The Goal by Eliyahu M. Goldratt

The Goal is an attempt to reevaluate business management from first principles. The book explains Goldratt’s Theory of Constraints, which states that in any business, the sole determinant of output is the bottleneck resource. To grow, a business has to identify its bottlenecks and reorganize business processes to address them. It sounds simple and perhaps obvious, but the lessons helped me to think about my own business.

The book teaches Goldratt’s ideas in novel format. It makes the book a light, engaging read, but it also feels like a lot of fluff to teach a few simple concepts. I admittedly felt silly that the author was doing the equivalent of hiding medicine in my doggie treats because I can’t pay attention to a real non-fiction book. That said, the last chapter is pure lessons and no story, and I found it difficult to pay attention, so maybe I needed the fluffy story.

The protagonist of the story is a struggling factory manager who’s mostly helpless without the guidance of a mysterious and near-omniscient Israeli physicist-turned-business-consultant. By sheer coincidence, the book’s author is an Israeli physicist-turned-business-consultant.

What I Liked ๐Ÿ”—︎

  • I found Goldratt’s distinction between science and truth to be profound and original.
  • It’s a light read, as the novel format makes it easy to stay engaged.
  • The Theory of Constraints helped me think about priorities in my business.
  • Goldratt’s metrics for judging a business’s effectiveness feel practical and well-reasoned.

What I Disliked ๐Ÿ”—︎

  • The novel format is a double-edged sword. Although it does make the book an easier read, some of the book is just for story, which dilutes the time you’re actually learning.
  • The protagonist’s marriage is a huge bummer.
    • He always prioritizes his job above his children and his marriage. The book makes his negligence out to be negative, but not very.
    • He does so little parenting or housework that his wife eventually leaves him.
    • I was expecting the protagonist to realize that he and his wife should discuss how to share responsibilities rather than him constantly breaking commitments to his family for his job. Instead, after his wife leaves, his mother moves in and just takes over all the childcare and household chores.
  • The protagonist meets an overweight child who struggles on a hike. Throughout the book, the protagonist casually refers to the child as “the fat kid” and uses him as the canonical example for a part of the system that holds everyone else back.

Key Takeaways ๐Ÿ”—︎

Science is not truth ๐Ÿ”—︎

  • Science is the process by which we create a minimum set of assumptions necessary to explain a phenomenon.

The Law of Conservation of Energy of physics is not truth. It is just an assumption that is valid in explaining a tremendous amount of natural phenomena. Such an assumption can never be proven since even an infinite number of phenomena that can be explained by it does not prove its universal application. On the other hand, it can be disproved by just a single phenomenon that cannot be explained by the assumption. This disproving does not detract from the validity of the assumption. It just highlights the need or even the existence of another assumption that is more valid. This is the case with the assumption of the conservation of energy which was replaced by Einstein’s more global โ€” more valid โ€” postulation of the conservation of energy and mass. Einstein’s assumption is not true to the same extent that the previous one was not “true.”

  • We’ve done too much gatekeeping around the “science” label.
    • We think of science as a way to explain physics, math, and biology, but we should accept the study of businesses and organizations as real science as well.

Basic characters and plot ๐Ÿ”—︎

The Goal explains non-fiction ideas through a fictional story, so it’s hard to explain without giving some basics on the story itself.


  • Alex is the manager of a ~100-person factory that manufactures some never-quite-defined product.
  • Jonah is an Israeli physicist-turned-business consultant. Alex studied under him in college and then runs into him many years later at the start of this story.

Basic premise:

  • Alex’s branch is underperforming, and company executives are considering shutting down his factory.
  • Jonah claims that the entire reason his plant is struggling is that everyone at his company blindly accepts conventional wisdom about business practices rather than thinking critically.
  • Jonah begins advising Alex on how to start from first principles and use scientific methodologies for saving his factory.

Local efficiencies are irrelevant ๐Ÿ”—︎

  • When Alex first runs into Jonah, Alex brags to him that new robots in his factory improved efficiency by 30% and reduced costs.
    • Jonah asks if they shipped more products -> No.
    • Jonah asks if they fired anyone (reduced labor costs) -> No.
    • Jonah asks if they reduced inventory (reduced warehousing costs) -> No.
    • Jonah claims that the robots have therefore not improved efficiency, as they failed to increase sales, reduce costs, or increase throughput.
  • When Alex tries to explain his metrics, Jonah claims that all of Alex’s metrics are meaningless.
    • Because everyone in the company uncritically accepts these metrics, nobody notices that the metrics fail to achieve the company’s true goal.
  • Alex reviews his sales figures and inventory records and concludes that Jonah is right.
    • Robots had no impact on sales, and they increased inventory (decreased throughput).
    • Because the robots were so expensive, the company pressured the factory to run them at 100% utilization, even when there was no demand for the parts they produced.
      • The result was a huge increase in half-finished inventory.

