Rockefeller's Bottleneck Strategy: how he built the one gate every barrel had to pass through, where the bottlenecks are in today's AI rush, and a worksheet to find yours
During America's first oil rush, thousands of men drilled and went broke. John D. Rockefeller skipped the drilling and went into refining, the one step every barrel had to go through before anyone could use it. Refining was not a locked gate when he started. Anyone with a still could join. He turned it into one. By around 1880 his company refined about 90% of America's oil. This page breaks down exactly how the Bottleneck Strategy worked, the four moves behind it, where the same narrow points sit in today's AI rush, and a 30-minute worksheet to find the one in your own industry.
Overview
In 1859 Edwin Drake drilled the first commercial oil well in Titusville, Pennsylvania. Within months the valley filled with drillers. Oil topped $10 a barrel in January 1861. By June it was 50 cents. By the end of the year, about 10 cents.
Rockefeller was a young bookkeeper in Cleveland. He looked at the same rush and asked a different question: what does every barrel have to pass through before someone pays for it?
The answer was the refinery. Crude oil was useless to a family. Kerosene, refined from crude, lit their lamps. There were thousands of wells and only a handful of refineries.
So he went after refining. In 1865 he bought out his partners at auction for $72,500 and owned the biggest refinery in Cleveland. In 1870 he founded Standard Oil.
The rush everyone else joined
Before Drake, people collected oil that seeped into Oil Creek by soaking blankets on the water and wringing them out. They sold it as medicine.
Drake proved you could drill for it. He drove iron pipe into the ground so the hole would not collapse, and his driller hit oil at 69.5 feet on August 27, 1859. Drake never patented the method. He ended up nearly broke, living on a small state pension.
Everyone who followed made the same bet. Lease a small plot from a farmer, drill, and hope. If you hit, so did your neighbors, and the flood of oil crashed the price you were selling at. The drillers carried all the risk. The farmers who leased the land, and the people selling barrels, tools and transport, often did better than the drillers.
The bottleneck
Look at the oil business as a chain:
| Step | Who did it | How crowded |
|---|---|---|
| Drilling | Anyone with a lease and a rig | Thousands of drillers |
| Refining | Crude in, kerosene out | A small number of refineries |
| Shipping | Railroads, later pipelines | A few railroads |
| Selling | Shops and wagons | Countless stores |
Drilling was crowded and risky. Selling was crowded too. Refining sat in the middle with far fewer players, and every barrel had to pass through it. Whoever controlled refining could set terms for the drillers on one side and the sellers on the other.
But it was not a natural bottleneck. Refining was cheap to start. The coal-oil plants that existed before 1859 simply switched to crude, and anyone with a still could join. Cleveland had dozens of refiners, and by the early 1870s there were too many and profits were thin.
Why refining still beat drilling: when crude crashed from $10 to 10 cents, a driller was ruined, but a refiner's raw material just got cheaper. Families kept buying kerosene for their lamps either way. The driller bet on what was under one plot of land and on the price. The refiner bet on steady demand.
So Rockefeller did not find a bottleneck. He built one, step by step.
The 4 moves behind it
Move 1: Pick the step everything passes through
He did not ask "where is the money?" Everyone thought the money was in the ground. He asked "what can nobody skip?" Every barrel had to be refined. It had fewer players than drilling, and it got stronger, not weaker, when crude prices crashed.
Move 2: Be the cheapest at it
His plants sealed each kerosene can with 40 drops of solder. He asked if 39 would hold. It did. Small savings like that, across millions of cans, let him sell cheaper than anyone and still make money.
Move 3: Use your size to get better terms
Because Standard Oil shipped more oil than anyone, the railroads needed it. He used that to get secret discounts his rivals did not get. This is also the part that later got him in trouble. The legal version today is simple: the biggest customer gets volume pricing.
The result: in about six weeks in February and March of 1872, 22 of the 26 refineries in Cleveland sold to him. People called it the Cleveland Massacre.
Move 4: Own the roads around it
A crowded step only becomes a gate once you control the roads in and out of it. So he stacked them. Pipelines, his own barrels, storage tanks, delivery wagons. Every time a rival could have gone around him, he owned that road too. By around 1880 Standard Oil refined about 90% of America's oil.
Where it broke
Ida Tarbell grew up in the oil region. Her father was a small independent oil producer hurt by Standard Oil. As a journalist she spent years documenting exactly how the company had grown, and her series in McClure's Magazine, starting in 1902, turned the country against it.
In 1911 the Supreme Court ordered Standard Oil broken into 34 companies. Rockefeller kept his shares in all of them. Once each traded on its own, investors could see what the pieces were worth. By most estimates his fortune reached about $900 million by 1913, around 2% of the US economy.
The lesson cuts both ways. Owning the bottleneck made him the richest man in America. Squeezing everyone who passed through it made him the most hated. The bottleneck works best when people are glad to pass through it.
