The Brand Measurement Stack

Three tiers, from research-grade instruments to a copy-paste prompt pipeline. Instruments produce measurements with noise floors and audit trails; prompts produce structured readings. Start at the tier that matches the decision you need to make.

Measure: The PRISM Instrument Family

PRISM (Perception Response Instrument for Structured Measurement) is the open instrument family for multi-dimensional LLM brand perception. A domain-neutral five-layer scaffold is the constant; each instrument inherits it, pre-registers its protocol before any data collection, and logs every API call. Each instrument below is a published paper you can run: fork the repository, substitute your constructs, execute the pipeline.

2026as

PRISM-B (brand)

PRISM-B: A Structured Measurement Instrument for Multi-Dimensional Brand Perception

The brand instrument, and the paper that defines the shared five-layer scaffold (PL0 specification, PL1 configuration, PL2 prompts, PL3 logged sessions, PL4 analysis) every family member inherits. Pre-registered, reproducible by construction, every API call logged.

GitHub Zenodo DOI
2026az

PRISM-M (metamerism)

Measuring Perceptual Indistinguishability: A Pre-Registered Metamerism Instrument for AI Brand Perception (PRISM-M)

A pre-registered dual-floor instrument for measured metamers: when do two different brands become perceptually indistinguishable to an AI observer?

GitHub Zenodo DOI
2026ba

PRISM-T (version drift)

Separating Instrument Drift from Brand Signal: A Pre-Registered Model-Version Tracking Instrument for AI Brand Perception (PRISM-T)

A sealed pinned-panel version floor: separates what changed in the brand from what changed in the model between vendor releases.

GitHub Zenodo DOI
2026bb

PRISM-C (choice gap)

From Stated Perception to Revealed Choice: A Pre-Registered Instrument for the Choice-Perception Gap in AI Brand Perception (PRISM-C)

Measures the gap between what a model says it perceives and what it actually chooses, against a choice-elicitation operator floor.

GitHub Zenodo DOI

PRISM-O (organizational) — Measures organizational self-description against the OST optimization hierarchy. The organizational member of the family is presented on the OST toolkit (2026bd). PRISM-H (human-subject confirmatory) is planned.

Read: The Brand Spectrometer

The applied instrument of the theory. The Brand Spectrometer reads cohort-resolved, eight-dimensional brand perception from dated public artifacts — its unit of measurement is the reflection, one artifact read once across all eight dimensions — and reports every cohort difference against self-computed noise floors. Open, in-browser, no account; a demo brand is loaded.

Open the Brand Spectrometer meter.spectralbranding.com · methods paper 2026ax

Start: The Seven-Module Prompt Toolkit

The entry tier: seven structured prompts that turn any capable LLM into an SBT analyst. Copy a module from GitHub, paste it into the model, specify your brand; each module produces standardized YAML matching its template. A Python validation layer enforces mathematical bounds from the research papers (metric axioms, metamerism conditions, cohort boundaries, channel capacity, trajectory convergence, diffusion dynamics, allocation optimality) on every profile the pipeline produces.

  1. 01 · Brand Decomposition Prompt Template
  2. 02 · Observer Mapping Prompt Template
  3. 03 · Cloud Prediction Prompt Template
  4. 04 · Coherence Audit Prompt Template
  5. 05 · Emission Strategy Prompt Template
  6. 06 · Re-collapse Simulation Prompt Template
  7. 07 · Resource Allocation Prompt Template

Supporting resources: Atom Taxonomy · Framework Specification · Glossary · Framework Reference · Brand Code

Related

Find your fast path through the corpus by role on the Guide. Understand the framework behind the stack at the theory page. See results from real brands in the case studies. The organizational counterpart — OrgSchema Consult and the org-side stack — is the OST toolkit.