Momentum Beats the Big Bet

The Core Insight

Eleven companies survived a screen that started with 1,435. Jim Collins and a team of 21 spent five years finding them. In 1996 a McKinsey managing director called Built to Last useless for most companies, because its companies had always been great. The question that fell out: can a good company become a great one, and how?

The screen was mechanical. A company had to return at least 3 times the market over the fifteen years after its transition. Over the fifteen years before, it had to return no better than 1.25 times the market. The shift also had to hold against its own industry.

Every company on the Fortune rankings of 1965, 1975, 1985 and 1995 went in, 1,435 of them. Return screens cut that to 126, cumulative-return patterns to 19, the industry test to 11. The eleven averaged returns 6.9 times the general market over the fifteen years after transition.

One dollar spread across them from 1965 and withdrawn on January 1, 2000 multiplied 471 times. The same dollar in the market multiplied 56 times.

Most managers explain a leap with an event: a hired outsider, a merger, a new technology, a named program. Collins argues the leap is an accumulation with no ignition point. Asked what they called their transformation, the executives had no name for it.

The Framework

The argument runs in three stages.

• Disciplined people come first: Level 5 leaders, then the right people on board before the destination is chosen.
• Disciplined thought follows: confront the brutal facts, then find the one concept where three circles meet.
• Disciplined action closes it with a culture of discipline and technology used only as an accelerator.
• The flywheel is what the three stages produce, turn after turn, until momentum breaks through.

Picture a metal disk on an axle, about 30 feet across, 2 feet thick, weighing about 5,000 pounds. Two or three hours of pushing produce one turn. Then two, four, eight, a thousand. Buildup took Circuit City nine years, Nucor ten, Gillette five, Fannie Mae three, Pitney Bowes about two. Circuit City drew no major articles in the decade before its transition and 97 in the decade after.

The comparison behavior is the doom loop. Skip the buildup, jump at breakthrough, get disappointing results, then lurch to a new direction. Warner-Lambert reversed direction four times between 1979 and 1998, cut 20,000 people, took a 550 million dollar write-off, and was absorbed by Pfizer.

The comparison design carries the claim. Eleven direct comparisons matched each company on industry, opportunity and resources at transition, and none of them leapt. Six more leapt and lost it. A variable earned a chapter only if it appeared in 100 percent of the eleven and under 30 percent of the comparisons.

Some findings are absences. Mergers ignite nothing, technology ignites nothing, and celebrity outside chief executives correlate negatively with the leap.

Key Ideas

Humility and Will Sit in the Same Person

Level 5 sits at the top of a five-level hierarchy of executive capability. Collins defines it as extreme personal humility fused with intense professional will. Every one of the eleven had it at transition, and the comparisons lacked it.

The diagnostic is the window and the mirror. These leaders look out the window to hand out credit when results are good, and in the mirror when results are bad. The comparison leaders ran it backwards.

Darwin Smith ran Kimberly-Clark for twenty years from 1971 and beat the market 4.1 to 1. David Maxwell declined the remaining 5.5 million dollars of a 20 million dollar retirement package. Al Dunlap took 100 million dollars for 603 days at Scott Paper, cut research in half, and sold the company. Ten of the eleven chief executives came from inside.

People Decisions Come Before the Destination

The eleven got the right people on the bus, the wrong people off, and the right people in the right seats. Only then did they decide where to drive. Collins amends the old adage: the right people are the asset.

David Maxwell interviewed every officer at Fannie Mae while the company lost 1 million dollars a day with 56 billion dollars of loans underwater. Fourteen of twenty-six executives left. The comparison pattern is a company built as a platform for one extraordinary individual.

Pay explains none of it. The 112 compensation analyses found no systematic difference on stock, salary, bonus or long-term pay. Six of the eleven recorded zero layoffs from ten years before breakthrough through 1998, and the comparisons used layoffs five times more often. Packard's Law sets the ceiling: no company grows revenue consistently faster than its ability to get enough of the right people.

Faith and Brutal Facts Have to Coexist

The Stockdale Paradox asks for both at once. Retain absolute faith that you will prevail in the end, whatever the difficulties. At the same time, confront the most brutal facts of your current reality. Jim Stockdale was the highest-ranking American officer in the Hanoi Hilton, tortured over twenty times across eight years from 1965 to 1973. Asked who did not make it out, he named the optimists.

Kroger decided in 1973 to eliminate, change or replace every single store and exit every region that did not fit. A&P opened one experimental store, disliked the answers, and closed it. Over twenty-five years Kroger returned ten times the market and eighty times A&P.

Four practices build a climate where the truth gets heard. Lead with questions and conduct autopsies without blame. Build red flag mechanisms that turn information into information nobody can ignore. Collins found no evidence that the eleven held better information than the comparisons.

One Simple Concept Beats Broad Competence

The Hedgehog Concept is a crystalline understanding of where three circles intersect. One circle holds what you can be the best in the world at. Another holds the economic engine, stated as one denominator, profit per x. The last holds what you are passionate about.

