (mechanic not trained)/Machine(equipment failed)/Method(pre-flight list confusing)/Material (Fuel)/Measurements(pilots watch didn't work correctly)/Mother nature(fog) - when mapping can also be used to determine sources of variation how each one affects the process
Control Charts
Best way to monitor
Champions and Sponsors
Called different things in different places, mean same thing/mid level or senior leaders/authority to deploy resources and remove barrier/responsible for chartering and selecting the process owner/makes regular contact with blk belt and green belts and process owner during improvement projects; when team ready to move from one tollgate to another/reviews the results from the team then approves or makes course correction for moving forward
The six blocker (Charter) -d
Can present a charter on one page; Business case Problem Statement;scope; goals; milestones; team; can also have financial goals , breakdowns for the business case , project plan with all 15 tollgates, list of stake holders beyond the team members if you would like to add a 7th blocker it would be constraints
Cellular Design-Key Lean Tool
Cellular design allows u to get things close together and to be able to follow one step to the other with our piles of inventory in between . -It helps reduce transportation waste and waste of inventory by more quickly identifying potential defects
Customers and the IPO Model (Inputs Process Outputs) -d
Customers can be the next person working on a product or service , Band -Set up company the band is there customer. IPO Model Band playing to the clients Band is the Input; Playing is the Process; Audience is the Output. the output leads to another input.
Defect Location Check Sheet -m
Collects data visually rather than with words , also know as defect map or measles chart used for simple count is needed to determine the defects. Place a visual cue such as an orange dot on an image (man with orange dots on him) Common in quality assurance Benefits -collects data with minimal effort , data can convert into very useful information weakness not being able to act on amount of error found. regular check sheet can keep track of number of error, but not their locations
In the improve phase
Come up with improvements that will close all of the gaps
Timeline
Common mistake when developing is being too aggressive , you have to find a balance if you carry the project too only you can lose momentum and team members to other projects, can't make too short will sent them up for failure. Must find a balance
complementary
Complementary events are the result has to be one or the other, the die example, the die will be wither even or odd, can't be both., has to be one or the other.
budget
Complexity of the budget depends on the nature of the solution , how large it is. The champion should work with the Green belt or Black belt to determine the budget upfront. often times it minimal or nothing and is really the level of effort of the individual on the team. however, larger improvement with large return on investments may be an internal rate of return that the company looks for. it depends on what the company is looking for.
The value of process capability, especially using Cp and Cpk includes;
Confidence that the process is performing well Ability to predict an outcome Better ability to predict an out come Better ability to communicate with customers Developing realistic expectations Establishing realistic expectations Establishing pricing for a product or service Better ability to plan The most valuable part of the process capability is that it is part of the Control phase of DMAIC. Cp and Cpk are great metrics to monitor the process
Variable Data -m
Continuous measurements- distance, heights, time can divide by 2, can't do that to variable or discrete data can divide people in half
Final Project Report
Control plans are an essential portion of the final project report. -It complies what you need to give to the process owners; it is their guiding documentation to keep the project going. -It is a reference to keep them on track It is a source of lessons learned and guiding information
Active Listening
Convey Clear message/ ask open ended questions/no one word answers yes, no/give personal responses. Read body language of audience. Falling asleep,looking down
Cost of Poor Quality- External Failures
Customer complaints, Penalties, liability suits, repair costs, time spent evaluation things gone wrong and responding to upset customer
oridinal scale
Days of the week, order but no defined differences, Friday, Thursday, Wednesday, Tuesday, Monday
DPU vs, RTY
Defects per unit how many defects exist in a given unit. multiple defects can exist on one unit RTY Roll throughput yield only considers that a unit was defective regardless of how many defects existed (a unit with one defect has the same impact on RTY as a unit with 10 defects) they share the same drawbacks -they don't consider complexity of the product or service being analyzed, only using theses two may cause an organization to choose the wrong priority when trying to deploy resources for improvement
Conclusion
Define Phase learned- Project Charter, customer needs and requirements and process mapping. -Measure Phase - introduced to identifying the right thing to measure, a plan to collect data and taking that before picture, baseline process performance. -Analyze Phase - how to come up with a long list of possible root causes, narrow that list down to a short list and then isolate it to one or two root causes, then arriving at the critical root cause you discovered how to improve that condition allot which took place in the Improve Phase. Finally once the process was improved upon, you learned how to keep it that way. that was in the final Phase of control.
How to construct and Pareto Chart -m
Define problem- 1. use SMART goals- specific, measurable, attainable, relevant, and time bound. 2. Determine the time frame that you want to analyze. 3. Chose characteristics to analyse - should lead to a conclusion that is actionable. 4. Sum up and sort the characteristics- addup the occurrences of each characteristic and sort them from largest to smallest. 5. Create a bar chart of the occurrences. Bar chart must pass a common sense test. Is the data relevant to the problem and is it actionable?
Steps for Hypothesis Testing A
Define situation or problem you want to examine State your null and alternative hypotheses State the level of significance-precision Select the proper tools for analysis collect data suing that toool calculate your test static look up your critical value Compare your test statistic ro your critical value Make a conclusion
What does DMAIC stand for
Define, Measure, Analyze, Improve, Control
Control Phase
Determine if all the previous work was successful -Measure how well improvements did -If improvements were not successful, you can always revisit -Implement Implement Implement -Get it done -a good plan executed today is always better than the perfect plan never executed - a bad plan should never be executed -Take any information that suggests your plan isn't going to work to your sponsor -Don't throw away any solution, even if you don't implement them
Tollgate Review Meetings
Different Organizations hold tollgate review meetings differently, some org. hold one at the end of each phase. Villanova LLS recommends having 15 smaller , individual meeting one for each individual tollgate. Here's why- 1. sponsors ( part of managment) review each step.(prevents rework) 2. Keeps information focused -small amounts at a time. 3. Easier to find time for sponsor. meeting 15min instead of an hour. 4. Keeps sponsor actively engaged. meeting more frequently.
Range -m
Distance between highest set of data and lowest set of data. fewest defects 2 the highest 10 the distance bt is 8 that's the range, best performance 2 worst 10 defects- works with two points of data, immediately shows highest and lowest point in data set, fast and easy to calculate, does not show where data piles up around middle , if shape important range not best choice
Tollgate Review Continual Improvement Plan
End of project - You and team identified root causes and generated a number of possible solutions; The team identified a few root causes and came up with solutions to eliminate them. What about ones left one the table? Should the team stop here? Continuous improvement is the core value, it's what process improvement is all about -Improve , improve, improve....Yes the team stopped immediate problem, but if they stop now, new ones will pop up or worse your competitors who are continuously improving will surpass your organization. So it's imperative to continue to improve processes. So the sponsor wants to see; -A prioritized list of future improvements that you are going to make- there are a lot to choose from you already have a list of potential solutions ready to go, so you can try out some of those things. -there are other root causes out there impacting the process; they just weren't impacting it enough to be called critical. So go out there and fix those. -Some dates and goals for thing you are going to do to keep continually improving the process. -Project is over so communication the results - let everyone know - Review lessons learned- DMAIC gain experience for next one
Cost of Poor Quality Prevention Costs
Error Proofing, Capability studies, Procedure writing, applicant screening, education and training
Inference Space & 6M's-I
Example with loading cargo onto a jet -Man when looking at the load levels , is it just one person or should it involve many people? -Machine was that just one jet or should it involve many jets? -Method was it just one method, or are there methods that spill across many, many jetliners within the same airline?(use one Method) -Material was it just one load type? gold bars or ping pong balls? one very heavy, one light -Mother Nature was it on just one day , maybe one climate, in the summer, or should it include what it is like in the winter or when it is rainy? If you wanted a really good standard deviation you would only involve one of everything. However, to get a really good understanding of the process and what it's capable of you need more than one day, one load type, one scale, one person. All of that is called an inference space. Narrow Inference space is one of each of the 6M's Wider load space includes more of the 6 M's Have one method but include more variations of the rest.
What Control Plan does
Explains; What we need to do to sustain our solutions -which variables we need to keep an eye on and monitor -when to respond/what triggers a response to action
Fredrick Winslow Taylor
First scholar of modern business first to introduce 3 core principles of project management and leadership- bench marking, process redesign and work out.
Ratio Data
Fixed starting point, temperature, distance or weight
Motion
Form of waste if you spend more time looking for an object to do your job, you spend less time doing your job.
