How the platform reads an organization
and what it will not claim.
The ten signals, the rules a claim has to clear before it appears, the frameworks behind them, and the research the whole thing is read against.
An organization is a living human system — people, relationships and cycles moving at once. More of it is connected than anyone can see from a single seat.
You do not have to adopt it wholesale. Provide the evidence, see the read, and test whether it explains something you already sense but could not name.
by Brooke Erickson
Not what you came for? Open the Research Studio
- What it claims
- Read the conditions people work inside together, rather than one symptom at a time, and the same few causes keep appearing behind problems that looked unrelated.
- What it refuses
- It chooses not to rate a person — the conditions people work inside explain more than the people do. Reading the system is how we stop blaming the people in it. It also will not call an association a cause, or fill a gap it cannot see.
- Where it came from
- Fifteen years of advisory work, held against a published research corpus that is open to you before you give us anything. It builds on organizational intelligence, organizational development, systems thinking, organizational psychology and behavioral research that long predate it; what is original here is the synthesis — the ten-signal model, the evidence rules, and this application of them.
Everything below is the long form. Stop wherever it stops being useful.
Ten interconnected reads of a living organization.
Not a scorecard. Ten conditions that keep turning up behind the problems leaders bring us, read together rather than one at a time. The weight next to each one is how much it contributes to the system read.
How effectively knowledge, feedback, context, and truth move through the organization.
The strength of trust, psychological safety, credibility, and collaboration across the organization.
The collective engagement, motivation, and capacity for effort — versus depletion, disengagement, and burnout.
The degree to which strategy, priorities, decisions, and daily work reinforce one another.
How the organization metabolizes experience, feedback, and failure into changed behavior.
The organization's ability to sense change, adjust course, and reshape itself without breaking form.
What the organization actually is in practice — the shared purpose, values, and behavior the system rewards.
How power, decision rights, and voice are distributed — where authority sits versus where information sits.
The finite attention, focus, and energy the system can allocate — and the slack it has left for anything new.
Structural, procedural, or behavioral barriers that make work harder than it needs to be.
Four principles the Method is grounded in.
Each principle states how organizations work, the research standard that validates it, and how that standard shows up in the platform.
Organizations are not made of separate functions. They are made of people and the relationships between them — so one decision reshapes trust, capacity, learning, and priorities for everyone at once.
Every claim in the Method must be traceable to research on how one part of a system reshapes another.
In the platform: Evidence intake never scores a signal in isolation. It always reads what that signal is pulling on.
Culture, safety, and performance do not live in a department. They are produced by how people treat each other under pressure.
We privilege research on system behavior over research on individual traits or isolated interventions.
In the platform: The Read surfaces the pattern between signals, not the signals alone, so the whole becomes visible.
History, incentives, and identity are carried by the people who lived them, and they outweigh most change efforts. Naming that gravity is often the intervention.
We include longitudinal and failure-case evidence, not only success stories, so the invisible field of forces becomes legible.
In the platform: The Advisor names what is already pulling before proposing what to push against.
People join, grow, plateau and leave while the organization is also moving. Any single reading catches both mid-stride — a photograph, where repetition is the movie.
The corpus is never finished. It is re-read and updated as the field evolves and new evidence appears.
In the platform: The Loop is designed for repetition, not one-shot fixes, so intervention becomes iteration.
A read is a claim under uncertainty, so it is required to say how strong it is and what would change it.
Every explanation the platform produces sits on a rung of the causal ladder and carries a three-part confidence grammar. Rival readings are shown next to each other, and the one you favor becomes the standing read until an intervention displaces it.
- Association
Two signals move together. Useful for attention, never for action on its own.
- Mechanism
You can name how one signal produces the other, and the mechanism survives contact with the evidence you already have.
- Intervention tested
A change was made on purpose and the read moved. This is the only rung that earns the word 'because'.
- Evidence
How much real evidence stands behind the claim, and whether the signals that matter are readable, emerging, contested, or not yet observed.
- Model
How well the framework being applied actually explains what is observed here, not in general.
- Judgment
How much the leader still has to supply. The platform names this out loud rather than hiding it inside a number.
Corrections, retired explanations, and reviewed interventions are carried forward. The Advisor cannot re-propose a read the organization already retired.
Open MemoryIt will not rate a person, predict an individual's behavior, or state a cause it has not tested. Where a signal is not yet observed, it says so instead of filling the gap.
Open ReasoningHow the free read decides what it is allowed to say
Situations ls-2.0 · reading rules ls-score-1.1 · selection ls-signal-1.0. Every version shown on a result page matches the rules described here.
Choices, not self-ratings
Fourteen situations, each one a forced choice between options that are all defensible. Nobody is asked to rate themselves. What separates leaders is the order they put good options in, and the order is the only thing measured.
Adaptive follow-ups
Where the core leaves a genuine ambiguity between two patterns — and only where an item exists that would actually separate them — up to three follow-up situations are asked. They are fixed, written in advance, and scored the same way as the core.
