Vivek Dahake

5Why and 5W2H are both used in problem solving techniques, but they are different in terms of their purpose and phases where they are used in problem solving work flow. * Over all understanding of these tool - 5W2H - It is a tool used in problem definition analysis, wherein multiple guiding questions like What, Where, When, Who, Why, How and How-much, are asked to gather all relevant information of problem. With this information team can see the exact nature of problem, and can then state it in precise terms. It is kind of problem assessment tool. 5Why - It is tool used in Root cause analysis. It is a series of interlinked "Why" questions asked one after another to reach to root of problem from symptoms of the problem. It gives team a chance to act upon a right & deeper cause, which will solve the problem from its re-occurrence instead of only surface actions * Purpose - 5W2H - To find out all important information about a problem so that SMART defination of a problem can be made. To define what is problem, and what is not problem. To limit the scope of problem solving. To get 360 degree view of the problem. 5Why - To find out the exact root 'cuase' of 'problem' by finding the exact chain of relationship between them To solve the problem permanently from its roots To eliminate the re-occurrence of problem * Where to use it - 5W2H -It is generally used at the start of problem solving in "Define" phase of problem solving work flow or a DMAIC six sigma project, while forming the 'problem statement' in that phase. It problem is well defined considering all directions, If problem is well understood by the team then, problem is half solved, means team will properly direct their next steps. 5Why - 5Why is used generally at the "Analysis" phase of problem solving work flor or a DMAIC six sigma project, while forming the 'problem statement' in that phase. * How to use it - Examples 5W2H - What - What is the problem? - Operator got injured on shop floor Where - Where is the problem? - In machine shop near machine number 4 & 5 When - When is the problem? - 2 Accidents happened in multiple shifts (1st shift as well as 3rd) in last 8 days Who- Who is involved in the problem phenomenon? - 2 machine operators working on machine were the affected persons, one slipped and injured, another slipped but did not got injured Why is the problem - probably floor conditions are not as per standard and slippages happening due to oil seen on shop floor as reported by shift supervisor How - How is problem happening? While working on machine operator has to take few steps to pick the parts and while doing it, he slipped down and got the elbow injury on right hand, in second case operator slipped but did not got injury How-much - How many incidences? there are total 2 incidences reported in last 8 days 5Why - 1 Why - Why we see this as a problem - Because one machine operator got injured which is reportable safety issue 2 Why - (Why Machine operator got injured?) Because he slipped from slippery shop floor 3 Why - (Why he slipped from slippery shop floor?) Because shop floor has oil spillage, reported by supervisor 4 Why - (Why shop floor has oil spillage?) Because drain outlet seal of oil tank of machine has leakage, seen by team 5 Why - (Why drain outlet of oil tank of machine has leakage?) Because seal was not replace during preventive maintenance (PM), verified by team Like this we can go further deeper like "why seal is not replaced in PM - because it was not defined in PM checklist, etc ....And action (actionable/feasible/economical Action) should be taken on Root Cause which is the last Cause of this chain. ~ Vivek Dahake

Yes, Key output can become key input for the same process. And that is what "Feedback" is. It can do miracles in keeping the process output in required/desired state and level. Key output of a process can be compared with "desired output", deviation can be known, and that deviation can be used as "input" to adjust or re-set the key process parameters. This type of process design helps to keep the process outcome in desired levels by taking continuous references of "actual outcome" and "desired outcome" and use as "input". Example - 1. Teacher can ask Student to prepare a topic and teach a class on sample basis. This way teacher is using his outcome (who is student), in the process (teaching), as an input, to improve his teaching process. By doing this, teacher judges whether his student has understood the concepts upto the expectations, accordingly improves his own teaching process. 2. Second example could be financial products, like investments, Money earned out of investments can also be used as investment and more money can be earned through same process. 3. In foundry, left out metal after casting (like gates, cleaned off parting line material from casting) is added casting process, as an ingredient. 4. Process of making Curd by adding small amount of curd into milk - Lactobacillus is a genus of bacteria which can convert sugars into lactic acid by means of fermentation. Milk contains a sugar called lactose, a disaccharide (compound sugar) made by the glycosidic bonding between glucose and galactose (monosaccharides). When pasteurized milk is heated to a temperature of 30-40 °C, or even at room temperature or refrigerator temperature, and a small amount of old curd or whey added to it, the lactobacillus in that curd or whey sample starts to grow. These convert the lactose into lactic acid, which imparts the sour taste to curd. Raw milk naturally contains lactobacillus. It is clear from above examples that, It creates Unique advantage like continuous improvement of process and maximum use of inputs & resources. Sometimes it helps in stabilizing the process. However it can also create some challenges like "Intentional" addition of smaller variation & impurities in process inputs.

