by David Jacoby
Continuous improvement (kaizen). The system is geared to provide a learning environment where each mistake turns into an opportunity for improvement, so that eventually no error should be repeated twice.
The systematic elimination of waste (muda in Japanese) and the related drive to ensure that every action and every effort drives towards adding value for the end-customer.
People-centredness. The TPS only works if people believe in it and execute it; it cannot be mandated and it cannot be managed virtually through information technology. People must internalise the norms and the values, and bring them to bear on the myriad problems that they encounter each day.
Simplicity, which means fewer breakdowns than in a complex system, and reliability: these are central to the TPS. Six norms guide execution within the TPS:
The whole system works to a rhythm (takt time), so that when everybody follows the same rhythm, it is harmonious and easier to control.
Control, stability, reliability and predictability: these are critical to the success of continuous improvement initiatives since a process that is in control can be adjusted much more easily and quickly than a process that is out of control.
Workplace organisation and cleanliness: these are prerequisites for control and stability, since dirt and clutter introduce randomness that causes unstable processes and results.
A one-piece flow: this is the best way to validate that there is no waste in the system. A wasteful process never produces a one-piece flow.
Repeatable processes: these ensure continuity and consistency of purpose and practice over time.
Worker empowerment: ensures the longevity and durability of the system, since top-down and autocratic systems are less effective at handling the wide range of problems that may occur.
The TPS is responsible for the creation of many of the tools that are associated with lean management, as follows:
The pull-based demand trigger eliminates waste by focusing all effort on satisfying customer needs rather than a forecast, which is inevitably erroneous.
Just-in-time (JIT) production minimises the waste that may occur when customer demands change, by eliminating buffer inventories throughout the pipeline.
Jidoka prevents problems before they occur, thereby improving system stability and reducing the need for problem diagnosis and remediation.
Visual controls, including kanban cards (physical cards that are placed at the end of a batch of inventoried items; the card itself triggers replenishment rather than an information system), ensure universal, real-time and easy-to-update access to information about the pace of production (takt time) and the status of errors and remediation efforts.
Capacity balancing and level loading ensure small lot sizes and hence minimum waste from changes in demand.
Root cause problem-solving, including diagnostic tools such as plan-do-check-act (PDCA) and the Ishikawa fishbone diagram (see Figure 6.2 on page 80), enables the working team to diagnose and resolve errors and restore normal operation following disruptions.
Elements of synchronisation strategy
Synchronisation strategy includes the following elements:
Constraints management and throughput analysis
Pull-based demand trigger
Just-in-time (JIT)
Perfect order fulfilment
Make-to-order (MTO)
Optimal inventory placement
Sales and operations planning (S&OP)
Collaborative inventory management
Everyday low price
An anchor player that ensures stability
Shifting demand and capacity
Better forecasting, less emotion
Risk mitigation
Constraints management and throughput analysis
To synchronise throughput in the supply chain, capacity must be aligned at each step in the process, or bottlenecks will constrain the output. For example, for a computer manufacturer that makes laptops in six steps – production, assembly, configuration, testing, labelling and packaging – the throughput will be constrained by the operation with the least capacity. As shown in Figure 7.1, testing is the bottleneck operation that constrains the system throughput.
Figure 7.1 Throughput capacity analysis – example
Source: Author
Aligning capacity across the partners in a supply chain requires collaboration and advance planning. Many countries have capacity limitations in their port and rail networks, which constrain the volume of imports and exports that can flow. The ports are creating incentives for off-hours pickup, upgrading to equipment with higher throughput and changing labour rules. The railways are transferring to inland transfer stations to alleviate congestion at the ports, and changing the way they load containers onto intermodal trains in order to speed up loading operations. However, some of these investments can take years (delivery of new cranes and locomotives, for example), so the ports and the railways need to collaborate or else they risk having misaligned capacity.
Theory of constraints
The theory of constraints addresses bottlenecks in operations and offers a systematic process for alleviating them. The concept is based on the premise that every process has one binding constraint (based on the throughput analysis above), and that throughput can be increased by eliminating that constraint. The process involves clarifying the mission of the organisation, identifying the binding constraint, eliminating it, aligning other processes to the new throughput levels and then pursuing the next constraint.
The theory of constraints is a hallmark in manufacturing thinking because it highlights the extent to which manufacturing has historically been driven by management’s desire to lower average standard unit costs by increasing the volume of production so as to amortise fixed costs over a larger number of units, and management’s behaviour in increasing plant capacity utilisation, both of which historically resulted in overproduction and excess inventory. It highlights the accounting mirages that have been at the root of decades of overproduction.
