The Goal by Eliyahu M. Goldratt
Reviewed by James H. Lui (copyright 1997)
A novel of extreme significance to any manager or parent who faces operational efficiency challenges.
The Goal, by Eliyahu Goldratt, is a story about a young, perhaps late-30's or early 40's, upwardly mobile executive named Alex Rogo who personifies the profile of a large number of typical people involved in business administration. He possesses an implied engineering undergraduate degree and an MBA. He is married and has two school-age children. Alex began his career at UniCo fifteen years ago as a project engineer under the supervision of Bill Peach. Peach has ascended the corporate hierarchy and attained the position of division manager, a vice-presidential level. Alex now works under Peach as a plant manager of a manufacturing factory in Bearington, his home town.
The factory produces machined assemblies furnished to other plants in the UniWare division as components of end-items, and also sold directly as spare parts assemblies to larger end-user customers. He has a competent and well-trained staff, each of whom is a subject matter expert in their respective functions. UniCo is led by highly experienced managers, each of whom has been exposed to increasingly broadened responsibilities for different UniCo functions.
UniCo's customers want quality products delivered on-time at a reasonable price. In order to control prices, UniCo's management has become increasingly efficient at reducing costs in several operational areas. Financial performance reporting is provided at every level of production in exquisite detail so that functional cost budgets can be produced and managed with great precision. Automation has been introduced at several levels to increase production efficiencies and have resulted in unprecedented savings in production time and station production rate. Even with the introduction of the robotics, somehow management has been able to satisfy the stringent demands of its unionized workforce to achieve parity in its labor relations.
Alex's plant also shares some other not-so-complimentary common traits with real-world organizations. He has orders bordering on two-months behind scheduled delivery date. He has over $20 million in unsold finished goods inventory sitting in a local warehouse. The items that can be delivered are being expedited through the factory with overtime and special handling involved at every step. As a result of the delivery slippage, sales are slowing down, material costs are skyrocketing, and every efficiency metric is heading downward at an alarming rate. Worst of all, the division is facing an increasing cash shortage -- if it continues, it won't be able to pay its bills, or meet its payroll.
Alex has to do something. He is politely informed that if these trends cannot be turned around in three months' time, the end of the fiscal year, UniCo has no other alternative than to shut down these terminally unprofitable operations and sell them off to the highest bidder. Clearly obvious is that the new owner would not want to keep an unprofitable or ineffective manager around to run the same organization. Alex's job and career are now on the line. As if this weren't enough, all the time Alex is spending trying to solve his plant's problems is taking its toll on his family. He's bringing his problems home with him and displacing his frustrations and anger on those who love him the most, slowly tearing his family apart.
Fortunately, Alex had, by pure chance, run into someone who introduces him to another way of thinking about his situation, his old physics professor Jonah. Alex, eager to demonstrate that his education was not wasted, describes his firm's investment in automation and how it has led to increased productivity. Jonah, in turn, queries Alex about some key identifiers of productivity such as decreasing inventory, reducing expenses, and selling more products. Although Alex can't affirmatively respond to any of them, at that point, he was sure he was only dealing with some annoying problems, not the demise of his entire organization. Jonah takes a moment to reveal the fallacy of logic in Alex's reasoning -- having accepted many things without questioning the common sense in their purpose and application. Late for a flight, Jonah leaves Alex to ponder the concept of productivity and what it means to a business. Productivity is defined as accomplishing something in terms of goals. But what is the goal?
Alex has a thinking session with his plant controller, Lou. They agree that the goal of any business is to increase net profit while simultaneously increasing return on investment and cash flow, or basically to make money. Unfortunately, neither one has a definite answer as to how to do all three at the same time.
Alex turns back to Jonah. Jonah reveals that operationally, a business must increase throughput, while simultaneously reducing inventory and operational expenses. Throughput is defined as the rate of sales. Inventory incorporates the money invested in things intended to be sold. Operational expenses include all money spent to convert inventory into throughput. When Alex asks how he's supposed to measure these elements, he is reminded that, "We are not concerned with local optimums," referring to conventional cost accounting reports. Jonah is a busy individual and must again hurriedly part from Alex without delivering all the answers. We begin to see Jonah's socratic teaching method appearing here -- Jonah will provide Alex with the constructs for him to divine the answers to the underlying concepts. Alex must think through the answers on his own.
Alex takes time to think about how the three measurements would apply to his particular situation. He gathers together his focus group which will eventually tie operational solutions together for the entire plant -- Lou, the controller, Bob, the production manager, and Stacey, the inventory control manager. Together they reveal that the source of the massive inventory is the result of overproduction of unnecessary parts being made to artificially keep the efficiency metrics up. Instead of producing what is needed for sales, the plant is producing every single part every stage can handle. Capacity for producing needed parts is unavailable because it is being tied up producing the excess parts.
