Not the values statement. Not the engagement score. This is the research CultureLogic rests on — why culture decides outcomes, why almost nobody reads it in time, and why the behavioural record is the thing worth reading.
What this page is for
This page is why it works. The research the method rests on, whose it is, and the limits of what can honestly be claimed from public evidence.
For what we do — the four steps in order, and which step each product buys — see Method. That page describes the procedure; this one argues for it.
“Culture eats strategy for breakfast.” Peter Drucker's formulation survives because it keeps being demonstrated: the best-constructed strategic plan is executed by an organization whose existing patterns decide what actually gets done with it. Leaders who treat culture as the soft variable are treating the expensive one as optional.
The reason it stays unread is that most culture work measures the wrong object. Surveys, values statements and engagement platforms collect what an organization says about itself — usually from the people most willing to answer. That is a measure of stated culture, and stated culture is exactly the part that is managed.
Culture is patterns of behavioural acquiescence to power — what people actually do under the local distribution of power, and what they quietly go along with. This position is argued hardest by Charles Dwyer at Wharton, and it refuses a premise most culture work depends on: that an organization is a living thing with values of its own. It isn't. An organization “cannot have objectives, goals, purposes, missions, visions, ideals, values or even problems” — those belong to people. It exists only in the minds of those attending to the concept of it.
Two consequences matter for anyone trying to read an organization from outside. First, power is not authority. Authority is assigned to you and shows on the org chart. Power is the ability to draw on organizational resources in service of your own values, and it accrues to whoever can link value satisfaction to specific behaviour in someone else's mind. The chart shows authority; the culture shows power. Second, an organization is always a means rather than an end — it attracts resources, transforms them into something of human value, and allocates what has been transformed.
Taken seriously, this is not a counsel of despair — it is the wager the whole practice rests on. If behaviour rather than public language is the signal, then behaviour's residue is the evidence. Press coverage, employee reviews, regulatory filings, leadership statements, hiring patterns, customer complaints: a great deal of that residue is already public, already dated, and was never negotiated with a communications team. That is what makes an organization legible from outside without anyone's permission.
The standard definitions of culture are not rejected here. They are subordinated to this spine: culture as a field, encoded in artifacts, many of them public, and therefore readable.
The four steps of the practice are on the Method page. This is the argument for why those steps are the right ones — and it is borrowed, deliberately, from a discipline that has spent a century learning how to read a subject that cannot be asked directly.
The interpretive core is adapted from depth psychology: the operations James Hillman sets out in Re-Visioning Psychology — personifying, pathologizing, psychologizing and dehumanizing — moved from clinical practice to organizational analysis. Naming the source matters more than borrowing its vocabulary, so the vocabulary stays here and does not appear in any reading we deliver.
Three of those operations do the work. Treating the organization as a character is what makes it a readable subject at all: without it you have data points, and with it you have something with a history, a temperament, and a way of behaving under pressure. Treating the symptom as a message inverts the usual consulting reflex — the attrition, the failed campaign, the deal that came apart are not malfunctions to be fixed but the culture communicating what it actually rewards and avoids. Seeing through the event to the pattern beneath is what lets one dynamic explain a retention problem, an integration risk and a partnership misalignment at the same time, rather than cataloguing three unrelated incidents.
The fourth operation is the discipline that keeps the other three honest: stepping back from the character so the pattern can be used. An analysis has to return something a room can act on and argue with, not a personality to believe in. This is also where the method deliberately parts company with its source — a clinical reading is for the subject, and this one is for a decision.
That adaptation is the load-bearing work, and it is documented rather than asserted: SJ King's dissertation research covers the bridge from clinical to organizational reading, and See What Others Miss sets it out for a general reader. The sources are public and listed below. How the work inside each step is conducted is proprietary.
The four moves require gathering and synthesizing a large body of organizational artifacts — press, filings, reviews, social signals, leadership statements. AI does that gathering at scale: what took a researcher six weeks now takes hours. The method hasn't changed; the throughput has.
What AI cannot do is the interpretive judgment at each move. Reading a symptom as a message requires reading meaning, not matching patterns. The AI gathers and surfaces; the methodology reads.
AI is a tool. It cannot take responsibility. AI scales the data work but cannot do the interpretive work that gives the analysis its depth.
It helps to think of culture as a field — a region of influence that shapes what the people inside it perceive, feel and do. The concept has serious lineage: a region in which a condition prevails in physics, the total configuration that determines the meaning of its parts in Gestalt psychology, and the dynamic system of forces shaping collective behaviour in Kurt Lewin's social psychology.
Three things follow, and each one is load-bearing for whether an organization can be read at all.
It is holistic. The meaning of any single element depends on its place in the larger pattern — as a word's sense depends on its sentence. This is why a pile of facts about an organization is not an analysis, and why individual data points routinely mislead.
It is generative, and self-reinforcing. A field directs attention and action toward some ends and away from others, largely below the level of deliberate choice. Over time those grooves deepen, because people internalise and reproduce the prevailing pattern through ordinary daily decisions. This is precisely why a culture is predictive: what an organization did under pressure last time is strong evidence for what it will do next time.
It is not monolithic. Most organizations carry subcultures by function, geography and level, which may align with the dominant pattern, diverge from it, or actively oppose it — and the whole thing keeps moving in response to leadership changes, acquisitions and outside shocks. An honest analysis names the dominant pattern and says where it does not hold.
Carl Jung described archetypes as the deepest patterns of psychic functioning — “the roots of the soul governing the perspectives we have of ourselves and the world.” They show up not only in mythic figures but in the recurring roles and storylines of organizational life. Jung also gave us the shadow: the unattended element that quietly undermines what a subject says it is trying to do. Both ideas do real work here.
But an analysis does not sort an organization into one of a fixed set of characters. There is no menu of twelve, no quadrant, no type to be assigned to. The pattern is named from that organization's own evidence, in language specific enough that it could not describe a different company — and it comes with what the pattern is protecting, and what it cannot see about itself.
That is the point of a name rather than a score. A score tells you where an organization sits on someone else's axis, and there is nothing to do with it but accept it. A named pattern makes a claim you can test against what you already know, and disagree with in one specific place. Every reading states its sources and its limits so that disagreement has something to land on.
What an analysis is, and the limits it states in every report, are set out on the Method page.