Process that is efficient but not effective - 1) When I read a book at very high speed, I finish it in straight 3 hours, but I may not understand its full concept; means I am efficient but not effective. 2) Making popcorn in cooker at home has very low cycle time (efficient) but defect rate is high (not effective) Process that can be considered effective but not efficient - 3) Process of transporting big size blades of wind-mill by road on trucks is effective (last mile delivery) but not efficient (consumes high resources & takes long time)

Continuous data is measured because it can take any value between two values. There is infinite possibility to measure reading between two readings. For Example - 1) Time taken to process a part – it can be 1hr or 1.1hr or 1.123hr, depending upon how much detailing is essential for problem solving. 2) Temp, Pressure, Speed, Time, Length, Space, etc Where as Attribute data is counted because there it takes only integer & finite values. For example – 1) Number of defective cars in Sept-17, Car with even a single defect will also be counted as defective car, but there will not be fraction or infinite possibilities like 0.5car, 0.25car etc. Car is defective or not defective, thats all. 2) Number of pregnant women in a city; there cannot be half pregnant women. Its like pregnant or not pregnant thats all. Confusion starts when we go in more details of defining attribute or its refinement. Practically it is possible to have infinite ranges of attributes, For example - 1) No of persons who have passed SSC in a particular age group range, now here observer can note number of persons in range of age in years, like persons of 15 years old, 16 years old, 17 years old etc…. But practically any two persons will not have exact same age, there is definitely infinite differences, in terms of Months, days, hours, minutes, seconds, milliseconds, microseconds and so on (its infinite)……. there is infinite possibility of making these “counting ranges” hence attribute data can be sometimes confused as continuous data in specific cases. In this example, it depends on problem statement, how much detailing is significant is to be decided by problem solver. Age in terms days is OK but in terms of hrs and beyond detailing will be insignificant and so one should group it in closest attribute range. 2) Other examples are, huge data counts or large amount of discrete entities like Plants biological data human population specifications data grains material quality in mining industry marine science data etc We may not give significance to 50000000009 and 50000000007 and its difference, practically these are continuous but we may consider in to closest discrete range.

Part A) Correction means rectifying/correcting things which have gone wrong; just like quenching the fire. It is not only desirable but indispensable to perform correction to start the system. Correction has to be done immediately or depending upon the allowed TAT by system/customer. Correction’s examples – Correcting the mistake by rectification by doing re-work / readjustment / re-setting / re-assembly Replacing the faulty part and re-staring the machine again Stopping and plugging the leakage of tank and filling up the liquid again in the tank Replacing the bearing with same type of bearing Part B.... Corrective Action (CA) means finding out the root cause of problem which has happened and taking action on it, so that problem will not repeat again in future. Corrective action is taken after issue/problem happens. It is desirable to perform Corrective actions but it depends upon the cost / life stage of system (age) / complications / knowledge availability / customer preference etc. Organization can take a call whether to go for CA or only depend on corrections, depending upon cost-benefit analysis. Corrective Action’s Example – Mistakes happen due to usage of wrong tool being used, then Corrective action would be to define the specifications of correct tool, give right tools, train operators on right tool, modify the SOP, if costs & customers permit Like What caused fire & eliminate that cause from system, if costs & customers permit If low lubrication caused the bearing failure, then replace the bearing with self-lubricating bearings, if costs & customers permit Part C) Preventive Action means finding out the potential cause of the problem before occurrence and taking action on cause to eliminate it from system, so that problem wont occur at first place. It is absolutely desirable and it is ideal to assess; plan & perform preventive actions before problem occurs but it comes with cost, so depending upon the costs involved / complications / knowledge availability / customer preference etc organization can take a call whether to go for PA or only depend on corrective actions / Corrections in later stage of product life cycle. Preventive Action’s Example – Doing Process-FMEA of assembly line and taking actions on those RPN causes which crosses predefined criteria of RPN (say RPN>100) (e.g using torque controlled battery operator tools instead of hand tools) , if costs & customers permit Doing Design-FMEA and modifying the design so that potential causes does not affect design performance (e.g using male-female notches to match the sheet metal parts being welded for achieving perfect geometry of final welded assembly) , if costs & customers permit Part D) Are there situations where both preventive action and corrective action are undesirable and correction is the only preferred action? Yes; there are situations where cost of CA and PA is huge or CAPA completion timeline is so high that it will dissatisfy the customer; wise a wise cost and time of correction is relatively less and customer is OK with correction instead of CAPA then in that situation it is desirable to go with only correction. For example – 1) Replacing blown off fuse (correction) would be the only cost effective and fast option, customer would be very happy to go with it, instead of changing/modifying the whole circuit for eliminating high current (CA) or to redesign the entire system (PA) to prevent load variation / high drawn current in the system. 2) Replacing wheel tyre due to normal wear or tear is desirable/preferred action (its Correction); instead of replacing that tyre after normal wear with very superior rubber (its CA, wear/tear is cause) which does not wear out (it has very high cost), as well as it is also not desirable to design a car in first place with a very special & sensitive suspension coupled with non-wearable tyre material (Its PA) (it will also have very very high cost and customer may not be able to pay for it)