Context and own words are never scored
Scope, time in the current role, current pressure, your own account of how you lead, what drops under load and the two open questions are all kept separate. Time in role means exactly that — it is not a measure of how long you have been leading. They affect wording and are shown back to you. They never create, strengthen or suppress a finding.
Relationships between observations, not a list of traits
Where two observations that each cleared the evidence gate relate to one another, the read names the relationship: two sources agreeing, two sources pulling apart, a pattern that holds until the situation changes, pressure turning a pattern up, or pressure changing it. Where the evidence conflicts, that is said plainly rather than resolved. Blind spots and compensating behaviour are deliberately not attempted here — they need a second observer or movement over time.
Research explains, it never manufactures
Your evidence creates the finding. Published research may then be shown, named and linked, to say why that relationship is worth noticing. It is always the last step and always optional to open. General statements about people are either directly supported by your own answers or carry a named source; the read makes no claims about how common anything is, here or anywhere else.
Told and shown are kept apart
Your account of how you lead is recorded before the situations, so the situations cannot prime it. Where it differs from what your choices showed, the read names the difference rather than averaging the two.
No traceable evidence, no finding
Every observation on a result page carries at least two specific answers you gave, shown under “What we saw”. A finding that cannot name the answers behind it is not shown at all — which is why some reads have two observations rather than three.
Nothing is padded to a number
There is no forced top three. Where your answers spread evenly, the read says so and declines to rank. Where an area did not get enough of your choices, it is named once, plainly, and nothing above rests on it.
Fixed rules, not a model
Every sentence in a result is written in advance by us and selected by rules. No language model writes, ranks or edits a finding. The same answers always produce the same read.
What changed in the current version
Reading rules ls-score-1.1 recalibrated how the same underlying evidence is read against the scale this instrument actually produces. Item weights, what each option counts toward, and the penalty for options passed over are unchanged. Selection ls-signal-1.0 replaced the old area-by-area report with the small number of observations the evidence can carry.
The deeper Leadership Intelligence work — richer self-report, personal and formative context, input from people around you, movement over time and organizational evidence read together — is deliberately not part of the free read, and none of it is inferred from these fourteen situations.
Every finding traces to a framework. Every framework traces to research.
Two groups: native frameworks Brooke authored inside the Method, and drawn-from frameworks the Method reads through — Edmondson, Meadows, Kahneman, Heifetz, Nonaka, Grant, Brown, Gladwell, Argyris, Schein, and the growing Method tradition among them.
The anchor stance every reading in the Method is grounded in.
The three governing principles — Interdependence, Emergence, Gravity — behind every framework here.
The frameworks that actively power the diagnostic, the dashboard, and the advisor.
How the ecosystem actually behaves across 15 dimensions, using five behavioral modalities designed to reduce social-desirability bias.
One honest reading of the whole system, weighted from signals — not performance metrics.
The gap between hours worked and strategic capacity actually available.
The reinforcing forces — history, incentives, identity, defensive routines — that quietly outweigh most change efforts.
Early signals — questions fading, meetings lengthening, decisions centralizing — that appear long before performance moves.
Ten interconnected indicators of ecosystem condition — an evidence-gated reading, not a balanced scorecard.
How much change the ecosystem can actually absorb without breaking.
How well functions, layers, and geographies actually talk to each other.
Lenses the Method reads results through. They shape how findings are explained — they don't score directly.
How dimensions combine — safety × workload × incentives — to produce outcomes no single variable explains.
The nested ecosystems — individual, team, function, enterprise — as one interacting whole.
The conditions around decisions — information, incentives, cadence, follow-through — not the decisions themselves.
Leadership as a system property — formal authority, informal influence, and network leadership all read as one flow.
How humans and AI share work, judgment, and accountability inside the ecosystem.
Where information moves — and where it gets trapped, distorted, or lost.
Trust as a network property — where it is thick, thin, one-directional, or broken.
The coined concept of reading organizations as living human systems — one of the early influences on the Method.
Being sharpened in the research lab. Visible to executives, honestly labeled as evolving.
The tacit knowledge, patterns, and decisions the enterprise is silently losing.
Whether the enterprise is generating energy or extracting it from future capacity.
The cognitive and emotional weight the ecosystem is asking humans to carry.
A continuous rhythm — lightweight signals plus targeted deeper reads — that keeps the ecosystem legible over time.
Which prior strengths — informal trust, founder judgment, ad-hoc coordination — quietly stop working as the ecosystem grows.
The library every read is held against.
31 domains, 96 works — from organizational psychology to complexity science to documented failures (Challenger, Boeing, Wells Fargo, Nokia). The corpus is curated, not scraped. The Advisor answers from it and is instructed to name what it draws on; where it cannot name a source, treat the answer as reasoning rather than citation.
Limits, plainly: a read describes conditions, not people. It cannot see what you did not give it, it does not predict what any individual will do, and one reading is a single point in time. Where a signal has no evidence behind it, the read leaves it unread rather than filling it in.
Browse domains, authors, and the exact works each framework and signal leans on.
Open ResearchTwenty-five items, about fifteen minutes. What the rest of the Method interprets against.
Open EvidenceA standing conversation grounded in your diagnostic, the framework library, and the research corpus.
Open the Advisor