Most people think they’re busy when they’re actually just spinning their wheels. You’ve seen it. The warehouse floor is humming, the developers are typing like mad, and the "To-Do" list is getting checked off with aggressive satisfaction. Yet, the bottom line doesn't move. Profits stay flat. Why? Because most managers are obsessed with local efficiencies instead of the one thing that actually matters.
In 1984, an Israeli physicist named Eliyahu M. Goldratt released a business novel—yes, a novel about manufacturing logistics—called The Goal. It introduced the world to The Goal: A Process of Ongoing Improvement. It sounded dry. It looked like a textbook. But it became a cult classic because it exposed a painful truth: being "efficient" at a task that isn't a bottleneck is a total waste of time. Honestly, it’s worse than a waste; it’s a distraction that hides your real problems.
The Theory of Constraints is Kinda Common Sense (That Everyone Ignores)
Goldratt’s central premise is the Theory of Constraints (TOC). Think of your business like a literal chain. If you want to make the chain stronger, there is only one place you can work to get a result: the weakest link.
Strengthening any other link is literally useless.
If your marketing brings in 1,000 leads a day, but your sales team can only call 50, hiring more marketers is insane. You're just piling up a backlog of angry potential customers. This is the heart of The Goal: A Process of Ongoing Improvement. It’s about finding that one "Herbie"—the slow hiker in the book who determines the pace of the entire troop—and making sure everything else serves that pace.
It sounds simple. It’s actually incredibly hard to do in a corporate culture that rewards "looking busy."
Stop Measuring the Wrong Things
Most companies live and die by "Standard Cost Accounting." They want to see every machine running 100% of the time. They want every employee "utilized."
Goldratt argued this is a path to ruin.
If you force a non-bottleneck machine to run at 100% capacity just to keep the "utilization" metrics looking good, you end up with massive piles of Work-In-Progress (WIP) inventory. That inventory is just cash sitting on your floor, gathering dust and getting damaged. In the context of The Goal: A Process of Ongoing Improvement, the focus shifts from local "optimizations" to three specific metrics that actually dictate success:
- Throughput: The rate at which the system generates money through sales. Not production—sales.
- Inventory: All the money the system has invested in purchasing things which it intends to sell. This includes intellectual property and equipment.
- Operating Expense: All the money the system spends in order to turn inventory into throughput.
You want throughput to go up while inventory and operating expenses go down. If a "productivity hack" doesn't do one of those three things, it’s not an improvement. It’s just noise.
What Herbie Teaches Us About Your Office
In the book, the protagonist, Alex Rogo, goes on a Scout hike with his son. He notices the line of hikers keeps stretching out. The fast kids are way ahead; the slow kid, Herbie, is trailing behind. Alex realizes that the "output" of the hike—the whole group arriving at the destination—is dictated solely by Herbie.
To fix it, he didn't tell the fast kids to run faster. He put Herbie at the front. He took the heavy gear out of Herbie’s pack and distributed it among the others.
Suddenly, the line stayed together. The group moved faster as a whole, even though the "fast" kids were technically moving slower than their maximum capacity. This is the The Goal: A Process of Ongoing Improvement in action. You have to subordinate the rest of the system to the bottleneck. If your software engineers are the bottleneck, your product managers shouldn't be throwing 50 new features at them every week. They should be asking: "How can I spend my time making sure the engineers never have to sit in a useless meeting again?"
The Five Focusing Steps
Goldratt didn't just point out the problem; he gave a five-step loop for fixing it. It’s a cycle. It never ends. Hence, the "ongoing" part of the title.
- Identify the system’s constraint. Find the Herbie. Is it a machine? A specific person? A legal approval process? A lack of market demand?
- Exploit the constraint. Don't buy a new machine yet. First, make sure the current constraint isn't idling during lunch breaks or working on "junk" projects.
- Subordinate everything else. This is the hardest part. It means telling non-bottlenecks to slow down or wait. It feels wrong to see people "doing nothing," but it's necessary.
- Elevate the constraint. Now you buy the second machine or hire the second specialist.
- Prevent inertia. Once the constraint is broken, it won't be the bottleneck anymore. It will move somewhere else. Don't let old rules apply to the new situation. Go back to Step 1.
Real-World Nuance: It's Not Just for Factories
While the book is set in a plant, these principles are everywhere. Amazon’s fulfillment centers are a massive exercise in TOC. They don't care if a robot is fast if the packing station is backed up.
In digital marketing, your bottleneck might be your landing page conversion rate. You can spend $100k on Facebook ads (Inventory/Operating Expense), but if the page doesn't convert, your Throughput is zero. In that scenario, the "Process of Ongoing Improvement" dictates that you stop the ads and fix the copy. Period.
However, there is a catch. Sometimes the constraint isn't a physical thing. It’s a "policy constraint." These are the "but we’ve always done it this way" rules. These are often the deadliest because they are invisible. Goldratt’s later work, like It's Not Luck, dived deeper into these logical traps.
Why This Still Matters in the AI Age
We’re currently obsessed with AI productivity. Everyone wants to use LLMs to write emails faster or generate code faster. But if your bottleneck is "Human Approval" or "Legal Compliance," generating 10x more content just creates a massive clog at the approval stage. You haven't improved the system. You've just stressed out your legal department and created a pile of digital WIP.
The Goal: A Process of Ongoing Improvement teaches us that technology is only useful if it addresses a specific constraint. If it doesn't, it’s just a shiny, expensive toy.
💡 You might also like: Pounds to Cedis Ghana: What Most People Get Wrong About Today's Rates
How to Actually Apply This Tomorrow
If you're feeling overwhelmed, you probably haven't identified your bottleneck. You’re trying to optimize everything at once, which is mathematically impossible. You need to pick a fight with one specific problem.
Start by mapping your workflow. Literally draw it on a napkin. Where does the work pile up? Where are people waiting for an answer? That's your Herbie.
- Audit your "utilization" mindset. Stop judging your team by how "busy" they look. Judge them by how much they are helping the bottleneck move faster.
- Protect the bottleneck. If your lead developer is the constraint, they shouldn't be answering basic tech support tickets. Every minute they spend on a non-constraint task is a minute of lost throughput for the entire company.
- Limit Work-In-Progress. Cut the number of active projects in half. Watch how fast things actually get finished when people aren't context-switching.
- Institutionalize the Five Steps. Make it a monthly meeting. "Where is the bottleneck this month?" If you can't answer that, you aren't managing; you're just reacting.
The goal isn't to have a perfect, frictionless business. That doesn't exist. The goal is a process—a relentless, slightly messy, and deeply honest cycle of finding the next weakest link and making it stronger. That’s how you actually win.
Actionable Next Steps
- Identify your Herbie: Look at your project management tool (Jira, Trello, etc.). Find the column with the most cards. That is your current constraint.
- Calculate your Throughput: Focus on one metric—money hitting the bank—and trace back exactly what prevents that number from doubling today.
- Run a "Subordination" experiment: For one week, tell your non-bottleneck departments to do nothing except support the bottleneck. See if your actual output increases. It usually does.