Spectral Branding

Two frameworks. One system.

Businesses are simultaneously perception systems and specification systems. We publish open-source research frameworks for both sides.

Spectral Brand Theory (SBT) models how brands are perceived — eight perceptual dimensions, observer-specific spectral profiles, math-hardened validation, an AI-native analytical pipeline. Organizational Schema Theory (OST) models how businesses are specified — a reverse-design TDD methodology where customer-experience goals are acceptance tests and each operational layer validates the layer above it.

Aaker drew the map. SBT built the microscope. OrgSchema built the specification. Neither alone is complete.

Everything is open: 50+ working papers on Zenodo (CC BY 4.0), machine-readable sources on GitHub, runnable toolkits.

---

Start here — find your path

The fastest route is the Guide. Pick your role — CMO, COO, CFO, researcher, AI agent, press — and it routes you to the seed papers, the decisions you can act on, the runnable instruments, and exactly how to widen from there. It is a generated projection of this corpus (not a hand-kept list), so it never drifts from the papers.

Two shortcuts that skip straight to a tool:

• Measure a brand → the live Brand Spectrometer — your brand's shape in minutes.
• Specify a business → the orgschema-toolkit — a complete business in ~40 minutes.

Your own AI can consume the same map directly: corpus-guide holds the machine-readable corpus-map.json (papers ↔ terms ↔ claims) and guide-routing.json (role accelerators) for seed → scope traversal.

---

Use the corpus with your own AI (the unbundled think-tank)

A think-tank bundles two things: the people who think, and the knowledge they think with — the tank. This corpus unbundles them. The published frameworks are the tank — a calibrated, versioned body of knowledge, method, and runnable instruments. You bring the think — your own AI — and your data stays with you.

• The lens (ours, open) — the papers, each repo's paper.yaml/ONTOLOGY.yaml substrate, and the instruments (the Brand Spectrometer, the OrgSchema validator).
• The intellect (yours) — any AI, local or cloud, pointed at the lens.
• Your case (yours, private) — your question and your data; never uploaded here.

Point your agent at the relevant reading path (or a repo's AGENTS.md), give it your case, and it reasons through the lens — and, like the instruments, tells you what it cannot resolve instead of inventing an answer. It diagnoses and analyzes; you decide and act. Measurement, not management.

---

The core ideas (concept map)

Spectral Brand Theory. A brand emits signals across eight independent perceptual dimensions — Semiotic, Narrative, Ideological, Experiential, Social, Economic, Cultural, Temporal. Each observer filters those signals through a personal spectral profile (a weight vector) and constructs a perception cloud — a probability distribution over possible convictions, not a single "brand image." Observers group into cohorts (perceptual, not demographic). Two different emission profiles can produce the same perception for a cohort — metamerism. Measurement, not management: SBT measures the perception you do not control.

Organizational Schema Theory. A business is a specification cascade across six tiers: customer-experience goals at the top are acceptance tests; each lower layer (positioning → capabilities → processes → roles → sourcing) must validate the layer above. Verification is an operator that projects the sprawling implementation back onto the specified subspace. Distinct configurations that yield equivalent outputs are organizational metamerism. Specification, not improvisation: OST makes the business you do control explicit and testable.

The bridge. Both are instances of the rendering problem: a compact specification (genome / org-schema / brand guidelines) is rendered into a vast implementation (organism / operations / signals), from which an emergent layer arises (phenotype / culture / perception). The meaningfulness program generalizes this to any knowledge artifact: a small substrate (the claims) renders into much larger prose.

            SPECIFICATION  ──render──▶  IMPLEMENTATION  ──emerge──▶  EMERGENT LAYER
  biology    genome                     organism                     phenotype
  business   org-schema (OST)           operations                   culture
  brand      guidelines                 emitted signals              perception cloud (SBT)
  research   substrate / spine          meaning                      meaningfulness

Per-term definitions and their formal relations live in each paper's ONTOLOGY.yaml + GLOSSARY.md.

What this work covers — and what it does not

Every claim here is a claim about one stage of one chain. Most disappointment with research like this comes from a claim being read as covering more than it does, so the boundaries are stated up front rather than discovered later.

On the brand side, the chain runs budget → media units → impressions → attention → perception → memory → behaviour → revenue.