It is an understanding, not a goal or a plan. The eleven took about four years on average to reach theirs. The mechanism was a standing Council of five to twelve of the right people, with the final call left to the leading executive.

Walgreens moved from profit per store to profit per customer visit. Wells Fargo used profit per employee, Nucor profit per ton of finished steel. The shift put nine Walgreens stores inside a one-mile radius in downtown San Francisco.

Eckerd had the same 1.7 billion dollars in revenue and bought a home video business instead. That business lost 31 million dollars and sold to Tandy 72 million dollars below book value.

Five of the eleven sat in bad to terrible industries, and Pitney Bowes and Nucor sat in bottom-five-percent ones. Both beat the market by well over five times. Over two thirds of the comparisons chased growth with no concept underneath it.

Discipline Removes the Need for Bureaucracy

The order is disciplined people, then disciplined thought, then disciplined action. With disciplined people, hierarchy stops being necessary. Disciplined thought removes bureaucracy, and disciplined action removes heavy controls.

Budgeting gets redefined as elimination: which activities support the concept and get funded fully, and which get cut to zero. Stop doing lists outrank to do lists. Darwin Smith unplugged titles, annual earnings forecasts and layers of management.

Nucor grew into a 3.5 billion dollar company with four layers of management and a headquarters staff under twenty-five people. In the 1982 recession, worker pay fell 25 percent, officer pay 60 percent and the chief executive's pay 75 percent. One dollar in Nucor beat one dollar in Bethlehem Steel by over 200 times.

Every unsustained comparison rose under a personally disciplining Level 4 leader and fell when he left. Technology accelerates momentum and creates none. Of the eighty-four executives interviewed, 80 percent did not name technology among their top five factors.

Your Own Flywheel Fits in Six Components

Turning the Flywheel converts chapter 8 into a method. Collins wrote it after teaching the flywheel effect to Jeff Bezos and his team in autumn 2001. Lower prices bring more customer visits and more sales volume, which spreads fixed costs, which raises efficiency, which funds lower prices again.

The method runs in seven steps. List the replicable successes, then the failures beside them. Sketch the loop from what the two lists share. Test it against both lists and against the three circles.

More than six components means the loop is too complicated, and the fix is to consolidate. Each component has to be close to an inevitable consequence of the one before it. That test separates a flywheel from a list of objectives drawn as a circle.

Vanguard is the example. Lower-cost funds produce superior returns, which produce client loyalty, which grows assets under management. Scale follows, and scale lowers costs again. Vanguard more than doubled assets under management from 2009 to 2017, past 4 trillion dollars.

Score each component from 1 to 10. A loop scored 9, 10, 8, 3, 9, 10 stalls at the 3. Bring the 3 to at least an 8 before adding anything to the loop.

Practical Applications

Write the three circles before anything else. Name what you can be the best in the world at, and name what you cannot. State the economic engine as one denominator, profit per x, and change the x until one version sorts your decisions.

Draft the flywheel from evidence. List the wins that repeated and the initiatives that beat plan. List the failures beside them. The components sit in the gap between the two lists, and the loop is wrong until it explains both.

Hold the loop to six components and put them in causal order. Then score each from 1 to 10 and fix the lowest one first. A single 3 caps the whole loop, and a seventh component does not raise it.

Run the budget as elimination: fund what supports the concept fully and cut the rest to zero. When in doubt about a hire, keep looking, and check the seat before deciding the person is wrong.

Who This Is For

The books pay off for operators running a business that already works. The study excluded start-ups by design: a company needed twenty-five years of operations and ten years of stock data at transition. Founders before product market fit get more from the 2019 method than from the eleven case studies.

Skip the case histories if you want a causal model of corporate performance. The research design has a hole in it. Collins picked the eleven by past stock performance, then searched for the traits they shared. That method cannot separate cause from coincidence. Any eleven companies chosen for a fifteen-year run share traits, and nothing in the design says which traits produced the run.

Collins answers with the comparison set and the admission bar. An applied mathematician on the team put the odds of the pattern arising by chance at less than 1 in 17 million. That defends the pattern against luck. It does not establish cause.

The record is in the books themselves. Circuit City posted the best figure of the eleven, 18.5 to 1, and died in the winter of 2008. Collins names the deeper mistake: spinning CarMax out as a separate company instead of treating it as an extension of the same wheel.

Gillette stumbled in 1999, and Collins flags it inside the book. Nucor peaked at fourteen times the market in 1994 and then fell off during management turmoil after Ken Iverson retired. Collins wrote that the future of Nucor as a great company remained uncertain. The other seven are not described as declining anywhere in either text.

What survives the criticism is the flywheel as a design tool. Drawing your own loop, naming its components in causal order, and scoring each one is useful whether or not eleven companies proved anything. If component four does not cause component five, the loop is wrong.

The Decision

Block two hours this week and draw the wheel. Put the wins that repeated in one column and the disappointments in the other. Name the four to six components that explain both columns, and write them in the order where each one causes the next.

Score each component from 1 to 10 and circle the lowest. That component is the plan for the next quarter, and the rest of the loop waits.

A wheel nobody can draw is a wheel nobody is pushing.