Common Project planning process and tool
Gantt chart - forces you to think of end goal, what future state look like, which the team has already decided and what are the steps to get there? and you work your way back. You also identity, know what things have to happens before something else can happen
Value Add Rule 3
Get it right the first time- ER run 10 lab tests in an hour but did 5 incorrectly. now have to pay twice as much to get those 5 corrected also caused back up in ER
Define Tollgate Review-Project Charter -d
Green Belt should Follow agenda, provide sponsor with clear concise and complete charter. should be sent to sponsor a few days before so they can review. Present a compelling business case that shows why this needs to be a six sigma project and what will happen if not done $ effect. provide clear problem statement well defined scope statement. green belt may be missing some information, it's ok this one time , let sponsor know what you need to tighten up charter
Dr. Kaoru Ishkawa
He deveoloped the Ishkawa diagram= fishbone diagram. He proposed to achieved complete immersion in a quality control moevemnt a commiment of 2-5 years woudl be required from the company and it's employees. He refereed to this as CWQC companywide quality control./ orignated quality circles which used PDCA cycle of Plan do check and adjust. He used qulaity control adudits to discover oppotunites for imrpovemnt OFI. s . He used stats and believed a companys problem could be solved using 7 basic tools. Flow charts,/ Ishikawa (fishbone) diagrams/charts/Oareti cgarts/histograms/scatter-grams/control charts
green belt facilitator
Helps in forming storming norming and performing cycle. Sets expectation, keeps team energized, not discouraged and leads team to performing. corrects team if they r3vert back to previous stage
Voice of the Process -d
Helps us understand what's going on in the process- helps u understand what went wrong and why the other voices are complaining VOB 9voice of business) & VOC; considers the activies with in the process , the order they occur in, level of quality, cycle times and more.
SIPOC chart-d
High level chart used to capture VOP ( voice of the process); shows who is involved; what steps are, and the relationship between the inputs and outputs. Suppliers, Input, Process, Output, Customer
Dr. Genichi Taguchi
His contributions are, loss of function/theory on off line quality control / innovations on statistical design of experiments.
Level 2 map- process map-d
Identifies the process steps from the workers point of view., gives one a good idea of the "as is" process. see workarounds and beyond the "as designed.
Force Field Analysis
Identify change or potential solution/brainstorm retaining & driving forces; forces pushing & pull you away from certain actions. can be useful in trying to develop a charter; what you should and shouldn't do
Customer needs & requirements -Define Tollgate
Identify who the customer is
Let your tools guide you-scenario
If project high risk, complex and has significant cost, use as many six sigma tools as needed -Let the tools answer the question you are asking of your process Keep you champion and black belt mentors well informed
Extra Processing
If you inspect a product three times and it was perfect the second time, then you are wasting your time, not adding value to the product
two numbers that describe a normal curve are -m
Mean & Standard deviation Mean- average, middle of the curve, has equal amounts of data to right and left standard deviation - quantifies the variability of the curve. how broad or how ;narrow the curve is
Why you Monitor more than the Project Y
If you only monitor the Y can lead to creation of a defect, which when out of control creates more defects and is costly; Monitoring the input/process level (X) is more robust than just doing so at the out put level (Y) If the X can be caught as it is going out of control, there may not be a negative impact on the Y -Creates an ongoing action place for process participants to follow -Ensures a timely response to processing problems, decreasing the risk of defects
Piloting solutions -I
Implementing a live solution in controlled environment; Reasons; Confirming expected results -evaluating variations -working out the bugs in a solution -Identifying potential problems -minimizing risk-gaps in plans -uncovering additional improvement opportunities -identifying any gas in implementation or control plans WHEN to use- The scope of the solution is very large -Full deployment failure will result in large cost to the company( possible far reaching or unintended consequences, blow something up on a small scale or injure someone)
balancing participation
Increases engagement more people involved better results more people engaged when solution put in place
qualitative data -m
Information describing color, odor, shape, or some other physical characteristic 6 M's source of data Man Machine Material Method Measurement MotherNature
The Mean for Xbar R and XmR Charts
Mean for Xbar R chart is x dble bar= Xbar's/n Mean for XmR chart is xbar=x's/n Same method for the Xbar, R chart; only difference is that it uses the moving range or MRbar.
unhappy Customer -m
Measured by central tendency Your centering is off, your variation or dispersion is so wide it goes off to the high or low side
VOC 5 ways to obtain
Interviews & Focus groups - ask questions to current & Potential customers- may not always be truthful ;Surveys paper or electronic; may not be representative of sample; Market Research very expensive: Complaints and Kudos -reactive; Data Observation- very expensive
Green Belt Corse
Is first half of foundation of process improvement. It has focused on making a product better and more efficient . The second 1/2 is making a product faster by reducing waste, making the process more efficient,. Waste removal this component is found in the Lean Six Sigma course.
Pp cal
Is the VOC compared to the VOP VOC is the numerator, while the VOP is the denominator Pp = USL-LSL/6 x s use x bar =1.25; s=.05; USL= 1.32; LSL= 1.07 1.32-1.07/6 x .05 .25/.03 = .83
Define
Is the foundation for everything else; VOC NOT the VOB- 3 tollagates used to evaluate your entire project. Charter has your problem statement, your customer needs and requirements are your yard stick you measure against, process map reflects exactly where the changes your making are going to occur
Variation-A
Is the number one thing to eliminate when brainstorming a problem. Special cause variation must be reduced with root cause analysis. The problem must be understood thoroughly in order to ensure that a permanent fix is put in place instead of a quick fix
Data Sensitive- m
Mean is sensitive to extreme data, unlike median Interquartile range is not sensitive to extreme data Standard deviation is sensitive to extreme data
Measures of location
Measuring the mean median or mode of the data set also called central tendencies
Importance of Process Capability
Knowledge of process is determined by process capability; -Allows product development engineers to select the right equipment for a proposed design -Provides the ability to schedule jobs better with the right machines -Lowers work in process -Reduces stoppages -ensures that process problems are visible and quantifiable -provides a means to communicate with the customers - helps develop realistic tolerances based upon actual performance -Helps set pricing of products
philosophy of Six Sigma
Measure twice, cut once we do not want to keep fixing the process. We want to make a significant improvement and get it under control and operating the way it should to deliver the quality your customer needs.
Allied Signal
Lawrence Bossidy became CEO in 1991 after working for GE using six sigma as a business enabling tool. His idea was to improve operating margin . Each project was hand picked by an executive that was trained in six sigma. he believed that defect removal and cost reduction were just by products of improving the operating margin of a company. Every different than Motorola. His strategy had 3 core processes. The people process/ the strategy and the operations process. identify the process most broken and create a tailored project to have the highest impact on operating margin. They also linked Six Sigma to the customer.
Lean and Six Sigma
Lean- eliminate waste- lean when discussing waste focuses on the OWNTIME or DOWNTIME Six Sigma- is about efficiency, speed and the elimination of variation, when discussing waste six sigma focuses on Defects with is the D in Downtime
Lean and DMAIC
Leqan tools are found through out the DMAIC cycle, don't get too caught up in which DMAIC phase you are in, just do the right thing at the right time in order to improve. -Measure, Lean metrics are all about measuring -Analyze; analyze the system to see where there is waste -Improve many lean tools come into play during this phase Cellular design TMP total productive maintenance Kaizen events -Improve ; kaizen events and other lean tools are used during this phase. -Control ; know that everything is going to be the same every time and every where there is standard work Six Sigma -elimination of defects Lean- 0 time of processes
The Response Plan
Like fire escape for your process, it tells you where to go and what to do. It will be handed off to the to the process management team b/c they will need to know what to do and where to go if the process starts going off track Having a response plan is critical bc it outlines those processes that you need to follow when there is a defect or out of control condition that you are tacking in the monitoring plans
Continual Improvement Plan
Long term Plan that includes multiple improvement projects all aimed at a common goal. Keep team focused on short and long term goals and set the expectation for continually improving . Six sigma is all about continual Improvement , so you should never settle for the status quo or "it's good enough"
Project charter -Define Tollgate
Made up of 1.Business case describes how the project is good for business, impact on key objectives and why worth doing.non quantifiable 2. Problem statement more specific, how big is problem, how long it exists, gap bt now and where we want to be in the future. quantifiable 3. Scope Statement starting point and ending point what's in scope of project and out of scope. 6 blocker charter has Business case Problem Statement;scope; goals; milestones; team; can also
Defect location check sheets -m
Make a picture of the part and make a form out of it and give it to someone that is in inspection. Instead of counting number of defects put a spot of where they are. then you can go back to the process and ask what is happening to cause the cluster of defects in a certain spot.-snap shot
Black Belt skills
Managing multiple projects and multitasking/completing advanced stat. analyses/being an expert facilitator/communicating ideas and ensuring that individuals understand those ideas
Design of Experiments -I
Manipulation factors at different levels to see their effect on some desired response.