The goal ๐Ÿ”—︎

  • The goal of any company is to make money.
  • Productive activities are those which move the company closer to its goal.
    • Any activity within a company must be measured in terms of whether it helps the company earn money.

Metrics ๐Ÿ”—︎

Three main metrics indicate whether a company is successful at making money.

  1. Net profit: How much the company earns in absolute dollar terms.
  2. Return on investment: Net profit must be judged in the context of initial investment.
    • A $10M profit is great for a $1M investment but a poor return for a $50B investment.
  3. Cash flow: Even with healthy profits and return on investment, a business can go bankrupt if its cash flow is poor.

Jonah suggests the following metrics, which express equivalent ideas but rely on numbers that are more practical to measure:

  1. Throughput: the rate at which the system generates money through sales.
  2. Inventory: all the money the system has invested in purchasing things that it intends to sell.
  3. Operational expense: all the money the system spends in order to turn inventory into throughput.
This part confused me, as the two sets of metrics seem like they measure different things.

The myth of the balanced plant ๐Ÿ”—︎

  • Every factory strives to be perfectly “balanced.”
    • Balanced means that the capacity for every resource matches the market demand perfectly.
    • At a balanced plant, every employee and machine is busy all the time, and there’s a customer ready to purchase the finished product as soon as it leaves the factory.
  • A balanced plant is unattainable in practice.
  • A plant with zero idle time is inefficient because it means people must be working on products faster than the rate the company can sell them.
    • They’re optimizing for local efficiencies rather than optimizing for what will most improve the company’s bottom line.
  • If you perfectly match capacity to demand, productivity will decrease.
    • This is due to two factors:
      1. Dependent events: Manufacturing a product involves multiple steps that depend on earlier processes.
      2. Statistical fluctuations: There’s always random variance in the time it takes for any particular step in a pipeline.
    • Example:
      • A product passes through five departments for a task that takes, on average, an hour in each department.
      • Due to statistical variance, some departments will finish their work in a little more than an hour, and some will need less than an hour.
      • If batches of products go through this pipeline, backups will form because delays in one stage impact all subsequent stages.
        • Finishing early has no benefit unless the next department is ready to process the incoming batch immediately.

Dependencies and fluctuations: hiking single-file ๐Ÿ”—︎

  • Alex leads a Boy Scout troop on a hike and observes dependent events and statistical events in the boys’ pace.
  • The path is narrow, so everyone has to walk single-file.
  • Even though most boys walk at a speed of roughly ~2 mph, the group’s collective pace is markedly slower.
  • One heavyset boy is slower than the rest and begins falling behind.
  • Alex realizes that the average pace doesn’t matter because the slowest team member sets the pace of the group.
  • The line of boys has dependent events and statistical fluctuations.
    • Statistical fluctuations: Boys don’t walk at constant speed but rather a mixture of faster and slower paces averaging about 2 mph.
    • Dependent events: You can’t move faster than the person ahead of you in line, so your speed is limited by theirs.
  • Solution: The troop discovers the slowest boy is carrying a ton of excess weight in his pack, so they divide it up. The boy can then move faster, which means the troop moves faster overall.

Two types of resources ๐Ÿ”—︎

  • Bottleneck resource: Any resource whose capacity is less than or equal to the demand placed on it.
  • Non-bottleneck resource: Any resource whose capacity is greater than demand placed on it.
  • Businesses should match their bottleneck resources’ capacity to the market demand for the product.

Finding bottlenecks ๐Ÿ”—︎

  • Alex tried to calculate bottlenecks in his plant by measuring the capacity for each resource and the operational time it needs to meet demand.
    • This failed because it required rigorous record-keeping, which the plant didn’t do.
  • Instead, Alex used heuristics to identify bottlenecks:
    • Ask foremen and expediters about resources with frequent issues.
    • Look for resources with long work queues ahead of them.
  • Alex discovers two bottlenecks in his plant.
    • One is an advanced machine that requires a specialist to operate it.
    • The other is the heat-treating station.
  • Alex can’t easily fix the bottlenecks.
    • He can’t move either bottleneck to another position in the production pipeline.
    • He can’t purchase more of either resource because it’s too expensive.
    • He can’t offload work to other resources because those machines are the only ones of their type.