BNSF Railroad in the United States underwent a detailed study of bottlenecks in its end-to-end supply chain, using the principles of the theory of constraints, when the tide of Asian imports streaming into the country brought the transportation system to its limits. To enable the free flow of cargo from the West Coast to population centres, which were mostly in the east, it identified the port of Los Angeles/Long Beach and its Chicago gateway and hub as bottlenecks, and took actions to eliminate the constraints. The actions included, for example, increasing the percentage of railcars that it loaded on-dock, requiring terminal operators at the ports to release full trainloads instead of partial trainloads, and differentiating between northern and southern trains by specifying that they must consist of cars destined for either northern (Kansas City, Chicago or the north-east) or southern (St Louis, Texas, Memphis, the south-east and Florida) stations. After eliminating the bottleneck in Los Angeles, it turned to the congestion at Chicago and the possibility of making joint public–private investments to alleviate that bottleneck.
Bottleneck elimination can occur at a macro level as well. When raw materials such as ore and minerals are in short supply, downstream companies consider acquiring the source in order to ensure supply. Nampak, a South African packaging company, acquired a paper mill in order to eliminate a bottleneck in its supply chain.
Spares management
Companies that operate in capital-intensive businesses like transportation and oil and gas need to place expensive spare equipment in the right place to minimise the risk of downtime at the least cost. Optimisation of the number and placement of highly capital-intensive spares is a branch of throughput analysis that requires special tools. Statistical analysis of failure is required to know how many spares are required and where they should be placed to maximise the chance of their availability. Logistics plans are needed to ensure their movement to the points of need. For example, aircraft engines are needed in repair centres around the world, but given their
high capital cost there is a limited number that can be available at any time. FedEx conducted detailed modelling that permitted it to more precisely forecast the number of spare engines needed, so that it would be sure to have enough to keep planes flying at the Christmas peak. It discovered that it had more than it needed and was able to eliminate spare engines.
Pull-based demand trigger
Inventory management techniques and technologies frequently reduce inventory by 20–30% in major companies, which often equates to 1–3% of sales. Single and multi-echelon inventory management software applications determine the optimal amount and type of inventory to hold at each location. A regional utility reduced its inventory by 28% by implementing a software model to set the optimal reorder points.
The overarching shift that has driven dramatic inventory decreases is a move from push to pull demand triggers. In the old push-based system, the product was pushed forward towards the point of sale, based on forecasts. Since the forecasts were inevitably wrong, there was excess inventory in the field.
Pull-based demand triggers used not to be feasible, but today they are the principal mode of replenishment, largely because of the improved ability of information systems to transmit production, sales and inventory data throughout the supply chain. The retail and automotive (see next section) sectors offer two examples of how pull has changed SCM for ever. In the retail industry, the real-time availability of point-of-sale data and the ability to reconcile consumption with inventory levels have made pull-based replenishment the only way to run a modern retailing operation. Wal-Mart has been able to reduce its in-store inventory from 10–12 weeks to 3–4 weeks by transmitting actual stock counts via wireless handheld devices to the distribution centres, and replenishing weekly. The improved accuracy has also enabled more direct-to-store shipments (facilitating the DC bypass component of rationalisation strategy) by reducing the need for a safety stock buffer.6
Just-in-time
JIT is a logical extension of lean thinking, since having material in advance of when it is needed is wasteful. It is addressed here rather than in Chapter 6 because it results in savings in working capital rather than in the cost of goods sold.
JIT, when properly implemented, significantly reduces or even eliminates inventory requirements, which explains why the concept has been so widely implemented. Of the three principal types of inventory – cycle, seasonal and safety – JIT primarily reduces safety stock, which is the buffer that protects the business from the effects of variability in demand and supply. Cycle stock is meant to satisfy consumption while orders are being processed and are en route (technically, that would be in-transit stock), so if the JIT supplier is moved to a nearby or adjacent facility, JIT can also reduce these stocks. Seasonal stock is meant to push stock to market in anticipation of seasonal sales.
In the automotive industry, component suppliers co-locate next to assembly plants in order to offer JIT capabilities:
General Motors set up a large supplier park next to its assembly operations in Brazil to ensure tight communication and JIT replenishment.7 Fiat and other automakers have followed the same principle.
The “Smart Car” was another auto industry experiment with lean production and distribution. The company (Micro-Compact Car, or MCC), which was originally a joint venture between Mercedes-Benz and Swiss Watchmakers and was eventually taken over by Daimler Benz, attempted to build cars in 7.5 hours. The trick was modular assemblies and postponement. The chassis, power train, doors and roof, electronics and cockpit were modules that could be rapidly assembled.8
InterSwitch in Nigeria implemented a JIT approach to its electronic funds transfer (EFT) solutions across about 20 banks in 2004/05. Seyi Oluwehinmi, team lead for SCM, explains:
We employed a JIT approach. We made sure that equipment and supplies were delivered just at the time they were needed, hence eliminating high inventory costs and mitigating the risks of tied-up capital.
One-piece flow concept
JIT is fundamental to the concept of the one-piece flow. One-piece flow is an ideal state based on a philosophy of production that replenishes upon use and in the smallest quantities possible, with an ideal order quantity of one unit. The ability to place more, smaller, orders more frequently addresses the source and the magnitude of the bullwhip effect related to order-batching.