Level 3 map -micro process map or improvement view of a process map-d
Maximum amount of detail, searches for redundancies, workaround, and variations from the standard operation procedure. serves as the necessary stepping stone to arrive at he improved performance
lagging indicators
Y's The final check point
Response Plan Measuring more that the Project Y
Monitoring Y is intuitive bc it is the measure of success, but it tells you defects after the fact. If you only monitor the Y can lead to creation of a defect, which when out of control creates more defects and is costly; Monitoring the input/process level (X) is more robust than just doing so at the out put level (Y) If the X can be caught as it is going out of control, there may not be a negative impact on the Y -Creates an ongoing action place for process participants to follow -Ensures a timely response to processing problems, decreasing the risk of defects
People
Most important, drive change, when purchasing new material or training individuals you;ll need people to understand the process and buy into it to make changes. Biggest mistake is thinking that 100% of the teams time will be dedicated to the process/solution , it's actually 20% or less of the team
Key pitfalls to avoid when planning implementation
Most importantly is the availability and capacity of the individuals making the improvement is paramount in developing that implementation plan. You have to get the buy in from the resources, the people doing the work. You also have to get the buy in from the leadership.ensure they understand the time commitment- 3 months to complete the project and 5 hours a week from employees . Must have follow up meetings to discuss if you are on track or off track and why . also gives time to look at key measures and ensure that were getting the results that we want
Transportation
Moving things, changing location, but not changing products form or function
Goal Statement -D
Must link to the Problem statement. It is what fixes the problem, defines how much of the problem this project will fix. when defining how much you're going to fix, start low and work your way up. 50% and if you do more great - 3 key items here to there by when. keep goals short can add financial expectations here
nominal scale
Names only no order; color Yellow Green Blue Orange Red / we have least number of options with data, consists of categories ex. names, labels, telephone numbers, street signs, social security numbers, not arranged in any order or scheme , no arithmetic operations are performed, possesses no order distance or origin
Affinitizing-A
Narrow down root cause by grouping them , many may be similar or the same, allows only unique items to appear and removes duplicates Allow the team to rule out obvious items; team votes together on what to eliminate Allow team to make decisions, rely on their expertise. Don't waste time, make sure you challenge them and ask how they know, if seem legitimate, let it go.
Histogram -m
Normal distribution- In middle classic bell curve falls right down middle of customer specs. like to see bt red lines which are customer specs. Normal Distribution- off to one side normally shaped but not down the middle. with in customer specs but to one side, indicates that a defect may occur soon. Normal distribution Wide in the Middle- signifies a lot of variation in the process.if data gets larger it may be in trouble of falling outside of customer specs and then a defect has occurred Normal distribution- Too far to the Side is normally distributed but misses target. too far off. Defects have definitely occurred ND but too wide in middle - too much variation goes outside of customer specs, need to decreases variation in process. Skewed Distribution- not bell shaped piles up on one side something other than random variation is at work, misses the middle ,mean not in center Bimodal Distribution- typically data from two different processes mixed together. day or night shift; machine A vs machine B. Ex screws different sizes; Wait times too long or sometimes short won't go back don't know what to expect Truncated Distribution- missing left and right tails someone could have thrown away defective items, or data not collected properly. either way needs to be looked at. Missing the Middle - The tails are present middle is missing. Data is with in specs limit but tails are somewhat close to edge. Can happen when customer pays a high premium for something, wants high level of consistency and parts are pulled from center. Histogram represents left over parts Flat distribution- which indicates the exact same thing is happening over and over. typically doesn't occur unless someone makes it occur. even if data falls in specs. needs to be checked out.
Determinant of Quality -d
On product site Performance/Features/Conformance to guidelines or targets/Timeliness/Reliability/Serviceability
When can a project be closed out?
Once the achieved X's are identified and there are final improvements in the Y
Green Belt
Pivotal role in SS prjects, keeps team on task, facilitate ss tools required to accomplish all steps in DMAIC process, typically mentored by black belt coach for first 3 projects, dedicated part time to ss projects.
Analyze Phase
Potential causes of a problem are identified. Current processes are analyzed, relationships bt inputs, processes and outputs are identified, data analysis is carried out, waste and root causes are determined, root causes are verified to understand cause and effect.great deal of brainstorming , data analyzed to show trends, use multiple tools to judge accuracy of data from a statistical point of view
FMEA Priorities (Failure Mode and Effects Analysis) -I
Priorities are based on : 1-Repetitive issues that the company faces 2-Top down projects that management has initiated as improvement projects 3-Managements strategic vision and selection of priorities for the business 4- Innovation so that the business can grow
PDPC
Process Decision Program Charts-is a specific tree diagram, used to access risk, ideas identified through multi rounds of brain storming. similar to failure modes and FMEA tools
Quincunx -m
Process generator that shows variation in the process. red beads opportunities/ use 6M's to measure-represents randomness and opportunity for improvement
Define Tollgates
Project charter, customer needs and process mapping/ 1. Project charter, compelling business case/clear problem/define scope 2. Gathered VOC - who your customers are/their needs/ specific requirements 3.Mapping process- delivers good or services to your customers - SIPOC &/or more detailed flow chart
Interquartile Range (IQR) -m
Q3-Q1 shows the percentage of data that falls with in a certain area, USED With MEDIANS, not effected by extreme values in a data set. quartiles divide data into four equal parts or quarters. 25% of data in each quarter. IQR measures the distance bt Q3 and Q1 the middles 50% example wait time Q1 = 1 min and Q3=4 minutes 1-4=3 thats the IQR 4 steps the calculate 1. sort data highest to lowest 2. Find middle or median of data set if odd number find the middle number and don't use it 1234567 don't use 3 (if want the median just add two middle number numbers together divide by 2) If even numbers just divide in half 123|456 3. find the median of the upper half Q3=5 and the lower half Q1=2 - it's simply the middle number all rules of median apply 4. subtract Q1 from Q3 2-5=3 outlier does not effect data Measure of variation that goes with median/ not impacted by extreme data indicates where middle 50% of data is/ a Little more difficult to calculate only gives middle bend. Most useful with box plot
Attribute data(discrete data)
Qualitative in nature- only two out comes yes or no , pass or fail , a go or no go can be categorized in classifications, and we can then count how many show up in each one of the classes # of defects can't divide in 1/2 doesn't make sense
Joseph Juran
Quality Control Handbook, 1951 Viewed quality as movement and the tools associated with it. He proposed management should be present in qulaity councils or circles . beleived in SMART stating specific , measurable, achievable, realist and timely goals from start of project. Quality trilogy- quality planning, quality control, quality improvement
Formula to estimate standard deviation
Rbar/d2 or MRbar/2 MR = 2 Look at the table of factors for control chart to find d2; find d2 constant for subgrou size of 2; d2 =1.128 s^=2/1.128=1.8 see page 12 packet 8
Eight Types of waste 8 types of waste
Remember by using the word DOWNTIME-Defects, Over Production, waste, Non-utilized talent, transportation, Inventory, Motion, Extra Processing
Motorola
Six Sigma Originated here in 1985 Bill Smith wrote a memo about product field life and rework during manufacturing and called it 6 sigma. The Mikel Harry and a small group of Motorola engineers create a methodology called MAIC based on Smith initiative which stands for Measure, Analyze, Improve and Control. they used control charts, design of experiments and root cause analysis to analyse and remove defects. Harry also created Mototola University a primary Six Sigma learning center with in the company. Documented it in 1986. It focused on the removal of defect. Saved $ 15 millions from 1985 until early part of this century.Highest levels of the company supported it and placed undervalued personnel in cahrge of the projects.
Carl Friedrich Gauss
Six Sigma Originator -German mathematician from 19th century Europe introduced concept of normal curve metric as an error analysis tool for measurement , probability and hypothesis testing. one of six sigma corner stones
Process Mapping -scenario
Some companies already have sell established process maps and defined metrics, if this is the case you can move from the charter into root causing you already have the baseline performance
Tollgate Review- Implement Solutions -Improve
Sponsor looking for -An implementation plan for the solutions that you have selected - this should be a real, live physical document that outlines when the solution is going to be in place, what timeline is, what resources are , who is responsible , whether or not training and equipment are needed and if the stakeholder by inf from another department is necessary. I- if a stakeholder by in is necessary for implementation, let the sponsor know beforehand who those people are
Tollgate Review-Response Plan
Sponsor needs to know; -What things could go wrong with process? What could happen that could derail the performance of the new process? -A documented list of prioritized issues -Preparations for how to deal with those things when they happen -A failure mode and defects analysis (FMEA) would be an excellent thing to use to help you do that.