COPQ The cost of poor quality (COPQ) are total cost that are due to offering poor quality services and products. The cost of poor quality can be divided into four different categories: the internal failure costs, the external failure costs, the appraisal costs, and the prevention costs. The internal failure costs are usually associated with the defects that are found on the products and services before they reach the customers. The external failure costs are related to the costs that your company has when the product or service reaches the customer with a defect. The appraisal costs are the costs that your business incurs in order to check the actual quality of the product or service according to some pre-specified degree. The prevention costs are the costs are the costs that you need to have to prevent both appraisal and failure costs to a minimum acceptable. Really worthwhile to do a painstakingly detailed assessment of this metric? Off-course Yes !! What we can measure that only we can improve. COPQ is a very good measure, to set the targets to teams. Assessment of COPQ BRINGS FOCUSE to prevention rather than correction. How precise can one get on this? Yes, its true that one can not be so precise about it. Its like iceburge, 10% visible and 90% under the surface. Practically one can not measure more than 50% of the COPQ, other costs are totally hidden. It is still important to work on it and get it in control. Efforts will have effect on 100% COPQ automatically even if it is invisible in the business. Thanks and Regards, Vivek M Dahake

Difference between common and special cause of variation – Common Cause - Variation caused by chance causes, by random variation in the system, resulting from many small factors. Common Cause variation is created by many factors, that are commonly part of the process, and are acting totally at random and independent of each other. Their origin can usually be traced to the key elements of the system in which the process operates. (Materials, Equipment, People, Environment, Methods). If only common causes of variation are present, the output of a process forms a distribution that is stable over time. Example: Variation in work commute due to traffic lights, pedestrian traffic, parking issues. Special Cause - Variation caused by special circumstances or assignable cause not inherent to the system. Special Cause variation is created by a non-random event leading to an unexpected change in the process output. The effects are intermittent and unpredictable. If Special Causes of variation are present, the process output is not stable over time and is not predictable. All processes must be brought into statistical control by first detecting and removing the Special Cause variation Example: Variation in work commute impacted by flat tyre, road closure, heavy frost/ice. This Differentiation is important because, unless we identify type of cause we can not truly resolve the problem from root and because they require different approaches to deal with them. There are different strategies to deal with it depends on type of cause. Common cause need long term strategy to know roots, manage it and improve the process. Special causes need immediate corrections and short term measures. Hence tools to be used are different. Misjudgement - If the type of variation is not identified then it is likely that the wrong tools will be used.