Kanban and reliable replenishment
Kanban, the use of a visual signal to trigger replenishment, has helped to convert forecast-driven managers to JIT by making them go back to the basics. Workers have a kanban card, which has all the necessary information listed on it. In a repetitive production operation, workers learn the information on the card by heart.
Rapid replenishment eliminates the need for buyers to hold safety stock for the demand during the lead time of the order cycle. Companies can also enable inventory reduction by increasing the reliability of the timing of delivery, especially on products for which the transit time is long. In a continuation of a trend that started in the retail environment, companies are increasingly dictating delivery time windows to their suppliers and the time windows are getting shorter: 15 minutes is typical in retail trades in the United States.
Making the delivery straight to the point of consumption removes all delivery uncertainty from the buyer’s inventory calculations. Automotive parts suppliers often co-locate next to the automaker’s plant in order to remove delivery time uncertainty, and in the Japanese model they are responsible for delivery not to the dock door but directly to the worker on the line who is doing the assembly. To minimise handling, companies carefully study direct delivery options that bypass distribution centres. Third-party logistics (3PL) companies set up services that collect product at its source, package it in the sequence in which it will be unloaded when it arrives, consolidate it and ship it direct to the store using multiple modes of transportation. Direct delivery service involves careful planning and meticulous execution, especially for complex global and multi-modal services.
Takt time
Takt time is the rate at which the production operation produces output (for example, four cars per hour). But it has a much deeper connotation than the word “rate” conveys. Takt time is the rhythm or heartbeat of the operation, and the musical and visceral connotations are deliberate. It is the pace at which everybody in the process works in order to keep capacity aligned, at the pace of the bottleneck process. Without takt time, there would be inventory in between work stations and shortages of material (or services) between others.
Level loading
Level loading is the practice of scheduling work so that a little bit of every product is made every period, rather than in large batches of one product at a time. Applying it in an office environment might involve taking a day-of-the week schedule with five activities (one performed each day of the week) and scheduling that work so that one-fifth of each activity gets done each day. This requires more changeovers, which forces quick set-ups (see single-minute exchange of die, below). However, level loading allows the quicker identification of errors and defects, ensures the balanced use of resources and mitigates the impact of unforeseen circumstances. For example, if one person handles a given task once per week and he or she is off sick that day, the task will not get done until the following week, but with level loading every problem and irregularity is corrected within one day.
Set-up time reduction and single-minute exchange of die
To attain level loading, companies must frequently change over from one product or service to another. When companies move from infrequent changeovers to frequent changeovers, they often realise that their current pace of changeover will never work with the frequency of changeover that is required for level loading. Therefore, they need to learn how to change over quickly. The concept of single-minute exchange of die (SMED) stems from the automotive industry (Toyota, specifically, which uses dies in stamping presses to produce stamped parts like hoods), in which changeovers were reduced to a minuscule period of ti
me in order to facilitate flexible manufacturing.
Despite its automotive legacy, the concept of SMED applies to all industries. Even in pure service industries like banking and consulting, lean theory prefers flexibility and responsiveness to efficiency, on the basis that there should always be as tight a linkage as possible between demand and the pace of production, and if possible, a one-piece flow.
Postponement
Retailers with global supply chains often find that when goods arrive in stores demand for them is significantly different from what they had predicted when the store’s order was placed, which is often eight or more weeks beforehand. To mitigate the difference, most leading retailers establish an interim level of product build-up (variations of assembly or packaging) that can be finished off late in the supply chain according to current demand patterns. This strategy is particularly useful in fashion industries. For example, apparel importers can dye shirts upon entry into the country rather than at their point of manufacture, thereby eliminating this industry’s largest element of variability (colour) that causes inventory shortages and excesses.
Using real-time feedback loops between the stores and its production facilities, Italian clothing-maker Benetton can distribute the styles and colours that sell the most to the stores, while avoiding excess and obsolete inventory of pre-dyed styles that are not popular with shoppers. Sources at the company say it saves money through the avoidance of clearances and the lower inventory carrying cost involved with stocking undyed garments (the material has less book-keeping value than finished goods, and it needs to hold fewer units in stock to satisfy the same demand because demand for the undyed garments is less variable than demand for specific styles and sizes).9
A well-known fine writing-instrument-maker imports pens in bulk and creates multipacks of pens in the United States. This is an example of postponement that saves on inventory costs and allows the company to have a high fill rate by being more responsive to the market. Sherwin-Williams, a maker of paint, mixes paint at the point of purchase rather than at the point of production to reduce the variations of inventory that would have to be kept in stock. CEVA Logistics, formerly TNT Logistics, ships laptops from China to its German distribution centre, where it configures them for sale in European markets, to reduce the number of finished goods SKUs that its customers need to keep there, and to reduce theft.10