Alex sits down with his team to get them all to understand what he has learned from Jonah. They grasp the measures easily but are clouded by previously learned rules of productivity such as, "…if we don't produce, our efficiencies go down," and "…we can't have people sitting around idle because that's unproductive." At this point, something else is still needed, and Jonah is the agreed upon source.Alex learns from Jonah several concepts which are polar opposites from what he has been told before about business operations. Jonah reveals that:
Jonah points out that the only way to create excess inventory is by having excess manpower. By trimming excess capacity to cut expenses, without reducing inventory and increasing sales you trigger downward throughput and increased inventory. If you attain only one or two of the three elements of productivity measurement, you are not working towards your goal - to reduce operational expense and reduce inventory while simultaneously increasing throughput.
Jonah reveals that a balanced plant is where the capacity of every resource is balanced exactly with demand from the market. However, the closer you come to being a balanced plant the closer you are to bankruptcy.
Mathematical proof exists that when capacity is trimmed to marketing demands, throughput goes down and inventory skyrockets. The carrying costs of inventory, an operational expense, also goes up. This increase tends to offset the savings presented by the original attempt to lower operational costs through labor reductions. If you continue to trim capacity to demand, demand continues to drop, carrying costs go up, and eventually you have no more market left for a mountain of inventory.
Two specific phenomena are identified which cause this effect - dependent events and statistical fluctuations. Dependent events are a series of events that must take place prior to another one beginning, or in other words, the subsequent event depends on the ones prior to it. Statistical fluctuations are the result of certain types of predictive information that cannot be determined precisely. These fluctuations influence prediction of error percentages, market demand estimates, and attempts to measure productivity.
Alex takes his son, David, on a hiking experience with his Boy Scout troop. During the hike, he notices that the line of hikers exhibits an odd pattern of stretching further and further apart the longer they hike. He notices that one particular scout, Herbie, is probably the slowest hiker in the group. Herbie appears to be holding up the remaining scouts behind him. According to management science, even though these hikers are all at different rates, their average rate of progress should be estimable. This average rate should become the nominal rate of progress for the entire troop. Instead the troop is making final progress, or completing the hike, at the rate of its slowest member, Herbie. Herbie is being left behind at a longer and longer stretch because he can't go any faster.
The boy scout hike is analogous to a set of dependent events subject to statistical fluctuations. Over time, the fluctuations do not average out, but rather accumulate because the influence of dependent events limits the opportunities for gain fluctuations. The length of the line of hikers becomes comparable to the total production time of a process.
Alex tries an attempt at re-balancing the capacity by placing Herbie at the front of the line, that way the production length won't tend to stretch out as before. It doesn't stretch, but it's still going as slow as Herbie. Herbie must be made faster, or gain throughput capacity, in order for the whole line to gain throughput.
Herbie's backpack load is lightened and distributed among the troop and the entire troop doubles its pace as a result of the change.
Alex's observations as a result of the hiking experience can be summarized as follows:
Returning from the boy scout hike, Alex's wife, Julie, has decided to resolve her own conflicts by leaving him. From a timing standpoint, Alex and Julie's separation is introduced at the critical moment when Alex needs the most time to dedicate himself to solving the plant's problems. This time-out period between Alex and Julie probably does more to save their relationship than would have been accomplished by subjecting Julie to several months of seeing Alex spending night after night working on Jonah's ideas. Alex has his mother come over and stay with his kids for the time being.
Alex returns to the plant only to have his observations confirmed by a production capacity test. The team focuses on production capacity, and decides it's time to talk with Jonah again. This time, Jonah introduces the constructs of bottlenecks and non-bottlenecks.
There are nine rules expressing the relationship between bottlenecks and non-bottlenecks. One specifically addressed is to balance flow, not capacity in relation to demand. If bottleneck capacity is kept equal to demand, and demand drops, costs will go up resulting in a loss of money. The objective is to maintain capacity at slightly less than demand.
Alex enlists the help of Ralph Nakamura, the data processing manager, to identify the plant's bottlenecks. A few days later, and probably several thousand pages of output analysis later, the search for Herbie, the bottleneck, is still on-going. Obviously they need a simpler approach. Alex remembers the analogy of the hike. The bottlenecks should be identified by the backlog of work-in-process sitting in front of them. The two obvious bottlenecks turn out to be the multi-process automation machine and a heat-treating furnace. But the robotic machine was supposed to have a major production efficiency increase over the original manual processes.
