Today we read Jim Collins’ Good to Great (2001) alongside Phil Rosenzweig’s The Halo Effect (2007). Collins looks for the common principles that let good companies leap to greatness. Rosenzweig warns that the moment we take the traits of successful companies to be the causes of their success, both research and judgment can go astray.

The second book directly criticizes the kind of management research the first one represents. So the clash between them is not a matter of taste. It is a conversation about whether studying successful companies can give us principles to act on, or whether we are being fooled by the halo that success itself creates.

Good to Great Jim Collins

The Halo Effect Phil Rosenzweig

Jim Collins selected companies that had stayed ordinary for a long time and then leapt to outstanding performance, and compared them with companies in similar circumstances that did not make the leap. Rather than a single star CEO or a dramatic innovation, he focuses on what the organization did consistently, over and over. His central claim is that greatness can be built through deliberate choices and disciplined execution.

Level 5 leaders put the organization’s long-term results ahead of their personal reputation. Humility here is not avoiding decisions; it is refusing to hoard the credit. Professional will is the strength to face hard facts and carry out the choices the organization’s goals require.

Instead of fixing the whole strategy first and then placing people, Collins suggests getting the right people on board first. Even as markets and technology shift, a team you can trust is able to debate and execute a new direction. This does not mean strategy is unnecessary; it means building the organizational foundation that lets strategy change.

It is the search for the point where three things meet: what the organization is deeply passionate about, what it can be the best in the world at, and what drives its economic engine. Instead of doing everything it could do, an organization must choose what to focus on and what not to do. A simple direction should be the result of deeply understanding a complex reality, not a shallow slogan.

Collins stresses confronting reality without dressing it up while keeping faith in long-term goals. The flywheel is a metaphor for momentum built by the accumulation of consistent choices and execution, rather than a single innovation or reorganization. Technology, too, is not magic in itself but a tool that accelerates a direction already chosen.

Phil Rosenzweig examines a logical error that runs through management research and business journalism. When a company’s sales rise, its strategy, culture and leadership are all praised; when results fall, the very same traits are reinterpreted as arrogance and incompetence. He argues that observing what successful companies have in common and proving what caused their success are entirely different things.

A company with strong results seems to have a great culture and brilliant leaders. When results decline, the same culture can be read as complacency and the same leader’s confidence as stubbornness. Assessments gathered after the outcome is already known may not be data about causes that are independent of that outcome.

Just because an autonomous culture and high performance appear together, we cannot conclude that autonomy caused the performance. High performance may have created the slack to grant autonomy, or a third market factor may have produced both at once. Having many examples does not by itself establish causation.

Great strategy, strong leadership and good culture are presented as separate drivers of success, but in real assessments they are tightly intertwined. Counting impressions that come from the same halo as several independent causes exaggerates their explanatory power. What matters is not the volume of data or the sophistication of the analysis but whether the measures are independent and valid.

A company’s performance depends on competitors, customer demand, technological change, the economy and timing. The same strategy can produce different results in a different environment. What Rosenzweig emphasizes is not a secret formula that guarantees success, but strategic choices and execution made with an understanding of uncertainty.

The first clash is about method. Collins extracts management principles by analyzing the differences between companies that made the leap and their comparison companies. Rosenzweig warns that assessments of culture and leadership collected after performance is known can be contaminated by the results. Using a comparison group helps, but it does not automatically remove bias from the measurement itself.

The second clash is about generalization. Collins’ ideas of leadership and the flywheel give a direction to act on. Rosenzweig asks whether every company with those traits succeeds, and whether companies that failed had the same traits too. The error can begin the moment a useful hypothesis is declared a universal law of success.

The third clash is about how much leaders can control. Collins emphasizes that the right choices and disciplined execution make the leap possible. Rosenzweig argues that underestimating competition, luck, timing and risk leads us to believe we control outcomes more than we do. The former shows the direction of action; the latter shows the limits of judgment.

Both authors are wary of blindly following fashionable management slogans. Collins says charisma or new technology alone does not make a great company. Rosenzweig criticizes the fantasy that copying the traits of successful companies will make you successful. Both value actual choices and execution over appearances.

The two books also meet on the need to face reality. Collins demands the discipline not to avoid brutal facts, and Rosenzweig demands that we look for independent data that could refute a success story. A good principle can serve as a working hypothesis, but its effect has to be tested continually.

If a technology company is entering a new market, Collins’ view says it should define its core customers and economic drivers and build a flywheel. From Rosenzweig’s view, however, one successful demo or one large contract does not prove a business model that works for every customer. Repeat purchases and real return on investment need to be checked separately.

In automation projects, too, PoC success rate, production stability, engineering hours per installation, maintenance cost and partners’ ability to implement independently should be measured separately. Stretching one success story into a claim about the maturity of the whole platform is falling into the halo effect. What matters is identifying the conditions under which the economic value customers get is repeated.

In evaluating leadership, it helps to separate outcomes from process. If every decision made in a strong-sales quarter is judged brilliant and every judgment in a missed quarter is judged wrong, learning becomes hard. Recording the information, alternatives and assumptions at the time of the decision, along with the actual outcome, reduces after-the-fact success stories.

An AI strategy should not start from copying famous companies’ adoption stories either. We need to confirm which of our customers’ problems get solved, and what improvements in accuracy, cost, quality and exception handling can be reproduced. The moment technology becomes a halo of success, the purpose of solving customer problems can blur.

In my work, I feel there is a big difference between a good principle and a proven one. An organization that has lost its way needs Collins’ Hedgehog Concept, but the more confident it becomes about its direction, the more it also needs Rosenzweig’s questions. We have to check whether the evidence that we are doing well was measured independently of our results.

One large contract does not prove everything about a product’s competitiveness, and one failed project does not invalidate every judgment a team has made. Revenue as an outcome matters, but whether the process that produced it can be repeated has to be checked separately. Numbers are a standard of accountability and also the starting point for deeper questions.

What a leader needs is neither the cynicism that throws principles away nor the blind faith that clings to a formula for success. We should choose a clear direction and execute it steadily, yet be able to revise our principles when contrary evidence appears. Good strategy gives us the power to execute, and good doubt keeps that power from running in the wrong direction.

Is your organization better at explaining, after a success, why it happened, or at testing the hypotheses it set before succeeding against actual results? It is also worth asking whether you have recently praised a strategy or a person simply because the numbers were good.