Johnson V

Effectiveness relates to how good the outcome of the process is. Efficiency relates to how much resources the process consumes. An example from the software industry is the Review process. Effective review means one which finds most defects and leaks very few. The work product which is the outcome of the review will be of high quality i.e. it will have very few defects. Efficient review means using minimum review time to find maximum defects. It doesn't deal with how many defects will be left in the work product. Thus effective review needn't be efficient and vice versa.

"Whenever the consumer risk is decreased, the producer risk increases" - this statement is always true. Risk of making an error is proportional to the tolerance that we are prepared to allow for the test. A tolerance of 5% means a confidence of 95%. Let us go with the null hypothesis that all the samples are of good quality, and set the alpha at 5%. Now, if we increase alpha to 10%, it means the customer would need less evidence to prove that the product is bad, implying his risk has decreased. Conversely, the producer’s risk has increased.

The Voice Of the Customer are the needs, wants, expectations, and preferences, both spoken and unspoken, of business’s customers, whether internal or external. Examples of customer needs are a fast service, 24/7 support, no down time etc The Voice Of the Business are the needs, wants, expectations, and preferences, both spoken and unspoken, of the people who constitute (run) the business itself (e.g., shareholders, officers, or others involved in corporate governance). Examples of business needs are revenue, growth, market leadership etc There are many times when the wishes of the customers might be at cross purposes with that of the shareholders. For e.g. most customers would prefer to drive down billing rates or squeeze in as much scope as possible into the available budget. While the company management would want to increase billing rates and get more change requests. The key to reconciling these lies in transparent communication and the awareness that just as the customer is critical to the vendor, so is the vendor critical to the customer. A mature customer would want his vendors to be profitable and efficient, so that they don't try to cut corners in quality or service delivery. Similarly a mature vendor would pass on some of the benefits of process improvement to the customer even if the market forces do not force him to.

The Seven Wastes of Lean Manufacturing are; Transport Inventory Motion Waiting Over-Processing Overproduction Defects Let us discuss a few of these in the context of IT organizations. The biggest waste in an IT company, in my opinion, is Waiting. Poor process design or poor requirement gathering and project planning leads to a situation where half of the team members are waiting for weeks to begin their work. This is wasted effort. Since deadlines usually do not get extended, these hapless souls are then forced to complete their part of the work in the remaining available timelines, leading to the second biggest waste - Defects. These defects lead to Rework to fix them and some of the defects might even get leaked to the customer. Another big waste in IT companies is Inventory (or what is called as Bench). This is the surplus capacity that companies build up and maintain so as to start new projects as soon as possible. The key to improving organizational efficiencies is to get rid of these wastes by better planning. Team ramp up should be planned in such a way that team members would have work from their first day in the team. Upstream processes should have already generated enough input for them to work on by the time they join. This also gives sufficient time to the resource management teams to source the people, either from within the organization or from outside. In turn, this allows organizations to work with a lower bench capacity (or Inventory). Detailed planning and minute tracking is the key to eliminating these wastes and improving the profitability of the organization.

Rework is usually considered as a waste. In fact Cost of Rework is considered as Cost of Poor Quality and a lot of emphasis is laid on minimizing, if not altogether eliminating rework. However, there are some methodologies where some amount of rework is inevitable due to the iterative nature. Take Joint Application Development (JAD) for example. In this methodology, the vendor and customer collaborate to create the product. First the design team of the vendor creates a prototype, which is shown to the customer. The customer reviews/tests the prototype and provides feedback on how it can be improved. In this process, some part of the prototype would have to be dismantled and rebuilt. In other words, some amount of rework is inevitable. However this method is superior to having the design team spend a far greater time on design simply to avoid a small amount of rework. Even that is not guaranteed as the minds of the designers and customers work differently. The benchmark should be whether the cost of prevention is lower than the cost of rework + cost of failure. If that is not the case, it might be better to incur the failure and rework, rather than over-engineer the product by trying to prevent defects!