• Owned: perception, and its valence sub-layer. This is the only stage with executed, pre-registered, publicly logged measurement behind it.
• Touched: budget (as an allocation across perceptual dimensions, not across media channels) and behaviour (stated and revealed choice of model cohorts — there is no human behavioural data anywhere in this corpus).
• Not covered: media buying, delivery and impressions, attention, memory, and revenue. Attention is ceded deliberately to external models, not overlooked.
• Of the seven links in that chain, this corpus owns none as specified. What it owns instead is a substitute route — emitted signal → perception, formal with calibrated demonstrations, never a field measurement — and executed evidence on a collapsed link, perception → behaviour, that the classic chain does not contain.

The single most useful caveat: the advertising chain is a transmission ladder — how a signal travels to a person. This corpus's own ladder is a construction ladder — what a decision-maker builds on top of a reading. They intersect at exactly one node, perception, and nowhere else. Laying these results over an advertising chain therefore implies coverage of two stages and three links that are entirely empty. That is the largest over-claim risk in the whole programme, and it is a presentation risk rather than an error in any paper.

On the organizational side there are three chains, not one, because an organization, a transaction and a research artifact are different objects and a chain is a property of an object rather than of a theory: the specification cascade inside one firm, what crosses an ownership boundary, and how a knowledge artifact is authored and verified.

• Owned: the specification cascade and its levels; the six-layer decomposition of an acquisition target; the algebra of what a transaction transfers; verification and audit, which is the strongest region of all; and the machinery for specifying and checking research artifacts.
• Ceded by decision: execution. It is not a layer in the stack but the rendering of the process layer by the people and systems that carry it out — a position, not a gap.
• Not covered: post-deal value realisation, and the prediction of deal success generally.
• Across all three of those chains there is exactly one link with field evidence behind it, and it is a bounded necessary-condition result about avoiding one specific integration-failure mode — a pre-close negative screen, never a predictor of a good deal.

The organizational caveat, and it is the counterpart of the one above: the practitioner ladder is causal — each stage is supposed to produce the next, and the last stage is money. The specification cascade is a traceability relation — each level is justified by the level above it, and its terminal object is a checkable contract rather than a result. They are read in opposite directions, and only one of them has an instrument here. This work verifies the upward relation and has no executed evidence on the downward one. Laying the cascade over a strategy-to-execution ladder will be read as a performance claim; it is not one.

The uncomfortable one, stated because a reader is entitled to it: there is no executed study here in which specification quality is the independent variable and firm performance is the dependent variable. The archival design for that question is complete and pre-registered; it has not been run. The nearest clean causal result on specification-beats-style has software agents as its subject, not companies. Anyone citing this work as evidence that a better-specified organization performs better is ahead of the evidence, and that includes us.

The full accounting, marked stage by stage with the paper keys behind each verdict, is published rather than summarised: the brand chain at spectralbranding.com/coverage, and the three organizational maps at orgschema.com/coverage. Both mark every stage as owned, touched, ceded or empty, with the evidence grade and the papers behind each verdict. The gaps those maps imply are listed at spectralbranding.com/open-problems.

Where to go instead. For delivery, reach and frequency, the media-planning tradition. For attention, the attention-measurement vendors and the published models this corpus explicitly differentiates from rather than competes with. For recall and mental availability, the memory and salience literature. For attribution and incrementality, the experimental and econometric marketing-measurement literature. A map that only says no is not useful; these are real questions with real answers elsewhere.

---

The repositories

Spectral Brand Theory

Repo	What it is
sbt-framework	7-module AI-native prompt kit with YAML templates + 7 math validators. Run a spectral brand audit on any brand.
brand-spectrometer	The open in-browser instrument — reads cohort-resolved, 8-dimension brand perception from any atlas and flags which differences between audiences clear the noise floor. Live at meter.spectralbranding.com. Methods paper: The Brand Spectrometer (2026ax).
sbt-papers	Research papers — framework spec, epistemological pipeline, mathematical foundations (R0–R22), AI-era studies, and the measurement instrument. 30+ papers on Zenodo.
brand-functions	Brand Function Registry — machine-readable brand specs (JSON) for 26 brands across the 8 dimensions. The schema.org of brand identity.
brand-code	Executable brand identity. The visual identity is a function, not a file.

spectralbranding.com · Substack · Zenodo

Organizational Schema Theory

Repo	What it is
orgschema-framework	Python validator + JSON Schema for the 6-level TDD cascade format.
orgschema-demo	Spectra Coffee — a complete business specified as organizational schema. 25 YAML files + CI/CD.
orgschema-toolkit	8-module AI prompt pipeline. Specify a complete business in ~40 minutes.
orgschema-papers	Research papers — six-tier ontology, projection cascade, tier-rotation, tier-allocation, verification-as-operator, specification-readiness, the rendering problem, the OST audit, and tier-bundle algebra for acquisition integration (with its large-N empirical test and multi-LLM coding instrument).

orgschema.com · Substack · Zenodo

Meaningfulness & Meaning

Meta-corpus on knowledge work in the post-AI era — the substrate↔rendering distinction operationalized for any propositional artifact (paper, memo, brand spec, decision log).