Tollgate Critical Root Causes -A last analyze tollgate
Sponsor wants Short list of critical root causes that will carry into improve phase not one hard to believe only one thing. Team believes that elimination of these root causes will close performance gaps . want critical few, not trial many How the team identified the critical root causes- use data driven tools to make final selection of critical root causes. An estimate of how much gap in performance cal be closed by eliminating each of these root causes. Hard copy of official list of critical root causes be prepared to visit prior phases
Last Tollgate of Measure Phase Baseline Performance Measurement
Sponsor wants ; 1. Charts and graphs with measurements, the target and how big the gap is. may involve conducting capability study by measuring process capability Pp and Ppk - make sure u can clearly explain. 2. Any necessary information to make final updates to the charter 3. An up-to-date version of the charter - Present before picture If sponsor doesn't like what they see may have to go back and re discuss data collection plan. Be ready to explain how accurate your data is and why
Tollgate 2- Analyze Phase- Narrow root cause-A
Sponsor wants Official list of the narrowed root causes; keep list bc it will be used later How the team identified the narrow root causes What tools the team used; ex, stat tools, graphical tools Pareto chart, scatter diagram... Data should drive the decision to eliminate potential root causes Any new information or updates that might change the scope of the project, the project charter
Define Tollgate Review VOC -2 Customer Needs and Requirement Tollgate
Sponsor wants a description of the customer group and specific segments that are being impacted by the problem identified in the charter. 2 what are the needs and requirements of the customer and what is not being met. provide some supporting documentation such as customer survey feedback, market reports, customer interviews whatever is appropriate . If anything u found impacted the tollgate let the sponsor know. or if the business case may have changed of. the scope of the project is expanding must tell sponsor
Tollgate Review- Monitor and Control Plan
Sponsor wants to see; -Performance Graphs and charts -A documented measurement plan -This lets the sponsor know that the process is being watched so it doesn't get out of control again -Cost saving and /or improvements in customer responses -The new process map give to the sponsor a few days before tollgate meeting -A control document - staff s/b using it shows sponsor that the standard operating procedures are being executed.
Tollgate review -Select Solutions -I
Sponsor wants; -Give your recommendations for a prioritized final list of solutions for each critical root cause. -How the team identified those solutions -The statistical tools used ex pilot test, design of experiments, etc -Cost effectiveness and resource estimate -These are required be the team is going to develop implementation plans for the solutions that are selected at this tollgate review.
Why You need a response plan
Systematic and quick-Sustained Gains-Paves way for improved processes-earlier response time for incidents designed to get things done in a systematic and quick way so that you don' have to redo the project, in the event things do go wrong and you do not respond appropriately. The gains from the project will be lost and you will have to stat all over. However if you monitor the process and take the appropriate response, the new and improved process has a better chance of staying on tract and maintain the gains that were made
Total Productive Maintenance-Key Lean Tool- TMP
TMP includes having the right people involved at the right levels -For ex. operators who use the equipment everyday should be involved in the process
Criticisms of Six Sigma
Takes too long Too constrained Requires too much data
Bruce Tuckman 60's
Team dynamic of forming. storming, norming & performing keys to successful team
Round Robin Brainstorming
Team takes turn providing ideas verbal or written down /facilitator poses problem and chooses lead to generate first idea Pros: each member has = number of opportunities/ avoids one individual from taking over. Cons people may forget their idea while waiting- tip: have them write down ideas while waiting
Present Performance
Teams can gather real time data to draw the process and get an idea of how the process is performing. 4 numbers needed 1. Average of data x with a bar over it x=1.25 2, Standard deviation (s) s=.05 3. Upper specification limits (USL) 1.32 4. Lower specification limit (LSL) 1.07
What does standard deviation tell us? -m
Tells us the variability or dispersion of data. If normal curve we expect 68% of data to be bt + or- one standard deviation; 95% of data to fall bt + or - two standard deviation, 99.7% of data or almost all to fall bt + or - three standard deviations
Voice of the Business (VOB) -d
The VOB and The VOC are not opposing voices need to marry, you're in business to make money and keep the customer happy
Process Capability
The ability of the process to meet the customers requirements; Made up of 2 things 1. Voice of the process or understand what the process can do 2. Voice of the customer,or knowing what the customer wants Process Capability allows team to get ideas of where the process is performing well somewhat predict outcome communicate better with customers develop more realistic expectations Plan better than if there was no data
Process Capability -Control Phase
The ability of the process to meet the customers requirements; Upper and lower specification limits must be clearly defined . Made up of 2 things 1. Voice of the process or understand what the process can do 2. Voice of the customer,or knowing what the customer wants
Variation -m
The fundamental principal of six sigma, how variation causes defects and focuses on identifying and reducing the sources of variation to increases the quality measure of variation or dispersion allows us to see where it occurs
p-value A
The probability
Control Charts and stability
The process must be stable in order to determine process capability. What does stable look like? Points fluctuate randomly between the control limits as shown below. see pg 14 of package 8 of 8. A process that is control is predictable. and can be unstable. A process that is not in control demonstrates instability and we are unable to predict how it will behave next-all over the place. pg 15
rejection region -A
The set of values for the test statistic that leads to rejection of the null hypothesis. The significance level is the total amount of data in the tail, or tails, depending on the type of test, the tails are also know as the rejection region. If the test stat falls in the tail beyond the critical value, known as the rejection region, reject the null hypothesis and accept the alternative hypothesis. If the test stat does not fall beyond the critical value, then we fail to reject the null hypothesis at the given significance level.
Phillip Crosby
The unconditional involvement of senior management , the culture that must be adopted for the new philosophy to take effect Four absolutes of quality management identify with which quality should be measured Quality comes from: conformance/ prevention/Quality standard is zero defects/ Measurement of quality should be the cost of non conformance
Critical to Quality Concept -CTQ -VOC -d
They are the key measurable characteristics of a project/service /process who's performance standards are established by the cuter There are four decision drivers that help u understand what you customer wants and why they buy from you Features - does the process provide what the customer expects and needs? How do you know and what are u using to help u understand what the customer expects and needs Quality- How does your customer define value? They define it bc they are the ones purchasing it. Delivery - getting the customer what they need in a timely manner. What is the customers time frame. How do you know what you're delivering is in the customers time frame? Cost-How much does the product or service cost? Does the customer perceive the value for that cost?
Team Members -d
Those who actually touch the process, internal staff , a direct customer, supplier ex. team lead, machine operator
interval scale
Tim, temperatures and calendar dates is an example a scale of measurement in which the intervals between numbers on the scale are all equal in size; similar to nominal or ordinal data but does not have a defined starting point. Order and defined differences but no definite starting point, arranged in some order for which differences in data value are meaningful, arranged by order and distance but with out origin. ordering scheme can be interrupted,- zero doesn't mean anything
Is data behaving Normally?
To determine this display data on a control chart: Allows us to see a process over time Provides more predictability than raw data Are there any trends? Are any points out of control. If data on charts is within the UCL upper control limit and the LCL lower control limit then it is predictable and behaving normally.
Meeting Rules
Turn off cell phones/reduce side conversations/be on time for breaks/ be on time for the beginning of meeting/ ask questions/Parking lot unresolved issues so don't get off track . consequences for rules... sing I'm a little tea pot is someone's phone rings.etc
F table-I
Two degrees of freedom each table is unique based on significant level .- degrees of freedom for variance in numerator and degrees of freedom in the denominator and have a significance levels such as .01 or 1% and .05 or 5% (table of cutoff scores on the F distribution for various degrees of freedom and significance levels)
Ppk cal How to center the process
Two formulas If center is closer to the lower specification use this formula ; xbar-LSL/3 x s If center is closer to the upper specification, use USL- x bar/3 x s use x bar =1.25; s=.05; USL= 1.32; LSL= 1.07 In this case use the second formula bc the mean is closer to the USL USL- x bar/3 x s 1.32-1.25/3 x .05 .07/.15=.466 Generally u want numbers close to 1.5 or greater, no where near that
Green belt qualities
Unbiased, preferably no stake in the problem, they have a basic understanding of the process, not an expert, excellent soft skills, knows how to communicate, good project management tools, can create a plan and keep team on task.
F Test-I
Use - if sample size larger than or = to 30 (use student T test for sample sizes less than 30) -there are more that two means (use t- test if only two means) Normal variation
Voice of the Process in DMAIC
Use it in each phase Define - use SIPOC-Suppliers, Input, Process, Output, Customer Measure -use sub process maps and drill down & add more detail, helps determine measurements needed Analyze- Use more detailed flow chart to identify the sources of the problem. It's important to know where the problem is occurring and what inputs reside in that part of the process. helps determine which input needs to be adjusted in order to solve the problem Improve-create a current state vs. future state map. state with current state map and make adjustments in order to create that future state map. Control-final phase, this map is the final version and goes to the process owners as their new and improved state of work.