A "stable" process is one in which there are no special causes of variation. The entire variation present in the process is contributed by common causes and no special patterns are observed if one were to produce a control chart. Some patterns which indicate an unstable process are: a single point outside 3 sigma, 2 out of 3 successive points beyond 2 sigma on the same side, 4 out of 5 successive points beyond 1 sigma on the same side, 8 successive points on the same side of mean etc. A stable process would be indicated by the absence of any of these patterns. On the other hand, a "capable" process is one which is consistently able to produce outputs that meet specifications. i.e. the UCL and LCL of the process should be within the USL and LSL provided. Generally, stability is considered as a pre-requisite since a stable process can be predicted. A process that is unstable cannot be predicted, i.e. it could be currently producing conforming products but at any time, it could begin to produce non-conforming products. Thus an unstable process cannot be safely assumed to be capable, even if it currently seems so. The only occasion when stability might not be much of an issue is when the margin of safety is very large. i.e. the specifications are wide enough to accommodate even the wildest swings in performance. Consider the example of a pizza delivery joint having strict delivery time SLAs. The USL for preparing the pizza is 20 mins so that the delivery boys can comfortably deliver it to the customer within 45 mins. Assume that the pizza making process is unstable and though it usually takes between 12 to 15 mins to make a pizza, sometimes it has taken upto 22 mins due to some problems with the oven. In this case, the process is good enough (capable) most of the times but it cannot be safely assumed to be capable at all times due to the inherent instability. The owner of the joint should try to fix the problem with the oven so that the process is first made stable. However, if the delivery SLA is relaxed to 1 hour, the same unstable process might be acceptable, as the delivery boys will still have enough time to do the delivery even if the pizza takes 22 mins to produce.

It is widely quoted that "Correlation does not imply causation". However, over emphasis on this statement often leads us to overlook the fact that the correlation might be due to a reason. i.e. it is more than mere coincidence. The real clue could be that there is some underlying cause which is common to both the events and that is the reason why both are moving hand-in-hand. Consider an example from a study which found a high correlation between consumption of ice cream and death by drowning. Common sense would suggest that consumption of ice cream cannot lead to drowning. However, if one wanted to do a proper root cause analysis of the drowning deaths and find ways to reduce them, it would be informative to explore the correlation with ice cream consumption. The real reason why both are correlated is that they are both related to the weather. Ice cream consumption increases significantly during summer. Similarly the hot weather leads to more people going for a swim, in turn leading to more number of drowning cases. Of course, there would be other causes of drowning which are unrelated to weather like the person not knowing how to swim! But that would not explain why more people don't know how to swim in certain months compared to rest of the year!! Thus exploring the correlation can give insights into the true root causes, which can lead to better targeted actions. In this case, it could be appointing more life guards during the summer months as compared to the rest of the year.

VOC is usually considered sacrosanct. After all, customer is the reason why we are in business. However, there are times, when we need to be firm with customers, even at the risk of temporary displeasure and lower customer satisfaction scores. One common example from the IT industry is the customer expectation of immediate turnaround of change requests. Irrespective of how late in the day the CR has been raised, it is always expected that it should go into the next release. Many vendors bend over backward to accommodate these requests and crush their employees with impossible deadlines, at the risk of not performing the necessary impact analysis and finally releasing unstable code into production. Mature companies would push their customers back insisting on proper time frames to follow the necessary processes so that only good quality, stable code is released. Customers are not born mature, they become mature over time. Thus many customers may be at lower stages of maturity than the vendor. In such cases, it is the duty of the vendor to invest in educating the customer and raising their maturity level for the long term benefit of the relationship. Even if the customer initially sees it as bureaucratic, persistent education will lead them to realize the value of what the vendor is proposing. On the other hand, if the vendor gives in to the demands and ends up with production issues, the customer will end up being more unhappy than if the request was denied in the first place.