SIPOC is a visual tool for plotting the summary of SUPPLIER-INPUT-PROCESS-OUTPUT-CUSTOMER value chain. It is done in tabular form. SIPOC is mostly used in DEFINE phase of SS Projects or Kaizen Event, when team members want to give birds eye view or represent / update / make up for the changes for people who were not part of the process recently or to help people to understand the process in totality before they begin further. Focus – During making the SIPOC, Focus can be internal or external (customer, suppliers) or mixture of both to the business; depending on the scope of project. Purpose & Benefits- It definitely serves the purpose of “CLARITY” to the project members. Also serves, Understanding Flow clarity Scoping Limits of project Parties involved Context - Actually its use sets the context right, at very beginning for entire team. Team then can not say we don’t about this perticuler party/aspect/input/output. Also while making SIPOC team actually once go through entire value chain and that gives them full idea about process. Example - SIPOC DIAGRAM - High Level Cost Data Integrity Project Supplier ► Input ► Process ► Output ► Customer Employees Employee Setup Data Setup Resources Assign Activities to Resources Assign Rates to Resources Enter Time Sheets Enter Vendor Invoices Summarize and Report Costs Active Employee Record Project Manager and Team Contractors Contractor Setup Data Active Contractor Record Project Manager and Team Employees & Contractors Planning Meeting Project Schedule Project Manager and Team Contracting Officer Statement of Work Project Schedule Project Manager and Team Payroll Department Employee Pay Rates Rate Table System Administrator Contractors Contractor Pay Rates Rate Table System Administrator Employees Timesheets Labor Cost Report Project Manager and Team Contractors (Time & Materials) Timesheets Labor Cost Report Project Manager and Team Contractors (Fixed Priced) Invoices Contractor Cost Report Project Manager and Team Project Management System Timesheet & Invoices Monthly Performance Reports CIO, Program Managers, VP Finance Instructions for Preparing the SIPOC Diagram: 1. Go through the process step by step, describing the major activities. List each of this activities under the Process column. 2. Add two columns, one to the left of the Process column Input and one to the right of the Process column Output. 3. Work to the left of each activity in the Process and describe the Input into this Activity. 4. Now work to the right of the activity in the Process and describe the Output from this Activity. 5. Add two more columns, one to the left of Inputs labeled Supplier and one to the right of Ouput labeled Customer. 6. Now work down the Input column and describe each source or supplier of the Input. 7. Do the same for the Output column and work down, describing who gets the output. Some points to consider: 1. The SIPOC Diagram should describe the existing process and not the "to be" process. 2. Have some Subject Matter Experts who know the process review the SIPOC Diagram for accuracy and completeness. 3. Include Sub-Processes and other key system interfaces within your SIPOC Diagram. 4. You can include additional descriptive information where appropriate such as applications involved, functional owners, etc. 5. If you need more detail, then create lower level SIPOC Diagrams.

In Manufacturing industry Jidoka / Autonomation mean "automation with human intelligence". Most common Jidoka is the practice of stopping a manual line / machine or a process when something goes wrong or is abnormal; and is detected by jidoka arrangement in process. This is done in order to minimize the losses further. Then immediately the operator present in cell, will find root cause and take corrective actions on root cause and restart the process. Whereas Full Automation is conversion of all human work content into all machine work. The main difference between two is, Jidoka seperates human from machine and has a form of automatic inspection and signalling mechanism along with automation. Example – A Press Machine produces defect then it is detected by jidoka (shape or weight sensor, in a jig on conveyor) before part goes to next machine and if abnormality found then previous machine is stopped and then operator take the appropriate action. Where as full automation means presses are connected with each other with a part handling mechanism (industrial robo) and parts are processed without auto inspection.

Pacemaker process is critical process in the value chain, which creates the pull in the value stream. Production planning team plans production schedule at the rate of takt time, and is first informed to pacemaker process, which is supposed to drive the entire value chain then. Pacemaker process creates ‘pull’ in other processes in value chain to avoid waste. Other processes must produce only when the pacemaker process demands so. Generally it is the last assembly process in mfg environment, but there can be exceptions. Importance & relevance of selecting an activity as pacemaker process in a value stream Pacemaker process is very important because, of following and thus hightlights it relevance in perfect selection from value chain, Pacemaker process responds directly to demands of external customers Gives trigger to internal customers to start the work Smoothens the entire value stream Helps organization to become lean, by producing only when ‘really needed’ by customer & reducing the wastage like overproduction, inventories, delays Pacemaker process is most susceptible with direct impacts on customer Pacemaker process cycle time must answer to takt time (equal or slightly less) Value stream must be balance with respect to pacemaker process