Data
Use it to support you decision making, never use gut feelings. The voice of the process is in the data, data table and picture rolled into one can show
ANOVA (analysis of variance)-I
Use to test two normal populations with the same standard deviation have the same mean - This tests for equality using the variance. (v= square of standard deviation). Answers the question: is variation bt the means significantly different? Use to compare the means of many different samples to tell us if that factor is a significant part of variation. calculate If the F-test statistic exceeds the F - critical value, reject the null hypothesis and accept the alternative that the means are not equal In ANOVA, MS is the mean square error term. When you divide the MS between groups by the MS within groups your come up with your F-test statistic
Dr. William Edwards Deming (1900-1993)
Used Shewharts tools such as statistical process control, operational definitions and the the Shewharts cycle called the Plan-do-Act or PDSA, he would also change the landscape of quality improvement. he was a government consultant who was instrumental in the post war economic miracle from 1950-1960. His teaching helped Japan be come the worlds second most powerful economy in just under 10 years. He is best know for Fourteen Points for Management, Seven Diseases of Management and Total Quality Management or TQM, ideologies. His methods were also responsible for Fords resurgence in the 1980's.
Run Charts -m
Used to plot time series data Useful in identifying trends Use dots to show data on chart and then connect to show highs and lows; Or you can make a frequency distribution by dragging the all the unconnected points to one side, which creates a picture. Run charts can tell u about trends, 1. 7 points in a row either in upward or down ward direction must be in a row, tells u that something unexpected happened 2. a shift of 7 points in a row above or below the center line or the average line it tells you something unexpectedly shifted the process up or down. Why 7 based on 50/50 probability. magic number 7 Use 6M's to help figure out , if plotting golf score why were you unexpectedly better that day, think, good mood, or worst another day, it was raining? Do the things that unexpectedly made you better and stop those that made it worst.
Shewharts Rules -A
Walter Shewhart is know as the father of statistical process control ; He and team created 8 rules to determine if a control chart variation was common , normal variation or random cause variation. Charts are divided into 3 zones Zone A is + or - one standard deviation from the mean Zone B is + or - two standard deviations from the mean Zone C is + or - 3 standard deviations from the mean. Rules 1) any data beyond the upper and lower control limits is probably special cause variation . MOst of data should be in Zone C 99.73% Rule 2) 9 points on either side of the mean is called a shift. investigate why occurred could be special cause variation Rule 3) 6 points in a row increasing or decreasing is a trend. investigate why occurred could be special cause variation ; If trend continues could end up outside control limits investigate root cause Rule 4) 14 points in a row alternating up and down may indicate seacial cause variation , investigate why - cycle Rule 5) two or three points are in Zone C, that may indicate special cause variation , Zone b covers 95% of data if common or random variation is in play Rule 6) 4 of 5 points in Zone B or C could indicate special cause vaiation Rule 7) 15 points in Zone A above or below the mean Could be special cause variation expect 68% of data in Zone A operators maybe throwing out defects instead of recording them. If process running that well redefine control limits- Rule 8) 8 points in a row on either side with no points in Zone A could indicate special cause variation. expect 68% of data in Zone A
SIPOC Chart-Supplier,Input,Process, outputs,customer -d
Way of mapping the process 1. Name the process -refines scope and clarifies what the team is working on; Keep it simple -name the process with a noun verb arrangement, like house-buying process or car-leasing process 2. As a team identify a start point and a stop point of the process & scope of process. ex starting point of buying a house start point forming a contract end point keys at closing 3. Identify the customer- start with customer bc we want to see their point of view.. buying a house , customer is person purchasing and their children or pet. 4. Identify the outputs of closing on a house should be that it is timely and the information correct. 5. List the highest level steps in the process, usually 5-7; for buying a home lean application, the inspection the appraisal, closing etc.. 6. Determine inputs needed in the process- when buying a house, comparisons to similar houses to assess value, credit information needs to be verified, various documents should be provided 7. Identify suppliers- who are suppliers- house-buying process include the mortgage broker, the inspector, the appraiser, the title company and more. ultimately one map with whole picture for team to use
Perspectives of a process-d
What we think the process is; What it actually is; what we think the process should be. When mapping the process you want to capture what is actually happening. It needs to be "as is" with no bias, capture good and bad, in order to be able to determine the problem. Focus on the process and fix it, the people with be fixed by default.
Correlation-A
When one value gets larger the other gets larger one value gets smaller, the other gets smaller one value gets larger, the other gets smaller one value gets smaller, the other gets larger Correlation is not causation. just bc ice cream sales and shark attacks rise at the same time does not mean they are related
Using Images for data
When possible convert a check sheet, discrete data into a picture, the Pareto chart is a good tool for this helps see what's going on in process Can use cycle time or continuous data to show frequency or distribution patterns
Termination Milestones/Checklist
When terminating a project you need to set some termination milestones and take budget and cost allocations into consideration. Termination Checklist 1. Plan a closeout meeting and a personnel rearrangement a strategy and then prepare some means of evaluating the project and the people who worked on the project, not just the ream members by the SME's, stakeholders 2 .Close out all work orders and subcontractors and contact the necessary finance and IT personnel who can help you close out all of the project books and who can make sure documents are traceable, in case you need them in the future. 3. Document delivery of products and services to your customers, submit your final reports to management- the control plan docs- and notify any stakeholders that are not in the direct loop that your project is complete. 4. Conduct a final review with key management personnel, including champions and get their signatures on all of the relevant documents. 5- Celebrate and promote food work that your team as done communicate throughout organization
Validate & Prioritize CTQ Characteristics -d
When you have good data validate specific requirements with your customer, ensure no gaps in what customer wants and what you are actually doing, confirm with customer
leading indicators -c
X's they are check points along the process -They should be reported frequently to ensure stability -Without the x's it is difficult to know if they Y actually being accomplished
(Cp) capability index/performance formula
Xbar=73.4; S^=1.8; USL-80; LSL=65 Cp= USL-LSL/6xs^ like Pp formula the numerator represents the voice of the customer while the denominator represents the voice of the process. Plug in the data to determine the value of Cp: Cp = 80-65/6x1.8 Cp=15/10.8 Cp=1.39 Cp like Pp will ideally be 1.5 or above. This out come is close. A Cp of 1.39 means the specification spread is 139% of the process spread. That mean there is a small amount of wiggle room on either side of the process or bell curve, in relation to the customer specifications, ideally, however, we would like more room on either side
equation y=f(x)
Y is a function of X Y's are: dependent; outputs; effects/symptoms/Monitored -see if out of control X's are; Independent;Inputs/processes/causes/ controlled Identify a measurable Y will satisfy the specific attribute you want to improve; there can be multiple measurable Y's, so you must fine the Y that will have the highest impact on the desired result
Six Sigma- Do not become trapped by details
You do not need to calculate sigma or design an experiment to find the optimal solution for every project. -If it is obvious your process in not going well- you have a lot of defects/variation- then the math becomes redundant
When do you use an analysis of variance or the F test that is embedded in the analysis of variance table?
You use it to test the null hypothesis , which is that normal population with the same standard deviation actually have the same mean ; the null hypothesis is that those populations do have the same mean, and the alternative hypothesis is that they do not have the same mean. To do that you use a one way ANOVA, which uses a ratio of the explained variance over the unexplained variance or another way to say it is the between group variance divided by the within group variance
Population -m
a complete collection of people things or data that you can analyze and make conclusions about. example ENTIRE elementary school
Statistical Difference (measure of variation)
a difference or change in the process that probably didn't happen by chance. we are looking for statistical differences when using hypothesis testing
z-score A
a measure of how many standard deviations you are away from the norm (average or mean) + or - 1 standard deviation is 68% + or - 2 standard deviation is 95% + or - 3 standard deviation is 99.73% I's called the standard normal distribution Statisticians call it a distribution instead of a curve The probability that a number chosen at random will show up at the left side middle or right. distribution is a normal , symmetrical curve. that is half 50% z score is the probability something is on the right or left of the mean Part of a standard score
Statistics -m
a measurement that represents a random sample or a subset of a population, if a pop needs to be measured and requires a sample to be analysed then it's a statistic- denoted by regular letter S represents the standard deviation of a statistic or a sample
Multi-Vari Chart-A
a tool that graphically displays patterns of variation. It is used to identify possible Xs or families of variation, such as variation within a subgroup, between subgroups, or over time. Time-time variation piece to piece variation and with in piece variation process
Brain Writing
a variation of brainstorming whereby participants write (rather than speak about) and share their ideas many variations, 1 idea published by Bernd Rohrbach 6 people 3 ideas per round 5 min each round end up with 108 ideas generated. PROs A lot of ideas, avoid group think/ Cons more prep involved and a lot of ideas to organize
alpha error A
accepting a bad hypothesis Type 1 error, by chance alone or bad luck not ok will be costly and safety. select a pretty large rejection region, run into a type 1 error where I reject the null hypothesis because my test stat fall our here even thought my null hypothesis is true
Roll Through Yield (RTY)
accounts for both scarp and rework, it's equal to the first pass yield (FPY) of the first process step multiplied by the first pass yield of the second process step , so on and so forth for all process steps Process A= 20units/4 defective= .80=80% FPY Process B= 20units/2 defective= .90= 90% FPY Process C = 20units/1 Defective = .95= 95%FPY RTY=.80x.90x,95=.684 = 68.4% not good The yields of a process multiplied together, only accounts for defective products or services, not if a product has multiple defects per process steps
The Vetting Process -A
after eliminating potential root causes the list will go through one more vetting process the next tollgate, tollgate 2 of Analyze potential causes will be verified before being designate as the critical root causes
good sample -m
are randomly select and are of sufficient size. in order to understand population as a whole
Defects
anytime you make a mistake and have to complete the work again
Activity Network Diagrams
as know as PERT program evaluation review technique, provides a weighted assessment of time or cost. 3 scenarios - Optimistic , best case/ Pessimistic or worst case or ML Most Likely scenario
Interrelationship Digraphs
as know as a network diagram, helps identify relationships and influences between various factors, areas or processes. It can be used to help identify critical issues and key drivers of a problem. Items or concepts with most connections will usually be the most important. outgoing arrow are roots causes, incoming arrows are key outcomes
Measure of Success
at least one goal is obvious, what ever is on the charter. It's the end goal that the team is trying to meet, profitability, an increase in sales. Typically improvements don't go from one performance level directly to the next, there is a gradual increase in performance and leading indicators. Identifying those upfront and ensuring the teams tracking them throughout implementation will ensure success. Set expectations with leadership at the beginning of the project and regularly through out the project will ensure that the team is set up for success.