Discrete data means that the values can be taken only from a finite set. For e.g. while marking students in an exam, examiners usually give a whole mark or half a mark. Thus for a 100 mark exam, the set of possible values are 0, 0.5, 1, 1.5 ..... 99, 99.5, 100. While this would fit the definition of discrete data, practically we can consider it as a continuous data set. As a rule of thumb, if the number of distinct elements in the set exceeds 10, the data set can be considered as a continuous data set.

These three terms were made very popular by the ISO 9001 standards couple of decades ago. "Correction" refers to the actions that are taken to rectify the non-conforming product after a defect has been found. For e.g. fixing the bugs in a piece of software. "Corrective action" refers to the actions that are taken to prevent the recurrence of the non conformity or defect. Note the emphasis on the word recurrence - it means that it has occurred once but that we want to prevent it from happening again. For e.g. updating a coding standards or guidelines document to include the most common causes of coding errors so that developers can avoid them in future. "Preventive action" refers to actions that are taken to prevent the occurrence of errors in the first place. This requires the ability to foresee what kind of errors are likely to occur and taking action in advance. For e.g. increasing server capacity in advance of launching an online sales campaign so that the increased volumes do not bring down the server. These actions also fall into the realm of proactive risk management, which then begs the question "why do we need a separate category called preventive actions?". Many QA professionals have tied themselves up in knots trying to distinguish between "corrective" and "preventive" actions with creative distinctions like: it is considered corrective, if the action is taken in the same project in which the error occurred, but if it is taken for a new upcoming project, then it is considered preventive. Some professionals have taken a little more harsher view saying, if the action is in response to an error that has occurred in any part of the company, then it can only be considered corrective, and not preventive. Which means that the action has to be in response to an event in another company or to something which has never happened before! Perhaps, in response to this lingering confusion, the ISO has dropped the term "preventive action" altogether from its 2015 version of the ISO 9001 standard. There could be situations where preventing the error or incident could be cost prohibitive and the result of a failure is not severe. For e.g. in the earlier example, if the cost of additional hardware is unlikely to be justified by the increase in sales, it might be a better idea to just bring the server up again, in case it fails, rather than spending money to prevent a failure. In this case, correction is a better option than prevention. Another case where prevention could be unnecessary is when the probability of the incident is extremely low, For e.g. not many would spend money to build earthquake resistant buildings in a seismic zone 1 or 2 region. However, in this case, the individual or organization is accepting the risk of a building collapse. Correction is not a better option than prevention, in this case.

To imagine that the concept of checksheet would be removed from the study of QC tools is like imagining that manual addition and multiplication would be removed from the syllabus in primary school because calculators and computers have become so common! Of course, the pen and paper will give way to electronic data capture means, like in most walks of life, and the new mechanism might not even involve check marks, but a serious student of statistical quality control would do very well to understand the different types of checksheets and their purpose with a view to bringing in innovation in how the data can be obtained more efficiently and with better accuracy. The new age checksheet might be an app in a mobile device and the check mark might be a tap on the screen in a designated area. This will also eliminate the time taken to process the data manually, as well as computational errors. Or it could be an automated robot inspecting a sample, recording various measurements and noting down those which are outside the acceptable range, by means of a message to a central server. Or it could be a scanner device scanning samples as they pass on a belt in front of it, recording the bar codes as well as all measures of interest in a fraction of a second. Needless to say, the data will be streamed to a server for processing. In such a scenario, we might end up collecting a lot of continuous data, which can be used as is or converted to discrete form as a checksheet would do. The concept of checksheet will continue to be relevant but the manner in which it is implemented will test the frontiers of human imagination!!

While just-in-time techniques or pull mechanisms help in reducing inventory and are considered more efficient, sometimes it may not be practically possible to have just-in-time techniques due to operational constraints. For e.g. it might make economic sense to order parts or raw materials in bulk, especially due to lower per unit shipping cost and bulk purchase discounts. In such situations, the money saved in terms of inventory might just get consumed in terms of higher shipping costs! Secondly, it is not always possible to have a perfect JIT technique that can deliver the product to the customer exactly when he wants it. Usually JIT techniques would involve some waiting time for the customer and in today's hyper competitive scenario, very few companies like Mercedes might be actually able to command customer loyalty to the point of making them wait for a product, even as competitors are flooding the market with similar products of comparable price and quality. Many a times, it is the logistics that makes a product successful, even more than its features. i.e. easy availability near one's place of residence, low waiting period, low price made possible due to large scale manufacturing etc. In such situations, flooding the market might be a better strategy than pull based manufacturing!