Specifically , the old machines' process times were: Machine1 - 2 minutes, Machine2 - 8 minutes, and Machine 3 - 4 minutes, for a total cycle time of 16 minutes. These machines required one machinist per station for a total of 10 machinists. The new NCX-10 robot can process the same item in 10 minutes using 2 operators. Less time plus fewer operators should have lower costs and higher efficiency.
But there is a six month lead time to train a NCX-10 operator because of the specialty position requirements. And trained operators are leaving the company faster than it can re-train replacements, so the machine isn't running at full capacity -- a non fully-utilized bottleneck. The furnace is being run at partial loads because of expediting -- another non-fully utilized bottleneck. Management will not accept a request for adding even more capacity for a plant that isn't making money in the first place. Another visit to Jonah is warranted.
There is a realism that one cannot turn a bottleneck into a non-bottleneck. One can increase capacity of a bottleneck to be closer to demand. During a plant tour, Jonah notes that the composition of much of the work-in-process waiting at each of the bottlenecks is actually non-saleable parts destined for warehouse storage. This is hidden excess capacity. He asks about alternative methods which could be used in addition to the present processes, addressing the old retired machines as a potential capacity source. He asks if every part actually needs to be processed by the bottleneck -- more hidden capacity.
Alex learns to consider using alternate processes or off-load to increase capacity. Quality controls should be placed prior to a bottleneck to ensure the bottleneck will not be processing defective parts and wasting valuable bottleneck process time. Rejecting materials prior to the bottleneck then becomes simple scrap rather than consuming additional capacity. Process controls at a bottleneck should be designed to ensure zero defect based processing to minimize re-work and system impact costs. Total plant capacity equals the bottleneck capacity. Bottlenecks should be optimized by eliminating time wasted through idle bottleneck time, processing defective parts, or producing parts which do not contribute to throughput.
Finally, Jonah asks if they know what costs are tied to these bottlenecks. Work-in-process bottleneck costs are not appropriately measurable by using standard material and labor component cost elements. True process costs incorporate the market value of the finished goods waiting for the part to be completed at the bottleneck. Bottleneck per unit cost thus equals the total plant operating cost divided by the total bottleneck production hours because the bottleneck defines the plant throughput.
The team determines that one of the underlying causes of their present parts pile-up at the bottlenecks is because the operator cannot tell the difference between a bottleneck-destined part and an ordinary one. The operator, in an attempt to keep busy, processes batch after batch of non-bottleneck parts when what they really need to do is work on bottleneck parts -- the parts holding up the rest of the process. They attempt a solution for this by placing identifying tags on the parts which are destined for a bottleneck process.
Bob manages to recall some of the old machines which used to process the NCX-10 parts and return them to service -- resulting in an increase in bottleneck capacity. The furnaces are not being manned by dedicated personnel to keep them operating and reloaded during the idle times, so additional personnel are assigned to them on a full-time basis. Both of these measures worry Bob since they both result in an increase in direct labor at the bottlenecks.
Some of the bottleneck foremen come up with methods of streamlining their processes to increase throughput at their stations. And for a time, things seem to be improving - inventories are slowly shrinking and more backlog orders are being filled. Then Stacey reveals a problem with shortages of non-bottleneck parts now occurring in addition to the bottleneck parts. Could a new bottleneck have emerged as a result of overtaxing the rest of the system? Another call to Jonah is in order.
After learning about the identifying tag solution and finding several backlogs of non-bottleneck parts, Jonah illustrates the four basic models of manufacturing, shown as follows, where X represents a bottleneck and Y represents a non-bottleneck:
As it turns out, activating a resource and utilizing a resource are non-synonymous. Because Stacey was continuing to feed non-bottleneck material into the system in order to maintain the production efficiency quotas, non-bottlenecks began turning out maximum units of non-bottleneck parts clogging the work-in-process inventories at bottlenecks and at non-bottleneck stations. Activating a resources is simply turning it on. Utilizing it means making use of the resource in a way that moves the system towards the goal.
Ralph and Stacey work together to develop a material release system which triggers release of bottleneck material only at the rate at which the bottlenecks need it, rather than being triggered by non-bottleneck idle time. Jonah shows that they can use the same methodology to develop a release system for materials throughout the system. By knowing when the bottleneck parts will reach final assembly, the release of the non-bottleneck materials can be timed to coincide along the other routes.
Alex reports his results to Peach, who is impressed, but not sufficiently to call off the division sale. Alex agrees to improving another fifteen percent improvement in the net profit in order to prove that the changes are not momentary or unique. Another call to Jonah is made.
As it turns out, Jonah indicates that after load balancing is performed to meet market demand without excess production, the next logical step is to reduce the batch sizes to reduce the total capital commitment used during production. Reduction in batch sizes also reduces the total time spent in work-in-process. Less time spent in production increases the speed of throughput as well as a faster turn-around on customer orders. Shorter lead times result in better response to the market demands.