Collect data
based on return on your investment, how much improvement will occur on project variable Y if data collected Determine, what data how often , how much and where in the process am I collecting it, collecting in it;s not granular form helps set up for Analyze
VOC -Customer needs -d
basic problem that customer is trying to solve. hungry, need a place to eat, travel need shelter, company needs electronic document storage, needs can change based on new problems , external pressure and completion Need to eat Needs Comp need electronic doc storage system Needs Blood Transfusion Most important Voice
scope _ D
be very brief and identifies the problem process, be specific and set boundaries, where does project begin and end. What's not included in the project. where project starts and stops. Keep short
DPMU Defects per million units -calculate
call center 500 calls listen for right answer given, 75 wrong answers given. 1. Defect rate = 75/500= .15 = 15% 2. DPMU .15 x 1,000,000= 150,000 defects per million units 3. Use Sigma table One column shows defects per million the other shows sigma, find a number slightly larger than 150,000 and one slightly less than 150,000 look in between those numbers see 2.5 sigma and three sigma. calculator confirms 2.54 sigma
DPMO (defects per million opportunities) calculation
call center looking at total service still monitoring 500 calls and 75 defects but looking at answer speed, courtesy of service and correct response 1. Opportunities for a defect = 500 x 3 (speed, courtesy of service & correct response) = 1500 2. Defect rate 75/1500 opportunities = .05 = 5% 3. DPMO = .05 x 1,000,000=50,000 4. Use sigma table to find that 50,000 dpmo is somewhere bt 3 and 3.5 sigma . calculator confirms 3.15 sigma
Process Capability-Control Phase
can be used to determine past performance, present performance and future performance. We are focusing on Future Performance
constraints-d
can derail, limit a project progress; Need support of your champion, your supporters, IT to collect data, Finance (need help of champion to get this), not having adequate resources. unavailability of people, resources and raw materials
Baseline Performance Measurement -m
can look at baseline six sigma- reaching six sigma means 3.4 defects per million opportunities; ex If we live to 95 years old divide our live into hours. at six sigma level only 2 hours were bad. It starts with a yield, how many out of 100, out of 1000 out of a million/ the bar is continually raised/ sigma helps break it down into sizable chucks
Causation -m
causation does not mean correlation.
independent data
data that is not related to each other, out put of one does not effect the other
measure
describes how to identify the right thing to measure, how to create a plan to collect that data, and how to baseline the process
Improve
describes how you can improve that condition
Chi-square test-I
distribution is where we test the null hypothesis that 2 factors being studied are independent. ex of factors.. - Age bracket of drives and choice of automobile manufacture they buy. -Median house hold income and level of education. When you reject independent accept dependence- ex 11.3 does not fall into the critical region of 12.6 (it's before it) or the reject region, therefore, you would fail to reject the null hypothesis that house hold income and education level are independent of one another. therefore they are dependent on one and other
Team Facilitator
doesn't know out come respects and trust know people personally & professionally, need expect to lead team there
Swim Lane Map- deployment map or cross functional chart-d
emphasizes who performs which task and which task and what steps are done concurrently, clear visual perspective of customer hand offs across functional areas, show relationship interconnection bt g groups involved in the process. show were X are located, easy to see gaps
Process mapping - Define Tollgate
entire process on one large map includes inputs and outputs, who suppliers are , who customers are
Pareto Chart -m
excellent way to visually graph a large amount of data; Pareto 80/20 rule 80% of the wealth comes from 20% of the people, same thing with the process 80% of the process performance or issues are drive by 20% of the activity. Use this chart when you have multiple root causes or drivers, if it is continuous data, or discrete data you can chunk it into finite number of buckets or characteristic.
Inventory
excess inventory waste storing cost money and space, also if something happens to stored inventory, you will have defective products which are also waste
Trails -I
experiments, runs, experimental runs
analyze
explain how to come up with a long list of possible root causes to a problem, narrow that list down to a short list and then isolate it to one or two root causes
T-distribution _A
explains how probabilities occur determines what the probability is of a particular value occurring based on the test statistic. degrees of freedom = n-1 n=sample size 8-1=7- can be used to help find the root cause and help find solutions -ex want to know which training is best, train two groups differently and then measure their average level of performance can use it during define phase to help determine customer satisfaction
Control
explores how you can maintain the improved process
Free-Form Brainstorming
facilitator poses topic or problem, team provides ideas orally or written down, facilitator organizes ideas/ group discussion held to decided on best ideas and action plan/ Pros: simplest technique; few resources and little time to conduct an effective session Cons: Vocal individuals over power team
Concept Testing -I
gaining consumer response to the solution Involves; -evaluating customer acceptance of a product or service idea -Identifying and prioritizing customers perceptions, want, needs and benefits -Communicating what is in it for customer -Refining ideas for greater market acceptance -Gaining insight for optimizing designs of a product or service This testing is the easiest to complete in the concept phase ;it is easiest to make changes when there is a physical product or service available
Rules for Nominal Voting
gather group of willing people/set an achievable goal/determine the limitations of the group/develop guiding rules/ create good atmosphere/ identity questions / address one topic per meeting
Forming
gathering of a team , not emotional define ground rules and expectations for team
Brainstorming
generate a lot of ideas and to pare those down to the ones that provide the biggest bang for your buck
Box and Whisker Plots -m
graphical method to summarize the data set by visualizing the min value, 25%tile, median, 75%tile, Maximum value and potential outliers, 3 points create box 1. median, splits data in half 2. median of 1st half of data 3. median of second half of data. three points create quartiles, create by putting data in order from least to greatest, find median of data, find median of first half and then median of second half of data Benefits allowing for the comparison of data from different categories for easier , more effective decision making.
Pareto Diagram-A
great way to rank issues, which occurs most often/ collecting data is simple for this diagram, many need more data Use if the question is "Which of these bad things happens most often"
Green Belt working with process owners
green belts not process owners boss/green belt can't tell them what to do/green there to help process owner solve a problem/ Goals and time lines should be should be communicated through sponsor or champion/ green belt needs to engage the process owners early in the Define stage
Shift - A
if seven points above or below the average- control chart
Use RTY
if there are checks and yields from multiple process steps and if multiple defects per unit are rare and not impact to the cost of quality
One root cause -scenario
if you have a very obvious root cause then you do not need to spend your time generating multiple solutions and pilot testing them
Variation -m
if you measure fine enough you will see it (bowl example) Also example if we measure height in the nearest yard it would like like we are all 2 yards tall/ if measure to nearest foot bt 4 ft and 7 ft. Inches tells true variation You may not think there is variation in your process, but you just have to measure fine enough to see it
Prototypes -I
implementing the exact solution on a smaller scale WHY USE- Confirm product works -Gain opportunity to assess the robustness of the design, where's the products breaking point. -Testing failure modes -Testing customer feedback -Validating whether or not the product is compliant with regulations -Evaluating production plans to decide if the product can be manufactured -Comparing the product to your competitor's product, how can I set my product apart from others. WHEN to use- when SAFETY is an issue -Need to confirm proof of concept testing -evaluating the manufacturing cost
FMEA Prevention -I
includes activities such as Error proofing Taking into account the process capabilities and performance of a current process to understand its characteristics
normal distribution -m
indicates that random variation is operating in the process which is different then when something is systematic bell curve
market research -Voc -d
internal data as well as industry data to understand products, customer needs, services and competitor, often third parties provide research solutions because their specialty is complex data analysis Pros- large amounts of data, gain industry data not accessible to your comp. Cons very expensive requires third party consultant who may not understand your firm
Project Sponsor
is Project Champion
Parameter -m
is a measurement like standard deviation or mean which represent an entire population(Football team, ecomm site, know how many) if population small enough to analyze it's a parameter; note by a Greek letter almost looks like a sideways P . sigma is standard deviation of a population
Process- d
is a repetitive and systematic series of steps or activities where inputs are modified to achieve a value added output. the intent of each process is to add value as perceived by the customer
Sample -m
is a subset of the population example 100 randomly selected 2nd grade students
Waste -I
is any activity that consumes resources, but creates no value for the customer 2 types -Required waste are non value added but necessary for their current process. -Pure waste add nothing to the process
Good Data
is important to help us understand the potential effect of the x's . When we have narrowed down the X's to a vital few, we need to be able to validate them with data. Data is also used after the project to help sustain the gains. Once handed back over the the Process Owner they need someway to track performance to make sure things don't go back to the way they were, data does this
Change Management
is people understanding why an improvement is being made. If people feel the personal reason or certainly the real business reason that the change is being made, where it's to improve safety, profitability or productivity it's very easy for them to buy in. If they feel it's being done bc that's what HR wants or bc their boss it did that way in his old comp. they won't be as willing to do it.