Repo	What it is
meaningfulness-papers	Spec-Based Research in the Post-AI Era (theory) + Same Meaning, Different Prose (empirical) + Invention as a Structural Operation (abduction as structural re-coordinatization).

Philosophy of Science

When a claim about where a system ends has empirical content — and what a domain must supply for such a claim to be testable at all. Uses organizations as instruments rather than as subjects.

Repo	What it is
philosophy-papers	When a Boundary Claim Becomes Testable (2026bo) — a three-condition criterion for adjudicating scale-relative boundary claims, with one dual-coded acquisition as an existence proof. Data.

Applied specifications

Repo	What it is
paper-spec	Paper Spec v0.1.0 — machine-readable YAML standard for scientific claims. What a paper claims.
paper-repo	Paper Repo v0.1.0 — Git-native protocol for scientific publishing. How a paper is managed.
activity-spec	YAML schema + LLM prompts for machine-readable professional profiles.
negotiation-spec-experiment	Empirical study — specify an LLM negotiator's objective, or style it? 7 arms, 4,920 dyads. Zenodo.

---

Reading paths (researchers & students)

Understand SBT (brand perception):

1. The framework — spectral-brand-theory (2026a): the 8 dimensions, emission profiles, perception clouds, cohorts.
2. Where it sits — r0-literature-survey (2026c): SBT vs Aaker / Keller / Kapferer / MDS. And spectral-identity (2026n): formalizing Aaker's identity model.
3. Why eight — r11-dimension-justification (2026r).
4. Mathematical spine — r1-formal-metric (geometry) → r2-spectral-metamerism (projection bounds) → r3-cohort-boundaries (concentration of measure) → r4-sphere-packing (market capacity) → r5-specification-impossibility.
5. Dynamics — r6-diffusion-dynamics, r9-nonergodic-perception, r12-coherence-resilience, r18-spectral-dynamics, r22-spectral-gap-restoration.
6. AI era — r15-ai-search-metamerism, r16-ai-native-brand-identity, r21-spectral-immunity, r19-rate-distortion.
7. Measure it — the PRISM instrument family (prism-b-brand — PRISM-B (2026as) for brand perception, plus the pre-registered siblings PRISM-M (2026az, metamerism), PRISM-T (2026ba, version-floor drift), PRISM-C (2026bb, stated-vs-revealed choice), and PRISM-O (2026bd, organizational optimization depth)) and the Brand Spectrometer (2026ax — the applied instrument built on PRISM: a reproducible pipeline that reads cohort-resolved 8-dimension perception from public artifacts under ground-truth-absence, with operator/artifact noise floors), and the applied cases r10-dove-case-study, r17-brand-triangulation.
8. When the score suffices — brand-correspondence-principle (2026au): the regime where an aggregate brand score is sufficient for a decision, and the measurable cost when it is not.
9. Reach a perceptual cohort — reaching-a-perception (2026av): a perceptual cohort has no native media address and does not need one — three measurable bridges from perception to delivery (broadcast a dimension and let self-selection route it, follow atom provenance, or quantify the proxy-join loss), and why the address-free route is the post-cookie tailwind.

Understand OST (how businesses are specified):

1. Six tiers — six-tier-ontology (2026ag) + the OST main paper.
2. The rendering problem — the cross-domain bridge (biology / org / brand).
3. Mechanics — verification-as-operator (2026ae), projection-paper (2026m), org-as-metadata (2026af, organizational metamerism).
4. Where to invest — tier-allocation (2026aj), tier-rotation (2026ai), brand-as-modular-layer (2026ah), capability-as-projection (2026al).
5. Readiness — specification-readiness (2026am) + its empirical companion (2026an); diagnose with orgschema-audit (2026ar).

Understand the meta-program (research after AI): meaning-meaningfulness (2026ao/2026ap) · r14-paper-as-repository (2026u) · canon-as-repository (2026w) · paper-spec · paper-repo.

