The Culture Code by Daniel Coyle: What the Research Says

What the book is

The Culture Code: The Secrets of Highly Successful Groups is a book by Daniel Coyle, the author of The Talent Code. Bantam, an imprint of Penguin Random House, published it on January 30, 2018; the publisher lists the hardcover at 304 pages (Coyle, 2018). The publisher's page calls it a New York Times bestseller and a Bloomberg and Library Journal best book of the year. Those are the publisher's descriptions, which I have not checked independently.

One line to avoid confusion: this is not The Culture Code by Clotaire Rapaille (Broadway Books, 2006), a book about national cultures and buying behavior (Rapaille, 2006).

In the introduction, which the publisher posts as an excerpt, Coyle says he spent four years visiting and researching eight of the world's most successful groups. He lists a special-operations military unit, an inner-city school, a professional basketball team, a movie studio, a comedy troupe and a gang of jewel thieves among them (Coyle, 2018).

The advice, in plain words

The book's starting claim, in the publisher's words, is that "Culture is not something you are—it's something you do" (Coyle, 2018). Culture, on this view, is not a trait a group is born with. It is made by what people do with each other, day after day.

The introduction names three skills that structure the book (Coyle, 2018):

1. Build safety. Small signals of connection tell people they belong and that the group is theirs.
2. Share vulnerability. Habits of mutual risk, such as admitting what you don't know, build trust that lets people cooperate.
3. Establish purpose. Stories the group tells about itself create shared goals and values.

Coyle says the three work from the bottom up: first the group connects, then it turns that connection into action. Each part ends with concrete suggestions. The opening example is an exercise in which teams build a tower from spaghetti and a marshmallow. Coyle reports that kindergartners beat business school students, because the students spent their effort working out who was in charge instead of trying things together.

Put plainly, the advice is this. Make it obvious to each person that they belong. Show your own mistakes first so others can show theirs. Say clearly, and often, what the group is for.

Why it became popular

I could not find a reliable source on how widely the book is used in companies or leadership programs, so I leave that question alone. What I can source is the reception. Beyond the publisher's bestseller labels, Kirkus Reviews called it a useful primer on building better organizations and pop science mixed with a business pep talk, summing it up as "Nothing world-shaking" but worth carrying on a trip (Kirkus Reviews, n.d.).

In my view, three features help explain the reach. It is built on stories about famous groups, which are easy to retell. It gives three short labels a manager can remember. And it carries a hopeful message: if culture is something you do, then any group can change its culture by changing its habits.

The main criticisms

Case selection. The book's method is to study groups already known as highly successful and to work out what they share. That design has a known weakness. Denrell (2003) argued, according to the abstract, that the organizations we can observe are survivors, and that books and the business press tend to focus on successful ones, so failure is undersampled. This can produce false beliefs about management: a risky practice can look linked to success in a sample of survivors even when it is unrelated to performance across all organizations. The question the book cannot answer from its cases is how many groups had the same habits and still failed. To be fair, the publisher says the book also includes stories of failure that show what not to do (Coyle, 2018), but those are illustrations, not a comparison group.

Anecdote doing the work of evidence. The publisher describes the book as combining leading-edge science with insights from leaders (Coyle, 2018). Research is cited, but the argument is carried by stories. The introduction, for example, cites a figure about strong culture and net income from what it calls a Harvard study of more than two hundred companies (Coyle, 2018). I did not trace that study, and a reader should treat the figure as the book's claim rather than a settled number.

My view. Several of the groups are elite and highly selective, and all of them made the book because they were already admired. Most teams are not like that. The book is strongest where its advice matches what research on ordinary teams finds, and weakest where it asks a reader to copy the habits of a championship team.