Trend -A
is seven points in a row in an upward or downward direction -Control Charts
Value Stream Mapping -Key Lean Tool
it helps us see the flow of material and information throughout the process, - complete a current state map to see where you are today. -complete a future state map to see where you want to be - Use the future state map to note the improvements you need to make.
Root Cause Analysis RCA -A
it;s a tool to understand the underlying causes of a problem, not one cause. the leaves are the symptoms of the problem or the weed, it is above the surface and very obvious; The root is the underlying cause; it is below the surface and is not obvious RCA is the process of finding the and eliminating the cause which would prevent the problem from returning
How to capture VOC -d
make list, decide how much data u need, how you'll use the data, tool of data collection, collect data
Leading and lagging indicators -m
leading indicators are things you can change in order to get better results in your lagging indicators . lagging indicators cannot be changed, so focus more on the leading indicators
Workout
less complex version of six sigma; Define the problem Decide what you need to measure Get you baseline Brainstorm toot causes Affinitize root causes Prioritize root causes Find the root cause Brainstorm stolutions Affinitize solutions Prioritize solutions Implement solutions Measure again
Cpk cal
look at centering statistic. Like Ppk there are two formulas for this; If the center is cloer to the lower spcification, use this formula; X bar-LSL/3 x S^ If the center is closer to the upper specification, use this formula; USL-xbar/3 x s^ In this case use the second formula , bc the mean is cloer to the USL: Cpk= USL-x bar/3xs^ Cpk=80-73.4/3x1.8 Cpk= 6.6/5.4 Cpk=1.22 Cpk isn't quite 1.5 either with means the data is crowding the upper specification.
Kaizen Event-Key Lean Tool
make a very substantial changes and improvements in operations in a small number of days instead of months
Inputs market research -d
market demographics of an area, customer specifications of similar products, the margins of your competition vs your company, and CSI Customer Service Index
Data not center in bell curve
mean falls with in the customer specifications but part of the tail is outside. That is a problem
central tendency -m
mean, median, mode/not about which number is most accurate it's about which number best reflects central tendency of your data. More variation distance of curve wider
Measure Phase-
measurement systems are developed and evaluated/current process performance is measured, key performance indicators are determined, data is collected quality of data is checked for accuracy, distribution of data can be determined to see how the process is behaving the amount of defects and variation present are determined-when and where to gather data from and what should be used
Scatter Plot _A
might tell us if a possible root causes has a relationship to the data Use if the question is " is there a relation ship be one of the critical X's and the Y or out come variable?"
Poka-yoke -I
mistake proofing developed by Shigeo Shingo, one of the fathers of Lean. 1961 he was visiting a plant noticed the springs were being left out of a part. It wasn't noticed until it got to the parent plant, by that time very costly error to fix, so he developed error proofing method such as check lists. He concluded that people forget for 2 reasons 1. Occasionally miss something. 2. Sometimes we forget that we even forgot something. examples; prompts to fill in on line forms, safe checks on medicines if they interact with one and other and a doctor prescribes it, an error message is sent/ manufacturing sorting devises to ensure right parts are passed along/ light sensors and distance measure machines in high volume production environments
Storming
most complex stage, different views role confusion, aggressive , not communicating, unprofessional, facilitator needs to listen to the team challenge them , don't get personal, remind team conflict, storming. is ok natural but work together toward problem
bell shaped curve/Normal curve -m
normal distribution-no unexpected variation from any of the six M's.- can't learn much by looking at one measurement. The more the process increases toward the center, the more it increases variation. Data distributed normally exhibits very specific characteristics, these characteristics allow us to make predictions or conclusions of out comes for a population
Factors -I
one of 6M's - something in the process that you want to try at a different level or a different setting, generally we want to simultaneously manipulate more than one factor to see its effect on a response variable, to see the effect on some desired result
Prioritization Matrix
organize data and prioritize resources and processes. Helps to narrow your options and compare choices in a systematic approach
Active Listener should
paraphrase the concepts being discussed/ provide conversation, back & forth/ask for clarification
Champion -d
people at the top of the organization, senior level
Example of Attribute data
people can't have 8 1/2 people at an event-Yes/No Questions
Tracking Costs Over Time
prevention cost goes up at 1st then levels out don't get caught up in numbers 5% is great
Data Observations -d
provide the most unfiltered and valuable data. Let Jane Goodall studying chimps, you study the customers; Pros observing problems that customers may not see as problems, unfiltered data, observe non verbal cues Cons- very expense and take a lot of time
Map the Process Tollgate 3 Define Phase-d
required SIPOC -be ready to explain Quick and concise explanation of the process even if it is damaged, convenience them you know what is happening Map process early indicators of what some of the causes might be, do not speculated, stick to the facts
X-Bar R-Chart-A
requires a rational subgroup- they have continuous production that is fast with not much changing ex bottle filling Not much is changing in terms of the 6 m's baking whole batch is exposed to the same processed components
Training and communicating the response plan
response plan creates on ongoing strategy for taking action if the process starts to get out of control. It is how to instructions to ensure a timely and appropriate response to processing problems as the occur, which allow the team to decrease the risk of defect getting to the customer. It also allows the employees to be trained so the know how to respond to the process one it has been subjected to the DMAIC methodology. The response plan should be complete prior to writing any work instructions and procedures bc generally, work instructions and produces are used to train employees. The response plan is no exception must align with your work instructions and procedures. It is imperative that the response plan is visible, acceptable, accessible and clearly communicated to everyone
In Control Phase
revisit things identified in the measure phase, return to baseline to see if you really changed your performance or not
Data Collection Plan
s/b easy to understand & read. Define all columns / Ensure that it can get right types of data in right amounts/identify what is already in document control for data collection two types of data discrete and continuous can use discrete data check sheet to show why delays exist. and frequency
Green belt pivotal roles
selecting team, ensuring everyone knows their responsibility, roles to ensure success of project
Before Meeting
send critical docs to sponsor couple of days before for review, Prepare back up docs(data collected and details)Find out sponsor's preferred format for the presentation- ex. power point, bullet summary, the sponsor is your customer.
Good enough is not good enough
shooting for target with as little variation as possible around those targets
Level 1 Map management or high level map-d
shows big picture shows how product or service is made and shows which map step is likely related to the problem. Not a lot of detail
nominal group technique
silently generated ideas in writing/record ideas with out discussion/ conduct group discussion for clarification of meanings, no arguments/ vote to establish priorities or rank ideas, discuss vote cast final vote.