POKAYOKE – It is mistake proofing technique, in which opportunities of blocking the possibilities of errors/faults/mistakes/defects/ill-impact are found out, counter measures are designed & implemented in such a way that system does not allow such faults to occur & becomes immune to it. Unquestionably it is a very useful & effective concept in industry to prevent errors and protect customers from undesirable effects. Poka means mistake and Yoke means prevention – these are Japanese words. Japanese manufacturing industry institutionalized the concept of Pokayoke and automobile industry world wide accepted and improvised by bringing more and more sensors, logical machine programming, feedback techniques, measurements and real time reaction system in automation. Further to add in its maturity many lean practitioners worked on autonomation (intelligent Automation) to prevent errors or if cannot prevent then at least correct it or if at all cannot prevent or correct then detect as soon as possible and give alert to process operator for acting upon it. Which of the given interpretations are correct? Explanation with valid examples. The human error will not happen at all. – It is very difficult for a human to not to make any mistake; since human are live person with a mind and emotions inbuild, we are bound to make some mistakes. Its because we humans are curious, and we learn continuously. We humans learn by trial and error. Hence errors are part and parcel of our behaviours and human life. Interpreting POKAYOKE as “human error will not happen at all” is thus incorrect. POKAYOKE mechanism will immediately identify errors made by human and prevent defect to happen or either correct it or alert the human to protect the customer (customer may be next process or next human in lean chain). Human error may be intentional or unintentional. Nobody decide to do mistake but it happens unintentionally. Mood of human is never constant and there is always a chance of deviation or abnormal behaviour by any human at work. No POKAYOKE in the word can “prevent” 100% human errors. Also % errors can be reduced by continuous training, high skill, Best workplace management and high motivation for perfection. Examples – Highly skilled workers, doctors, pilots, sports person and professionals make mistakes in their respective area, its evident from history of great people. Nothing can really prevent human errors to happen. It does happen but right POKAYOKE does not allow its consequences further. Human error may continue to happen but the defect will not happen. – Yes; this is appropriate interpretation of POKAYOKE. Having known that human errors will happen (because of human’s nature); sufficient process knowledge, training and skill is imparted into humans who are responsible for processes, to stick to standard operating conditions and errors are minimized. Nevertheless, it is essential for lean practitioner to design the mechanism (POKAYOKE) to not let human error result into defect. This can be achieved by sensors (by sensing the error/abnormality created by human) and building on the logic to change the course of process or stop the process (only if correction is not possible) to prevent defect to get created. Sensor’s signal is taken as input in POKAYOKE logic. This kind of POKAYOKE are best to have. Examples – (a) When operator loads the part in machine in wrong orientation, then a shape based sensors (e.g proximity sensor) can sense its wrong orientation and activates the correction mechanism and prevents process to develop defect because of this reason. (b) Even if you forget to switch off the lights while leaving the room unoccupied, Movement sensor senses the ‘No Movement’ in the room and then lights are automatically puts on OFF condition after a while; preventing the wastage of energy further. (c) When wrong coloured part is loaded in the machine for processing; then colour sensor stops the process, till correct coloured part is loaded, when colour is a CTQ; and defect is prevented. Human error may happen, the defect is less likely to happen. – Yes; this is also an appropriate interpretation of POKAYOKE. In this human errors are detected and appropriate actions are inbuilt into process steps to prevent defect creation. However, is not possible to capture all the errors or not possible to build logic for all the type of corrections. These type of POKAYOKE will be good to have but will result in some % of wastages; though considerably less than not having any POKAYOKE. Pokayoke logic is same as explained in case-2 above; however, most of the industrial level sensing/technology may not be 100% perfect all the time, moreover regular wear/tear, aging, environmental factors come in effect and degrades the sensing efficiency and reduces the probability of 100 % defect prevention. Regular TPM practices is the key to restore POKAYOKE to its optimum performance. Example – (a) Vision based sensors present in sheet metal press may not detect minor abnormality due to dust on sensor; and result into slightly less efficiency of POKAYOKE. But in most of the cases sensor will prevent defect to happen. (b) Even if a driver bangs the car on to a pole, airbags will immediately blow up and prevent injury to driver; there is less likely that driver will hit the head against the steering wheel and get injured. Human error may happen, the defect will also happen but will be detected and corrected automatically. - Yes; this is interpretation of substandard POKAYOKE than case 2 and 3 above. In this case human errors are not detected but defects are detected and appropriate actions are inbuilt into process steps which leads to correction or stoppage of the process, to restrict the defective part reach to next process/immediate customer. This is really a inferior Poka-yoke, but in some cases it is not possible to prevent defect to happen and we have to be dependent on correction only. Example – (a) In an automatic packing line, if box is not filled up to required weight and goes ahead (abnormality), then at next station position (before final packing lock) weight is checked, if found abnormal then this unit is not packed but send back to previous station through a dedicated route and again filled up to right weight. (b) In a laminator machine, edge of laminate is continuously detected and if found shifted then correction unit shifts in inverse direction with same amount, shift is the defect but it is corrected in next station automatically. Written by – Vivek Dahake Based on Auto-industry and FMCG industry experience.

