Leaderboard

Process yield is used to measure process performance. It shows how good process is working and how many defective parts process in producing. In another way we can say that Yield is percentage of defect free process. Formula for Yield Percent = Actual Yield/ Theoretical Yield x 100. There can be serval challenges while calculating accurate and consistent %Yield. It can be with human error while defining defects, Incomplete data, external factors leading to process variation or Dynamic changes in complicated process. For example you take a project to make a lemonade. For successful project completion you need to measure yield percentage and defects. This will help to improve this lemonade making process. To begin with you have to analyze key process steps like Squeezing lemons to get juice, Addition of Sugar and Water, steering or mixing and finally pouring into cup and serving. Here mistakes can be 1. Identical size of lemons- Sizes can be different. 2. Taste or quality- Some of Lemons may be juicy or too sour 3. counting mistakes- Counting of lemons or Coup 4. Difference of Opinions - Some people may like Lemonade, Some may find fault like too sweat, too sour. By understanding these defects you can improve the process and quality of Lemonade further to delight your stakeholders.

Defective % can be calculated by taking a ratio of Total No of defective items with Total observed Items and Yield is not thing but 100 % - Defective %. So, calculating accurate total no of defective is very critical in Yield % calculation. Below Is the table for challenge faced and counter measure (solutions) to address these challenge Challange Faced Counter Measure Data accuracy and consistency Inaccurate data can lead us in wrong direction of analysis Data collection system can be automized Standard data collection method can be set and operator to be trained on it Frequently changing operational definition Operation definition of inspection station CTQ has changed many times as per customer criteria. Which creates confusion between operators. CTQ definition can set on High standard, single definition can cover all customer requirement Operator skill level Deputing Un-skilled operator on inspection station in always a challenge. New operator many times takes time to understand all defects MSA must be done to ensure that data collected is reliable Multi skilling of operator can be done to handle such situation. AI based systems always make this process easy by auto identifying defects with location as well as defect name. All Defects are not visible. Sometimes all defects of a product cannot be seen due to eye fatigue or monotonous Job. AI based camera system can identify 100% defects. Job rotation in 2-3 Hr a shift of the operator on this type of station where eye fatigue is quite possible Example: I am considering a Solar Panel for this Example. In any solar panel manufacturing process below can be top defects. Poor soldering - Soldering not done up to required strength Ribbon Misalignment – Ribbon wire not on its place Cell Crack – Solar cell got a crack Dark Area in cells – Solar cell coming black in testing Gap Between Cell to Cell and other defects. Here in inspection process of solar module, two images are captured (visual image – to check visual defect & EL image – to check electrical related defects.). Operators need to look on the image and identify defects. Challanges faced in this inspection process are below Many times all defects cannot be logged due to manual process and data accuracy is not there. Not able to see because of human incapability of taking right measure by eye. Wrongly assessed defective as operator is newly deputed or, customer criteria are change and he is not trained. To address these challenge MES (Manufacturing execution system) can be used for fast data logging and data accuracy. On defined Frequency MSA to be conducted for operator. AI enabled system can be installed to identify all defects and criteria changes can be assessed by AI software only.

Let us try to understand a few concepts before we begin. Rolled Throughput Yield = Multiplication of Yield of Steps 1 to Step n = Yield 1 X Yield 2 .... Yield % of Defective % = Defects / Total ideal output = 1- RTY So ideally, we are talking about Absolute Defect % which is a useful but complicated variable to get in a E2E business process. Let us understand why it is difficult? 1- A sum total of process is built across various departments, sometimes vendors, and details of interest are often not reported, or not a part of concern. Let us see when this can happen? 1.1- Headcount approval process application might need to pass 8-10 stakeholders. When it goes to HR, they might just look at the applications from Department X, and not track number of times rework is needed on the application. In this case, even if Department X is looking at # rework loops, overall getting this metric will be next to impossible lest it is agreed across all stakeholders. 1.2- Let us look at the example of a vendor who must update the details of this application on a portal within a maximum of 8H or same working day. Now they do not care about # rework loops/ interaction they need to make with the HR department. 2- Multiple process steps add defective %, because as a rule of thumb, each manual intervention brings in varied opportunities of creating a defective output. Having understood this, let us see how we can get this metric in the LSS projects. A- Alignment of the stakeholders at the time of kick off and highlighting the KPIs that the project is trying to address; or associated KPI tree that supports the project helps get everyone to the same page. B- Assigning responsibilities/ Process owner across departments will also help in getting necessary data and hence proceed with the project in a streamlined fashion.

In Six Sigma projects we usually try to improve the capability of the process in order to achieve a better DPMO(defects per million opportunities) than the current DPMO. In simpler words, how much the input material/effort is converted to quality output. And since, variation is a given in the near perfect world, it is challenging to always achieve 100% good output. That's where most of us LSS practitioners delve in order to improve quality. In general, the formulas for : - % Yield = Actual Good Output/Maximum Good Output * 100. % Defective = Actual Bad Output/Maximum Good Output *100. These 2 metrics could be used for anything ranging for products manufactured, quality level of any particular sub-operation, raw materials received from a partner, etc in the manufacturing industry as well as quality of service provided, no. of bugs detected, no. of failed transactions, no. of dissatisfied customers etc. in the Service Industry. When this metric of %Yield or %Defective is used across various industries, various operations, various metrics, various people at various levels of maturity towards their quality journey it becomes very challenging to measure it for their projects as there could be multiple inputs, multiple processes & reprocessing, multiple interpretations of the data etc. Some of the Challenges are: - 1. Complex or Multiple Operations When the operations are complex or has got multiple operations going on simultaneously or in sequence there are challenges in finding the reason why a defect or variation has occurred. For e.g. in the brewing industry, the % conversion of the sugars to alcohol has multiple steps and is based on multiple factors such as source of starch/sugars, filtering conditions, microbial culture, pH of water, and a whole lot of other variables and it is still now an empirical calculation that translates to the yield % for the processes. This causes difficulty in identifying why a particular batch had low yield. Rolled Throughput Yield supplemented by cross-functionally accepted definition of Operations Parameters can be an effective tool to use in such cases as it would further breakdown the process steps and help identify the operation/process where the loss actually occurred. 2. Varying Skill Levels and Maturity in Business Excellence Very few organizations have matured in Business Excellence principles. Thereby, still having some elements of functional disagreements and siloed work processes. In such cases, it becomes very difficult to calculate the yield as conflicts would arise on the method and parameters to calculate the %yield and %defectives and spark off a blame game. For e.g., In many industries the output of a particular process of the Production department is dependent on availabilities of utilities which may not be aligned with the goals the Maintenance & Engg departments. In such cases, it is important to get a sign-off on the parameters and methodologies to calculate the yields and the definition of the Defects , In-process parameters and Service Level Agreements right at the Define Stage itself or at the Toll Gate reviews from the Team Members/Project Champion/Subject Matter Experts. 3. Unreliable/biased/inconsistent/Insufficient/Wrong Data Sources There is always a challenge in getting accurate, consistent, unbiased or correct data especially when there is a human element to the data collection techniques when initiating a LSS Project. It's important to clean and categorize the data and undertake a Normality study or any other applicable data reliability study & MSA in order to ensure that the data can be used for analysis.