The four primary time components include: setup time, process time, queue time (associated with bottlenecks where parts wait for a machine to become free), and wait time (associated with non-bottlenecks when a part waits for another part to continue processing). Time saved at a non-bottleneck is imaginary because when non-bottlenecks are being set up, the time spent is taken away from idle time, not production time. Economic batch quantities are calculated based upon the whole system and not the bottlenecks themselves. As a result, most batch sizes are not optimized to the stations most affected by them -- the bottlenecks.
Now that the plant has the potential of responding better to market demands, Alex focuses on the third component of productivity measurement, that of sales throughput, and gets the division sales manager, Johnny Jons, to market his plant's improved capacity. Together they manage to tie down a major contract using a combination of incremental deliveries and low quantity pricing. This sets the stage for accomplishing the fifteen percent improvement Alex promised to Peach.
Smyth, the division productivity manager and competing plant manager, sets out to identify what Alex is doing to his plant by initiating an internal audit. As a result of the labor changes and the non-bottleneck idle times the cost reports show an increase in per unit costs. Smyth calls Alex in to explain himself in light of the audit findings.
Alex illustrates specific points that are in direct contradiction with conventional manufacturing assumptions:
Alex is then placed in front of Peach with the division controller, Frost, and Jons, the division sales manager. Smyth presents his findings - that Alex's plant has decreased productivity, increased product cost, and improper adherence to procedures throughout the organization. Smyth is asked to address the contraindicating evidence that the plant has turned solid profits and lowered operating expenses, increasing cash flow. He states that it must be only a temporary phenomenon and that big losses can be predicted for the future. Jons points out that Alex's plant represents the ideal combination of delivery speed, low cost, and flexibility that the market really needs. With that, Alex is announced as Bill Peach's replacement as UniWare division manager.
It turns out that Lou has also been hard at work behind the scenes re-crunching the numbers and identifying an additional flaw in the conventional cost accounting process - evaluation of inventory costs. Even though using the cash method would clearly demonstrate the decrease in work in process and finished goods inventory, and decreases in purchased material costs, the traditional accrual method shows these actions as period losses since cash payment avoidance is not recorded until the next accounting period. In re-calculating the financial statements, Lou found a projected twenty percent bottom line improvement instead of the promised fifteen. But instead of giving these cumbersome accounting explanations to Alex to use in his defense, Lou took the results directly to Frost who understood the ramifications of the numbers. Thus Alex ended up having two vocal supporters in his meeting with Peach.
Alex announces his promotion to his team an solicits their interest in moving up the ladder with him. Lou will take a place with Alex as the new division controller; Bob will become plant manager, and Stacey will become production manager. They initially brainstorm about how Alex might reorganize the entire division to implement the improvements across the other plants, but in the process they identify that most attempts to organize result in arbitrary assignments without function. They determine that what is needed is a process or tool to reveal the intrinsic order of elements within an organization that could form a naturally occurring trellis.
Looking back at the process they went through in order to achieve their present state they identify the following five primary steps of the improvement process:
Stacey finds that the tag system originally introduced to solve the backlog problem has developed into a now superfluous tool causing abnormal fluctuations in the balanced production system. Removal of the tags should solve the situation.
Alex sits down with Jons to talk about further market expansion. Due to the improvements, the plant now has twenty percent additional capacity available to fulfill demand. Jons explains that Europe has many potential customers, but the prices they demand are so low below the domestic market, UniCo couldn't possibly take them without losing money. Alex points out that when production is used from spare capacity, the only costs are the cost of the materials and as such, any price above material cost represents profit. Combined with an unbelievably short delivery time to shut out unimproved competitors, Jons has a deal in-pocket to ensure the future sales of the plant.
With the introduction of the increased demand, Alex's new sale turns out to set the plant in a temporary tail-spin by quickly consuming all remaining capacity without providing for the required excess capacity to absorb the statistical fluctuations of production. Julie has taken up reading philosophy and points out the normal process of scientific deduction is based upon one key relationship "IF…THEN." Triggering breakthrough thinking is not enough. One must also address all of the resulting bad effects, without creating insurmountable new ones.
Alex and Lou identify three fundamental decision issues as critical to the success of any manager:
Alex resolves to learn the rest of the thinking processes to know that he's doing his job as a manager. Jonah has pointed out a path, which each person must travel alone, by thinking and learning how and why to think.
One should be able to hone in on a core problem even in a complex environment. One should be able to construct and check solutions that really solve all negative effects without creating new ones. And above all, one should be able to cause a major change smoothly, without creating resistance, but instead, enthusiasm.
Fortunately for the reader, these processes are identified in detail in Goldratt's following book, "It's Not Luck."