Modeling the cost of a bottleneck resource ๐Ÿ”—︎

  • The entire plant is only as fast as its bottleneck resources.
  • A bottleneck resource sitting idle for one hour is equivalent to the entire plant shutting down for one hour.
  • Cost at Alex’s plant:
    • Plant has $1.6M per month in operating expenses.
    • Specialty machine (bottleneck) can run for 585 hours per month.
    • One hour of idle time for the specialty machine costs the plant $2,735 ($1.6M / 585)

Fixing the bottlenecks ๐Ÿ”—︎

  • The specialty machine is idle when workers take their union break.
    • Fix: Renegotiate union contract so that workers rotate on the bottleneck machine to eliminate downtime.
  • The heat treat machine often runs below capacity.
    • Fix: Prioritize planning so that they can load each heat treat batch to near capacity.
  • The heat treat machine has idle time while workers load parts in and out.
    • Fix: Invest in more efficient techniques for loading and unloading the heat treat machine.
  • Quality control checks happen at the end of production.
    • This means that bottleneck resources wasted time processing parts that were already defective.
    • Fix: Do quality checks upstream of bottleneck resources.
  • The plant uses bottleneck resources to produce products that don’t have pending orders.
    • Fix: Reserve bottleneck resources exclusively for products that have outstanding orders.
  • Specialty machine is too expensive to buy more than one.
    • Fix: They find a set of older, less efficient machines that accomplish the same result as the specialty machine but slower. These machines supplement the capacity of the bottleneck machine.
  • Bottleneck resources didn’t require constant attention, so workers would often leave the machines to do other things and come back after the machine finished its work. This left idle time because workers never returned at the exact moment the machine finished.
    • Fix: Assign workers to be on constant standby at bottleneck resources.
    • Even though it seems wasteful to have workers stand around for hours waiting for machines to finish their work, it’s more efficient than letting the machine sit idle.
  • 20% of the heat treating station’s work is due to a change to an efficiency improvement on a machine earlier in the pipeline (a non-bottleneck resource).
    • Fix: Undo the change to the upstream machine to shift load away from bottlenecked resources.

After Alex’s plant implements all these changes, the plant ships a record number of orders for the month. But the following month, orders start running late again, and huge queues form in front of bottleneck resources.

  • Jonah points out that the new backlog is due to Alex chasing wrongheaded opportunities for efficiency.
    • Even though the bottleneck resources determine total throughput for the plant, the plant still runs non-bottleneck machines at maximum capacity.
    • If a non-bottleneck resource is upstream of a bottleneck resource and you run the non-bottleneck at maximum capacity, by definition, it generates too much work for the bottleneck to process.
    • For products that don’t pass through a bottleneck resource, the market demand is the bottleneck.
      • There’s no reason to produce a product at a faster rate than the market consumes it.
    • Fix: Send products through the pipeline only at the processing rate of the bottleneck resources.
      • Even if machines and workers sit idle for hours at a time.

Generalized process of bottleneck-oriented improvement ๐Ÿ”—︎

  1. Identify the system’s constraints (bottlenecks).
  2. Decide how to exploit the bottlenecks (i.e., maximize utilization of the bottleneck resources).
  3. Subordinate everything else to the above decision.
  4. Elevate the system’s bottlenecks (make changes that reduce bottleneck load or increase capacity).
  5. If, in a previous step, you eliminated a bottleneck, go back to step 1.

Profit by selling products for a loss ๐Ÿ”—︎

  • Alex’s plant becomes so efficient that they have 20% spare capacity and can’t find orders to fill it.
  • Alex offers to sell products below their production cost and insists that this will increase profit.
  • Co-workers are skeptical, but Alex’s logic is that the plant incurs most of its cost regardless of whether they manufacture new products.
    • They still have to pay worker salaries and plant maintenance.
  • If they’re paying recurring costs regardless, they should sell products as long as the sale price is higher than the cost of materials.
    • e.g., if a product costs $300 in materials and $200+ in labor, they should sell for prices above $300 because the labor cost isn’t meaningful when there’s spare capacity.
    • Letting parts or finished products sit in storage also costs money.