The concept of value is not universal. It depends on the perspective of the customer, situation and the nature of the industry in general. Let me illustrate a few examples where the traditional definitions might not hold good: 1. During these times of increased security risk, we have authorities perform secondary checks at ladder point in many airports. Going by the standard definition, this would be a non value added activity as it would mean that the first check was not done correctly (not first time right!). But I would argue that it is value adding as it reduces the probability of dangerous elements gaining access to an aircraft and putting the lives of passengers at risk. Here, the "value" is the reduction in probability of an undesirable event, though it might seem to be intangible. 2. Consider the delivery of a luxury car in a showroom. We can see the sales representative making a detailed inspection of the car to make sure that it is spic and span, no loose ends are visible etc. Mind you, this is not a technical check on the functioning of the car, which would have been done already by the technicians. This would be followed by the sales manager making another final inspection of the car to check for neatness and presentation. The multiple checks might seem redundant and non-value-adding to a quality professional but for an extremely discerning customer, a stray mop lying on the back seat might be enough reason to never come back to that showroom again. It would be silly to ask the customer to pay for each round of inspection as the customer is already paying a premium. The "high level of service" is an integral expectation. The "value" in these checks is in ensuring customer delight. 3. In the software industry, it is common to have multiple rounds of testing - unit, integration, system, performance, security etc. For especially sensitive systems like financial services, these tests would be even more stringent as the loss arising out of a security breach or unplanned system downtime could be disastrous. These tests are not transformational, rather they are destructive.In such cases, having redundant systems and multiple quality checks would be the "price to pay" for good quality of service.

Kaizen stands for small continuous improvements. It involves everyone taking up small improvements in their methods and procedures in their day to day work. Kaikaku stands for radical improvements. It involves big ticket changes aka initiatives usually sponsored by someone senior in the organization. Kakushin stands for breakthrough innovation. It can be the result of technological progress, market forces, customer demands or organizational R&D. Ideally a combination of all 3 techniques would be desirable to achieve optimal results. Kaizen is for short term results, Kaikaku for medium term results and Kakushin for long term results. Absence of Kaizen could lead to customer dissatisfaction as service levels will stagnate. Absence of Kaikaku could lead to loss of market share as other competing products will start eating into your share. And absence of Kakushin could wipe out the company like in the case of Kodak. The 3 can complement each other very well. Consider the example of a telecom service provider. Once a big innovation is brought in (Kakushin - for e.g. 4G), small improvements (Kaizen - for e.g. adding more cell towers or harmonizing spectrum) can be made over a period of time to stabilize the service. When the need is felt for a more substantial change, Kaikaku (for e.g. video calling) can be used for taking the service to the next level. Again Kaizen can be used to stabilize the operations. This cycle of Kaikaku and Kaizen can be carried out number of times till technology or competition forces another Kakushin (this duration will be different in different industries).