Values check sheet -m
simply write values on bottom and have people record the values they see.-snap shot
Tollgate -Generate Possible Solutions -Improve
sponsor looking for -What are some of the teams ideas to reduce to eliminate each of the critical root causes -How did the team identify those Sponsor may make suggestion or veto some of your solutions, you should have a few options for them. Provide official prioritized list of possible solution for each of the critical root causes. Also preliminary cost effectiveness and resource estimates are always welcome at this point
Tollgate Potential Root Causes- Analyze Phase
sponsor wants to know what are some potential reason for gaps in performance How did the team identify those? If list of potential root causes long group into categories for sponsor. give to sponsor before hand. This tollgate is to show as many root causes as possible, not eliminate any .
measures of variation
spread of data dispersion-use the range, interquartile range, standard deviation, or variance to
End of Measure Phase
step separates six sigma from a lot of other process improvement approaches. This will drive decisions from this point forward. At this point u should Have verified the situation identified in the charter during the Define Stage Know exactly how much u need to change your performance to close the gap Know how you're going to measure it and where to get the data. This is critical for the next phase of DMAIC
characteristics of a normal curve -m
the area bt + or - one standard deviation is 68% the area bt + or - two standard deviations is 95% he area bt + or- three standard deviations is 99.8%
Probability and student t distribution-A
student t or t distribution helps find the probability of a value being on one side or the other of the test statistic. First step calculate the test stat. then use the student t table to make determinations about the probability that stat occurred to just chance alone. Its important to know on the table being used if the probability lies to the left or the right
standard normal curve vs t-curve-A
t- distribution looks like a standard normal or z distribution an firs a bell shaped curve, but the tails in the t disruption have a little more data and are higher than in the standard normal curve which come down a little lower first and then come across
Practical Difference
take into consideration- stats may show one thing but is that best/practice for your company always make sure you understand your decision and it's not based on stats alone
Norming
team is aligning around the problem statement and the roles with in the team, result of storming, refocused on problem statement and data collect to help team move toward resolution
Standard deviation (measure of variation)
the distance from the mean
Lagging indicators -m
the outcomes of those leading things that go on upstream. ex stepping on a scale (no matter how often) is and outcome resulting from things that took place upstream in the process
standard cal standard deviation -m
the square root of the variance or the sum of the squared deviation; average distance any point of data is from the mean; s= V E(x-xbar2/ n-1 s= standard deviation V = find the square root E = sum x= point of data x-bar mean of the data n = number of sample items 6 steps data 12345 1. find the mean x bar of data= 1+2+3+4+5=15/5=3 xbar 2. subtract the mean from each point of data (x) 1-3=-2 2-3=-1 3-3=0 4-3=1 5-3=2 3.Square those numbers -2x-2=4 -1x-1=1 0x0=0 1x1=1 2x2=4 4 Add up squared numbers 4+1+0+1+4=10 5. Divide the whole sum by (n-1) started with 5 numbers - 1 5-1=4 10/4= 2.5 6. Take square root Use calculator to find square root 1.58 is standard deviation Measures variability that goes with mean, determines how close the the middle most of the data is shows certain % of data should be certain distance from the mean. impacted by extreme data/ not easy to calculate
Twofers -m
there are three areas your need data for, customers require quality; 1. input from suppliers, a. key requirements,b. quality of raw material 2.process design a. cycle time b. need data from each category and 3. outputs a. How well are requirements met b. products with out defects -in each of these areas there are two types of measures..twofers
Attribute Data-Data Types -m
things that are counted , discrete data, answer to questions are either ; Yes, No, go or no go or counted data ( ex # of people in stadium)
Why we need to get data
to help us understand the baseline performance of out process and the magnitude and frequency of out defects or issue
Joseph Juran
to reach a solution two journeys one is a diagnostic journey from symptom to cause and the other is from cause to remedy. easier is from cause to remedy bc if we find the right cause the solutions tend to gravitate toward the cause. Go through cause gate
Team -Good Team
together everyone achieves more. achieve a common goal/ cross functional every member has a unique role 5-8 people/trust achieve more/ 1. Sponsor/ Champion 2. Process owner 3. Green Black Belt 4. Team members -each team members roles & how much time s/b dedicated to the team
Affinity Diagram
toll that allows large number of ideas (brainstorming)to be sorted into groups based on natural relations ships
Cause & Effect Matrix
tools used to subjectivity quantity the relationships of several x's to several y's. the X's generally come from the cause and effect diagram across top enter output measures or effects (Y's)
mutual exclusive
two events are independent of one and other, no way for the outcome of event A to occur at the same time as Event B; two independent events can be mutually exclusive ex if you roll dice there is not way the results can be both even and odd..it;s either 246 or 135 but can't be both. out comes are mutually exclusive
degrees of freedom=A
two tailed test find degrees of freedom (n1 + n2) -2 n=sample (8+8)=16-2=14 degrees of freedom; use to find critical value on table. For testing at .05 level go across to the .05 column two tailed test and then you go down the column to 14 degrees of freedom where those two meet you find the critical value of 2.145 Since our value is of 2.7 is greater than the critical value of 2.145 reject the null hypothesis. Rejecting the null hypothesis means that there is a signification different bt the two data sets being compared and accept the alternative
Project Matrix
use criteria to make a project selection-create a list of potential projects and narrow you list down with a project performance matrix-boxes with L=1 M=3 H=5. Rank projects according to importance add up
Complaints and Kudos -d
use data to determine common problems customers face with a product or service they gather data from the source , returned products, customer letters and customer phone calls- comp may have systems to tract both. Pros- gain specific feed back on needs and requirement that are not met. Cons Reactive you must wait until there is a problem before understanding the data. Difficult to track and analyze. (can use internet reviews)
Estimate of Standard Deviation
use the same data to determine Cp and Cpk as for the Pp and Ppk. In this case used an estimate of standard deviation(S^) . This estimation obtained from the control chart provides more predictability that the raw data itself. To determine S^ two additional numbers are needed. 1. Average Range, or Moving Range (MR) bar the range and moving range are the same. When using XmR chart use MR bar. 2. Constant number (d2)
Surveys - VOC -d
use to collect information for a large sample of customers send out questionnaires paper or electronic Pros- inexpensive , gain a lot of infor in short amount of time, get quantitative and qualitative data Cons results are not verifiable, response rates low, responses may not be representative of random sample
Process Capability Pp/Ppk
used to determine past performance, present performance and (future performance cpk).
CTQ tree -d
used to drive analysis from the VOC back to those activities requested to help meet the needs of the customer. Ties directly into the house of quality or quality function deployment. Data should always come first from the VOC.-quantitative method
Hypothesis testing (How you test )
used when making inferences about what our data is telling us. helps integrate the voice of the process with the voice of the business. (use statistical difference with hypothesis testing)-concept ;use different testing sample T test ex can use to test for variance.
Third Party PI -m
ways to gather customers needs and requirements Data comes from outside the company, industry experts publish information, industry associations, competitor data, market research Advantages gain competitors data and research data, reach more people, less company time spent gathering data. Disadvantage; cost of consultants , consultants time line, too much data
Direct PI -m
ways to gather customers needs and requirements Your directly in contact with your customer you as a team leader conduct interviews, participate in focus groups, survey customers, direct observation, become the customer, various score cards. Advantages- firsthand information/targeting your groups/face to face contact Disadvantage; Cost (rent room, feed people) Group think, Survey response rate low
Indirect PI -m
ways to gather customers needs and requirements data comes from an internal third party- Complaints department, service department, sales people . ex. internet complaints Advantages Immediate resolution/ Explicitly expressed concerns/direct communication/control over turnaround time Disadvantage; unhappy customers don't complain, they just go away, dissatisfaction not expressed
VOC- voice of customer -d
what are customers specifications, targets, requirements, and tolerances
alpha percentage A
whats in the tail or rejection region-a value will fall out in the tail just due to chance. Alpha determines where the critical value goes and where the critical value goes determines where we reject or accept the null hypothesis or fail to reject it based on the relationship to the test stat-
Key questions, considerations regarding testing-l
when deciding on testing answer the following questions. -How long is it going to take to do the test? -How much data am I going to need? -Do I have a stable process in place that I am actually going to get useful information? -What are my verification and validation criteria for both my implementation and testing strategy? -What kinds of schedules are there and are we going to meet them? -Am I in the confirms of regulatory requirements?
Use DPU
when the cost of rework is high and multiple defects per unit are possible
Over production
when you produce more than you need or you deliver something sooner than needed-waste not changing form or function and need somewhere to store it
Scatter Diagram -m
work great for visualizing the enemy, it shows variation you could not see looking at numbers if angled into left corner and out positive correlation if scattered from top of L to bottom end of L negative correlation, does not show causation- snap shot
Draw the process
x bar= 1.25; s=.05; USL= 1.32; LSL= 1.07 draw the process on a piece of paper. to calucate the practicail end points to represent 99.73 of the data should fall add and subtract the result of 3 x's on either side of the mean. .05 x 3 =0.15 1.25 + 0.15= 1.4 1.25- 0.15= 1.1 draw a bell curve 1.25 center of highest point. 1.1 falls to the left side furthest point from center 1.4 is on the right side furthest point from center
Functions and Variables-A
y=f(x) relates to independent and dependent variables and by extension, Lean and continuous Improvement Y dependent variable X is the independent variable F is the function X goes into the function and comes out as y formula y=x squared -7 when x is 5 y=25 (5x5) y= 25-7 y=18 x can be positive, negatives, decimals ect.. x causes y put in as x comes out as y continuous function causation, different from correlation
why do you use random sampling -m
you can't always survey your entire population, it can be expensive and time consuming
Continuous data
you need at least 30 data points