In journey of pursuing Business Excellence, its practitioner must make sure that they put enough efforts in preventing occurrence of any type of error; by error-proofing the system. However in practical situations, one has to remember that there will be some % of probability of error; even in a very precise and matured system. But it must be then acceptable limit of Noise/Signal ration; which depends on sensitivity of the system. You have option in case of false alarm, to think and decide whether to react In case of missed alarm, you may not have the option and one has to face the consequences of event for which you missed the alarm I will prefer a false alarm over missed alert. Examples – On 26 September 1983, the computers in the Serpukhov-15 bunker outside Moscow, which housed the command centre of theSoviet early warning satellite system, twice reported that American intercontinental ballistic missiles were heading toward the Soviet Union. Stanislav Petrov, who was duty officer that night, suspected that the system was malfunctioning and managed to convince his superiors of the same thing. He argued that if the Americans were going to attack pre-emptively they would do so with more than just five missiles, and that it was best to wait for ground radar confirmation before launching a counter-attack. Ref - The Man Who Saved the World

Takt Time Concept - Takt is a German word, It describes the conductors baton (for example in music orchestra). It is the concept that all activity within a business is synchronized by a rhythm, set by the customer demand. To simplify this concept, auto-industry, where mass production is done in ‘pull’ manner, has generally described it as “Time Available to Complete Task (Task here means production of one unit e.g a car). An organization may decide and make suitable arrangements to produce a unit (product/service) in Takt-time or slightly less than Takt-time, repetitively in a cyclic manner, this speed at which production is done to ensure org meets the customer requirement is called Cycle Time. And total production time from start to end is called Lead Time. To calculate the 'Takt time', following formula is used, Takt = 'Production Time Available' / 'Customer Demand'. 'Production time Available' is derived by deducting the Lunch and tea breaks, Team briefing times, TPM breaks, Clean down time, etc from 'total available time', as per industry practices. Hence 'Takt' is in fact the goal or target rate. Takt & 'Resource planning' is closely related - As demand may fluctuate, a takt time can be then calculated for anticipated demand (or known demand) to plan resource levels. This applies across the board for any production operation, as well as Lean operations. Why Takt is important & decisions it affects/drives- Takt Time regulates following, 1. The pulse of the Production System, and so the resources required ! 2. Rate of fulfillment of customer demand or Pace of sales !! 3. Links production activity to actual customer demand !!! 4. Ensures all production activity will be synchronized from 1st process to final assembly process, which is nothing but the balance flow. Takt time concept helps to set the flow of production. Work balancing is done to match to takt (by applying lean tools & practices) and so the flow of production is in tune with customer needs. Hence Takt is enabler for work balance in the business. Successful Takt planning drives the business decisions which are strategic as well as tactical in nature. Takt drives organization to meet the strategic objectives of business, for example - Customer Satisfaction, Profits, Investment decisions, ROI, ROCE, Payback, Market Share and Growth. Also effective takt planning drives org to meet operational level objectives (Tactical), for example - Effective use of its people and plants, inventories, cashflow, productivity and various costs. Whole SCM related decisions are driven by takt planning decisions. Calculations/Decisions on Takt eventually drives/affects the decision on Technology selection, Equipment selection decisions, Capital Investment, Operations cost, Offloading decisions / in-house production, Manpower & skills decisions, etc. Also, Identifying the areas of under-performance or waste in the system is the total focus of Lean and there Takt plays a vital role & becomes driver of improvement efforts. - Written by Vivek Dahake, based on experience in Auto-mfg-industry.