Three ideas hold the map together
An ontology says what a company is made of. A taxonomy says how an industry is classified. Signal Theory says why the result is a network you can diagnose, not a diagram you can only admire.
Ontology
Five layers (people, process, software, data, decisions), the primitives inside each, the modules and interfaces under every primitive, and the components, values and edges that describe one real company on top of them. Every term is defined once and used everywhere, so two companies described by two consultants still read the same.
Read the glossaryTaxonomy
An industry schema groups its primitives into the parts of the business (winning work, delivering work, getting paid…) and names the standard functions inside each. Schemas are versioned and published, so a company mapped in 2026 stays comparable to one mapped later.
Browse the schemasSignal Theory
Every edge in the atlas is a signal: intent encoded by one part of the company, carried over a channel, decoded by another, producing action. Four governing principles (Shannon, Ashby, Beer, Wiener) say when that network is viable. The atlas views are those four questions, drawn.
See the methodA company is a network of signals. The atlas draws the network.
Every act inside a company follows the same architecture: intent is encoded into a signal, carried over a channel, decoded at a destination, and produces action. Signal Theory integrates Shannon, Wiener, Ashby and Beer into one framework around a single objective, maximising signal-to-noise, under four governing principles. Read onto a company, Shannon's model becomes the atlas's data model.
| Shannon's model | In a company | In the atlas |
|---|---|---|
| Information source | The team or role that forms the intent | A component on the people or decisions layer |
| Encoder | Creating the signal: a bid, a schedule, an invoice | A process component and the genre it emits |
| Channel | The medium it travels on: a tool, a meeting, a hand-off | An edge, with its flow type, or a software component |
| Noise | Re-keying, politics, overload, ambiguity | A manual hand-off, a missing interface, a low-confidence value |
| Decoder | The team that reads it and acts | The component at the other end of the edge |
| Destination | The action taken: work done, cash collected | A record on the data layer, closed by a feedback edge |
Four questions, four views
The diagnostic reduces to one question applied four times. Any “no” names both the failure and its fix. Each question has a view in the atlas that makes the answer visible.
Edges with no tool behind them, hand-offs re-keyed by hand, one role at the centre of too many edges.
- Failures
- Routing failure, bandwidth overload, fidelity failure
Primitives and modules a company is missing or runs manually while the industry's demands keep growing. The coverage states and the gap report.
- Failures
- Genre mismatch, variety failure, structure failure
Every part of the business is itself a viable system: operations, coordination, control, intelligence, policy. The hierarchy and layer-stack views show where a level is hollow, and where a signal skips a layer it should pass through.
- Failures
- Bridge failure, herniation, decay
Feedback edges: approvals, reconciliations, retrospectives, the returning-customer loop. A flow with no edge back is open-loop and drifts.
- Failures
- Feedback failure
- Views
- Swimlanes, flow
What each part of a company does in the network
Every organisational entity has two names: what it is, and what it does for signals. The atlas stores the first and draws the second.
- Company
- Network
- The company atlas: the bounded system every view renders
- Department
- Subnet
- A top-level primitive, a part of the business; a swimlane
- Team
- Cluster
- A people-layer component with children
- Person or role
- Endpoint
- A leaf component on the people layer
- Recurring process
- Standing channel
- A process component and its edges
- Project
- Signal chain
- An ordered path of edges through the map
- Partner, customer, regulator
- Bridge
- An external node outside the company boundary
- Knowledge base, records
- Signal memory
- A data-layer component
- Machine intelligence
- Signal processor
- A software component that encodes, routes or decodes
The five dimensions of a signal
Every interface in the module catalog records what it carries, who is on the other side and what triggers it. Those are the signal's dimensions, which is why a missing interface is a diagnosable failure rather than a blank box.
- Mode
- How is it perceived?Linguistic, visual, auditory, spatial
- Genre
- What conventional form does it take?SOP, proposal, invoice, change order, daily log
- Type
- What does it do?Direct, inform, commit, decide, express
- Format
- What is the container?PDF, spreadsheet, app record, meeting
- Structure
- How is it organised?Scope → responsibilities → steps → exceptions
Luna, R. H. (2026). Signal Theory: The Architecture of Optimal Intent Encoding in Communication Systems. SORX Research.