Both gurus are right in their own way! The first quote is easier to explain. It is exhorting managers to try and use measurements in everyday life to take decisions. For e.g. budget variances, productivity, cost of rework, and of course the business metrics like sales growth, profit growth, PAT, employee turnover etc. Measuring each of these gives a handle to see whether we are improving in these areas or worsening or remaining in the same place as before. To give a silly example, one would hardly find a CEO telling his employees that the company has done extremely well this year without comparing the current year's numbers with the previous year. Or a Quality Head claiming that the innovations he introduced, have improved efficiency without providing any numbers to back it. The problem arises when one reads too much into the quote, literally, and avoids managing things where the measurements are difficult. Drucker is not advising us to stop managing things that cannot be measured. He is only telling us to measure (as much as possible) whatever we want to manage! Deming, on the other hand, is saying that there is a life beyond numbers too! He says that we should not lose sight of key parameters, just because they are difficult to measure. Such parameters are extremely common in the area of human resources. For e.g. how does one measure the mood of an employee? Mood does have a material impact on the quality of deliverables produced in a knowledge economy. Or how does one measure the loyalty quotient of an employee? Such variables are qualitative and managers need to use management / behavioral theories to devise suitable approaches to manage them. Even in public administration, such examples are common. How does a police officer judge the threshold beyond which the crowd might turn violent? It cannot be a simple measure of the size of the crowd, it could be the extent to which their sentiments are aroused on a particular topic. In all such areas, managers need to still take decisions anecdotally or based on conventional wisdom. Gifted managers have a 'nose' for such variables. They can smell it, even if they can't measure it!!

Let's start with the basics. We assume we are doing the root-cause analysis for a problem that we want to prevent, like fire in the house or attrition in a company or a heart attack. A cause which is neither necessary nor sufficient, should probably just be ignored. Very little point in pursuing those. Even if you do manage to prevent such causes, they will probably have no impact on the problem that we want to prevent. Second, the necessary conditions should be explored. Eliminating a necessary condition is the surest way to prevent a problem, so if it is cost-permissive, we should try to kill the necessary conditions. For e.g. oxygen is a necessary condition for fire, so removing oxygen is the surest way to douse a fire. Mosquito bite is a necessary condition to spread malaria, so eliminating mosquitoes (or at least preventing them from biting!) is a sure way to attack malaria. Out of the two options, necessary and sufficient versus necessary but not sufficient, they should the weighed on the basis of practicality and cost-effectiveness. Either of these options will be equally effective in preventing the problem since the absence of even a single necessary condition will cause the problem to disappear. Lastly if it is not possible to sufficiently neutralize any of the necessary conditions, we may turn our attention to the sufficient but not necessary conditions. Eliminating these, of course, helps us solve our problem, but there is a catch!! We must eliminate all of the sufficient factors to eliminate the problem! Of course, life is not all so bad!!! So eliminating a few of the sufficient conditions also might significantly reduce the probability our problem occurring! And therein lies the reason why we keep popping anti-cholesterol pills...

Personal excellence: This is a personal trait. It may be God-given or achieved through rigor of practice and experience. We often hear of people who have excelled in one area or another. For e.g. Sachin Tendulkar, Lata Mangeshkar etc. This kind of excellence is inherent, not everyone can achieve this. However at a personal level, it also means ‘doing the best I can’ which means putting that extra bit of effort and not being satisfied with mediocrity. It also means becoming more competent tomorrow than I am today Process excellence: This involves putting less emphasis on the individual and allowing the process to take over. This can be observed in all kinds of industries, both traditional and modern. Even supposedly creative activities like R&D and innovation are channeled into a process framework so that the gains obtained are sustained. Continuous improvement led by small changes cause the performance of the organization to continuously improve. Learning and feedback loops help to keep the processes relevant. These kind of organizations are poised for sustained growth. Operational excellence: This is very common in traditional industries where the process of manufacture or service provision has been perfected through years of research and repeated execution with little variation. The workers only need to follow the instructions. Typically such organizations can achieve good performance with even moderately skilled workforce. for e.g. dabbawalas of Mumbai. The main difference with respect to process excellence is the response to change. An organization might operationally excel in one service but a foray into another service could be chaotic. Comparatively, a process excellent company would handle change also methodically. Business excellence: Here the focus is on outcomes, not on inputs or processes. Makes most sense in a competitive world where return on capital has to be justified to stakeholders. In such a company, process or operational excellence would not be an end in itself, rather they would be constantly subjected to the touchstone of RoI and cost benefit analysis. Methods like Six Sigma focus on business excellence and even traditional process improvement models like CMMI have begun to incorporate business goals into their process areas. Having said this, only a company with a solid base of process and operational excellence can have continued business excellence.