What the research says

Coyle's first skill, build safety, overlaps heavily with what researchers call psychological safety. In the Unfrozen Research Index (24,463 abstracts of research articles and reviews supplied by OpenAlex, covering 28 of the 29 journals (21 core, 7 expansion), 2007 to 2026, 2026 to date), 100 titles or abstracts mention psychological safety or being psychologically safe. These are keyword counts of titles and abstracts, not of full texts, so they understate how often the idea is used. Seventy-three of the 100 date from 2017 onward. Four studies bear directly on the book.

Safety and learning in teams. Edmondson (1999) defined team psychological safety as a shared belief that the team is safe for interpersonal risk taking. According to the abstract, in a study of 51 work teams in a manufacturing company, psychological safety was associated with learning behavior, and learning behavior mediated the link between safety and team performance. The same abstract says team structures, including support from the context and coaching by the team leader, shape these outcomes alongside shared beliefs. That fits Coyle's first two skills: safety pays off through what people then do, such as asking questions and admitting errors.

The meta-analysis. Frazier, Fainshmidt, Klinger, Pezeshkan and Vracheva (2017) pooled 136 independent samples covering more than 22,000 individuals and nearly 5,000 groups. Relationships with leaders and support from peers were among the antecedents linked to psychological safety, and the authors wrote that their results highlight "the salience of the direct leader in shaping the work context." Among individual-level outcomes, learning behavior had the largest estimated link. The same paper is candid about limits. The authors concluded that common method bias is a major concern in this literature, since only 13% of the correlations they used came from different sources. They reported that the link with task performance was much larger when performance was rated subjectively (an estimated .40) than when it was measured objectively (.07), and they noted that the literature lacks longitudinal designs. So safety is consistently linked to how people say they work together; the case that it raises measured results is thinner.

Leaders who invite contributions. Nembhard and Edmondson (2006) surveyed staff in 23 neonatal intensive care units. According to the abstract, leader inclusiveness, meaning words and deeds that invite and appreciate others' contributions, predicted psychological safety and moderated the link between professional status and psychological safety. Psychological safety in turn linked leader inclusiveness to engagement in improvement work. This is close to Coyle's second skill, with one difference: the study locates the behavior in the unit's leaders, not in a group's culture in general.

Climate varies team by team. Zohar and Luria (2005) studied safety climate, meaning shared perceptions of how much physical safety is valued, not psychological safety, so the parallel is partial. According to the abstract, organization-level and group-level climates were broadly aligned, and the effect of the organization's climate on safety behavior ran entirely through the group's climate. They also found meaningful variation between groups inside a single organization, which they attributed to supervisors' discretion in applying formal procedures when safety competed with productivity.

The manager closest to the team

Put those studies next to the book and a pattern appears. Coyle writes about whole groups and their famous leaders. The research keeps pointing at a smaller unit: the team and the person who runs it day to day. In the abstracts above, coaching by the team leader is among the structures that shape a team's outcomes, unit leaders' inclusiveness predicts whether people feel safe, and supervisors' choices explain why groups in the same organization end up with different climates.

My view is that this is the most useful way to read The Culture Code. Culture is not only set in a values statement at the top. It is also set in how a middle manager responds to bad news on a Tuesday afternoon, and that can differ from one team to the next under the same leadership. Leaders running a change lose something concrete when those managers get this wrong: the candid information the people closest to the work already have.

A balanced verdict

What the book is good for: giving managers memorable, behavior-level language. Its first two skills line up with a substantial research literature on psychological safety, and its central claim, that culture is made by repeated behavior, gives a manager something to do on Monday instead of something to wait for.

Where it misleads: by studying only celebrated groups, it cannot show that its habits separate winners from losers, and its stories can make the evidence look stronger than it is. The meta-analysis behind the safety research itself warns that most of the correlations it pooled came from a single source and that links with objectively measured performance are weak.

The best use I can suggest is as a checklist for the team you actually run, not a template copied from an elite group. Pick one behavior from each skill, try it for a month, and ask your team whether anything changed. For other books aimed at the people doing that work, see leadership books for managers, and for how teams react when change reaches them, see resistance to change.