With the implementation of an approach to manufacturing which is focussed on flow of orders on the shop floor rather that utilisation of capacity, it is expected that all plant managers align their thinking and decision-making to the paradigm of flow rather than local efficiencies. While it is intuitively easy to understand and appreciate the flow paradigm, it is very difficult to give up the efficiency paradigm. Not many production managers can easily give up the efficiency paradigm, even though they understand how efficiency measures come in the way of the flow. Many of them would agree to implement an effective flow system that prevents workers from working on tasks that should not be worked on at a given time. But at the same time, they present a valid obstacle: “How do we ensure resources are working efficiently on things they are supposed to work?” They have a valid point.It is important to ensure workers are working hard and are not slacking off when they are working on the right things. We need people to be efficient. While the need is valid, the mechanism of using efficiency measures to conclude on efficiency is not, particularly in environments with a wide variety of SKUs, differing processing times and variability. In most of these environments, people have developed elaborate formulas to determine the productivity of each work centre. The complexity of the formula increases when one tries to nullify the effects of production variables to arrive at a ‘fair’ way to compare work centres. But it is almost impossible to include all possible production factors (particularly variability) in the formula. So, most of the time, the formula offers enough scope for people to ‘manage’ it. These measures actually encourage ‘cherry-picking’ (selecting the easier-to-manufacture at the cost of others), batching, and postponing set-ups. Every work centre worker and manager knows the way to manage these numbers by selecting some items over others. People can hide slackening of efforts on some SKUs by ‘cherry–picking’ other SKUs to arrive at a better number for the period. Hence, one can never know from these numbers if there was any slackening of efforts in manufacturing. So, one never knows whether people were really efficient just by looking at the numbers. Not only do managers have no clue about efficiency from just looking at these numbers, but there is further damage to flow caused by the ‘cherry-picking’ actions of the work centre managers. We all know what ‘cherry-picking’ does to the flow of orders. The cherry-picking of items across various local work centres leads to desynchronisation at the convergent points (not all items are available together for assembly) or material stealing at the divergent points. It leads to some items being produced ahead of time; work-in-progress (WIP) goes up on the shop floor, and so does the lead time, while on-time delivery drops.It is difficult to conclude whether these measures really prevent a slackening of efforts (we can never know whether numbers were achieved through hard work, genuine improvement, or mere cherry-picking), but they do definitely lead to higher lead times and missed due dates. So, in a way, one is paying workers incentives to actually mess up deliveries.Now we need a way not only to improve flow but also to achieve real efficiencies (to prevent slackening of effort). This calls for the Road Runner work ethics (If there is work, do not slacken and move as fast as possible. However, if there is nothing, sit idle but do not produce excess or ahead of time). Clearly, there are two parts of the Road Runner behaviour – efficiency (doing what needs to be done fast) and effectiveness (not doing things that should not be done). Do we need new measures to ensure these two parts of the behaviours? Like the throughput rupee days (TRD) for efficiency, and inventory rupee days (IRD) for effectiveness. In my opinion, both these measures also have their own associated problems. Due to the rupee value attached to orders, they can also lead to cherry-picking. The importance of an order can never be judged by the throughput value of the order. There can be so many different factors associated with the long-term importance of the order. Ideally, all the factors should be considered while taking in the order but once the order is on the shop floor, the due dates of all orders are equally important. Hence, the shop floor should only be managed with the buffer signals which provide the relative risk of missing dates of various orders (a priority system which treats all due dates as equally important).  If we cannot use TRD and IRD, how do we ensure the desired behaviours?Do we always have to invent a new measure to get the desired behaviour? Our experience shows that a ‘poka-yoke’ process is a much more powerful way to ensure the right behaviours while preventing the wrong ones. The process of choking the release prevents workers from working on unnecessary items – we achieved effectiveness with this process. On the other hand, visible colour signals in the shop floor and daily buffer management help ensure ‘real efficiencies’ on the shop floor. With daily supervision of reds and yellows, any slackening is detected early and corrected. (In a conventional plant with a significant month-end skew, the detection of delays and corresponding expediting is mostly towards the month’s end.) The colour signals increase sensitivity to reds and blacks. The entire plant suddenly looks ‘more visible’ and ‘transparent’ to everyone. The plant develops a self-correcting system because no one wants to be seen as the villain who created the reds and blacks in the plant (and, in turn, caused the team’s failure).  So we have a solution in place to replace the wrong measures: • Chocking the release for effectiveness • Daily buffer management for efficiency Now comes the next question. Do we fight to remove all the efficiency measures before starting the implementation? We know wrong measures drive wrong behaviours. Then, as a first step, we need to remove them to ‘correct’ the behaviour. Right?To touch measurements as a starting point is a waste of time. Even though people would be convinced by that logic, sheer inertia will hold them back. The best way to overcome the inertia of the old paradigm is to show significant results of using the new paradigm. So, the best way forward is to reach an agreement to implement the two important processes to a level where they become routine.So, we implement the new processes, which, in a way, mitigate the impact of wrong measures. But the question still remains: What do we do with the old measures?It depends on the way the measure is being used. If, along with the measure, there is a defined upper limit as a norm, we need to remove the upper limit as it becomes a self-fulfilling prophecy and prevents the release of capacity. In most plants, once plant output rises, the efficiency (even under the old measures) of the work centres also rises. Eventually, the measures will become redundant – they do not drive any behaviour, positive or negative. Now is the time to remove them. You will hardly have any conceptual opposition, provided one has worked out a way to compensate the workers.Refer to the article ‘<a target="_blank" rel="noopener noreferrer" href="https://www.vectorconsulting.in/en/research-and-publication/articles/standing-on-the-shoulder-of-the-giants/">Standing on the Shoulders of Giants</a>’ by Dr. Eli Goldratt to understand the difference between flow and efficiency paradigms.Throughput dollar days (TDD) is a measure for items done behind schedule, measured as days delayed multiplied by the throughput (sale value - material costs) of the order and IDD (inventory dollar days) is a measure for items done ahead of schedule and hence inventoried, measured as the throughput value of the order multiplied by the excess days in the finished goods (FG).Choking the release is the first step of Drum-Buffer-Rope (DBR) implementation for improving flow in production) wherein the material is released to the shop floor only a stipulated time before the delivery date.The time between order release and delivery, also called the production buffer, is divided into three colour zones based on elapsed time. The three zones -- Green, Yellow and Red -- provide easy guidelines to the shop floor on priorities and triggers for expediting.

BA Read More

Problem: Traditional efficiency measures distort performance by encouraging cherry-picking, excess production, and gaming of metrics. This undermines flow, increases work-in-progress (WIP) and lead times, and hides whether workers are truly efficient, while inertia keeps old measures in place. Solution: Replace misleading efficiency metrics with simple, process-based controls: choke material release to ensure only necessary work enters the shop floor, and use daily buffer management with visible signals to detect and correct delays. Transition gradually, allowing old measures to fade once results from the new system are evident. Benefits: Improved synchronisation and smoother flow, with reduced WIP, shorter lead times, and more reliable deliveries. A transparent, self-correcting shop floor that ensures real efficiency without reliance on flawed measures, while easing cultural resistance to change.

Key Takeaways

Shifting to flow-focused production from an efficiency-driven model is challenging, especially in a multi-SKU environment. One of the obstacles is the continued use of traditional efficiency metrics, which often causes plants to revert to practices such as 'cherry-picking', a behaviour that disrupts flow. However, there is a way to measure effectiveness without derailing flow.