The bottlenecks in today's AI rush
Today thousands of people are building AI apps, the way thousands drilled in 1860. Map it the same way Rockefeller mapped oil.
| 1860s oil | Today's AI | Why |
|---|---|---|
| Drillers | AI app builders | Thousands of them, cheap to start, and the price of "an AI app" keeps falling |
| Refiners | Model makers (OpenAI, Anthropic, Google) and chipmakers (Nvidia designs, TSMC manufactures) | Few of them, and almost every app pays them on every use |
| Railroads and pipelines | Cloud providers (AWS, Microsoft Azure, Google Cloud) and app stores | They deliver the computing power and the apps |
| Families buying kerosene | People and businesses using AI | Steady demand that keeps growing |
The same split as 1861: when AI apps flood the market and prices crash, app builders get squeezed, but the model makers and chipmakers still get paid on every use. A crash in apps hurts the drillers and helps the refiners.
Most readers will not build a chip company. So here is where the narrow points sit right now, from giant to small, including the ones a single person can own.
1. Chips
Almost every big AI model is trained and run on specialized chips, and a small number of companies design and manufacture most of the advanced ones. That is why chipmakers became some of the most valuable companies in the world during the AI boom. The app builders are the drillers. The chipmakers are the refinery.
2. App stores and platforms
If your AI app lives on a phone, it usually goes through Apple's or Google's store, which take a cut of sales (Apple's standard rate is 30%, lower for small developers). The platform is the narrow point every app has to pass through to reach a customer.
3. Attention
AI companies have raised huge amounts of money, and they all need users. On October 20, 2026, Sam Parr of My First Million said he had talked to four creators making over $1 million a year in ad revenue, almost entirely from AI companies, each with fewer than 100,000 followers. The person who owns an audience of the right people is the narrow point between a funded AI company and its customers.
4. Trust in a niche
A dentist will not try ten AI tools. They will try the one their dental association, their favorite newsletter or the consultant they already pay recommends. Whoever owns that trust decides which tools a whole industry adopts.
5. Local and licensed work
AI can write the plan for a kitchen remodel. It cannot pull the permit or wire the house. Licensed trades, inspections and on-site work are narrow points AI makes more valuable, because every AI-generated plan still has to pass through a person with a license.
How one person can be a small refiner
Today, AI tools are the crude oil. There are thousands of them, they are cheap, and prices keep falling. Most business owners cannot use raw AI, the same way a family could not burn raw crude in a lamp. They want the finished result.
The small refiner turns raw AI into a finished result for one group of people. When AI tools get cheaper, the refiner's costs fall and customers still pay for the result. Examples:
- The setup person for one trade. Sets up AI booking and follow-up for dental clinics. The clinic buys more appointments, not software.
- The trusted filter for one niche. The newsletter a whole industry reads to learn which AI tools actually work. Every tool that wants that audience has to go through it.
- The done-for-you product. Raw AI and 3D software in, a finished property tour video out, delivered to real estate agents by email.
- The finisher. AI writes the first draft of a contract, a logo or a tax summary. A licensed or skilled person checks it and signs off. Every AI draft still has to pass through someone people trust.
The test: if AI got ten times cheaper tomorrow, would your business get better or worse? A small refiner gets better.
The numbers
| Figure | Number |
|---|---|
| Drake's well | August 27, 1859, Titusville, PA, oil at 69.5 feet |
| Oil price, 1861 | About $10 a barrel in January, 50 cents by June, 10 cents by year end |
| Refinery auction | $72,500, 1865 |
| Standard Oil founded | 1870 |
| Cleveland Massacre | 22 of 26 refineries in about six weeks, 1872 |
| Share of US refining | About 90% by around 1880 |
| Breakup | 1911, into 34 companies |
| Fortune by 1913 | About $900 million, around 2% of the US economy (estimates vary) |
Find your bottleneck in 30 minutes
1. Draw your industry as a chain (10 minutes)
Do this: write 4 to 6 boxes from left to right: where the raw material or work starts, every step it goes through, and the customer who pays at the end.
Example for local home cleaning: homeowner has a mess, finds a cleaner, books and pays, cleaner shows up, job gets checked, homeowner leaves a review.
2. Count the players in each box (5 minutes)
Do this: under each box, write roughly how many people or companies do that step in your market. Hundreds, dozens, a handful, one.
3. Circle the narrow box (5 minutes)
Do this: circle the box with the fewest players that nobody can skip. That is your bottleneck. In the cleaning example, it is often "finds a cleaner." There are hundreds of cleaners and only a few places homeowners look.
4. Ask the 5 questions (5 minutes)
- Does everyone in this chain have to pass through this box?
- Is it getting more crowded or less crowded?
- Could I own a small slice of it, one city, one niche, one kind of customer?
- What would make me the cheapest or easiest option at this step?
- Would people be glad to pass through me, or would they resent it?
Want 50 more breakdowns like this?
Real businesses with their revenue, tools, costs and how they grew, plus step-by-step guides to start your own. Free.
Get free accessMore breakdowns
Sources
- PBS American Experience, The Rockefellers timeline
- American Heritage, "Rockefeller and Pithole, Pennsylvania"
- James D. Hamilton, "Historical Oil Shocks" (oil price history)
- EBSCO Research Starters, Standard Oil Company is incorporated
- Columbia Law, "History of Standard Oil: The Cleveland Massacre"
- Wikipedia, Standard Oil; South Improvement Company; Edwin Drake
- Sam Parr, post on X, October 20, 2026