---

Evaluate a business case (executives)

A concrete four-step recipe, using both frameworks:

1. Place your business on the six tiers (OST) — use orgschema-toolkit; see orgschema-demo for a worked example. Output: an explicit, testable spec with the acceptance tests that define "working."
2. Measure your brand's 8-dimension emission profile and the cohorts that matter (SBT) — run sbt-framework; compare against the 26 specs in brand-functions.
3. Find the two gaps — the specification gap (operations exceed or miss the spec → OST verification-as-operator, orgschema-audit) and the perception gap (the cohorts' perception clouds diverge from intent → SBT r10-dove-case-study, r12-coherence-resilience).
4. Decide where capital/attention creates the most value — tier-allocation (2026aj, "where to invest within the firm"), tier-rotation (2026ai), and the AI-substrate-dependence lens for new bets. For brand-portfolio moves: r8 → r21-spectral-immunity (when portfolio interference does and does not hold), brand-as-modular-layer (2026ah).

CMO-first? Skip to steps 2–3. CFO-first? Steps 1 and 4. The point: you can only manage the specification you control once you can measure the perception you don't.

---

For AI agents

• Start from the machine map: corpus-guide — corpus-map.json (papers ↔ terms ↔ claims) and guide-routing.json (role accelerators). Route a task by role, then widen from its seed papers via shared terms and citation edges (seed → scope); ground the answer in the widened set and state what you cannot resolve. The machine view of the Guide is at spectralbranding.com/guide.json.
• Route a business decision to the right instrument via OrgSchema Consult — consult.orgschema.com/consult.json is the machine manifest that maps each decision to the corpus and the runnable instrument that answers it (a measurement question hands off to an instrument, not a literature summary). It routes brand-perception decisions to the Brand Spectrometer, an operating-model decision to the OrgSchema Toolkit, and the organizational decision "is a stated transformation real or rhetorical?" to PRISM-O.
• Every paper folder has an AGENTS.md — a file-by-file guide. Read it first.
• Citations are [@citation_key] keys, resolved against the bundled <key>.bib. A [@key] is never a typo — it renders to "(Author Year)". See each folder's CITATIONS.md.
• Structured claims (what a paper asserts, assumes, and depends on) are in paper.yaml (Paper Spec) — parse this instead of the prose.
• Terms are defined in GLOSSARY.md with formal relations in ONTOLOGY.yaml (owns / imports / refines).
• To answer a grounded question: route via the Guide → open the seed paper.md → resolve cites against the .bib → widen via corpus-map.json.

When to measure a brand, use the Brand Spectrometer — not the papers. If the task is to obtain cohort-resolved, 8-dimension brand perception from real public evidence, or to test whether two audiences genuinely differ, route to the Brand Spectrometer: live tool meter.spectralbranding.com, code brand-spectrometer, methods paper 2026ax.

• Use it for measurement questions — "what does this brand's perception cloud look like?", "do these two cohorts read the brand differently?", "which dimension separates them?"
• How it works (so an agent knows what it returns): dated public artifacts → reflections (per-artifact 8-D readings) → cohort vectors with self-computed operator and source noise floors; it declares a cohort difference resolved only when the signal-to-noise clears the floor (triangulated), and otherwise reports sub-resolution — never "no difference," never a finding.
• Do not use it as a single-number brand tracker, and note it measures cohorts, never individuals (no person is identified or stored).
• Pair with the papers: for the theory of what it measures, read SBT (2026a); for whether one score is enough and how to act, read 2026au; for reaching a measured cohort, 2026av. The emission unit is a signal; the Spectrometer's measurement unit is a reflection — distinct concepts (see sbt-framework/docs/GLOSSARY.md, Measurement layer).

When to measure whether an organization's stated optimization is real, use PRISM-O — the org-side counterpart. If the task is to test whether a company's stated transformation is actually enacted — not whether the strategy reads well — route to PRISM-O (pre-registered protocol + code prism-o-optimization-depth, paper 2026bd). It reads the two-channel public record (strategy prose vs reported actions), codes each improvement onto a four-rung optimization-depth ladder under a fixed rubric, and returns the signed stated-actual gap against a cross-family operator floor — floored and no-rescue, so a claim that cannot clear the noise is reported as sub-resolution, never manufactured. It joins the SBT and OST corpora at the producer–observer seam; OrgSchema Consult routes the "is this transformation real?" decision to it.

---

Cite · License · Contact

• License: papers CC BY 4.0; code MIT (dual-licensed per repo).
• Cite: each paper folder has a CITATION.cff and a Zenodo DOI.
• Author: Dmitry Zharnikov — ResearchGate · Zenodo (SBT) · Zenodo (OST)