Methodology
Every score published by Brand & Signal is produced by this methodology. It is versioned, and results are only comparable within one version.
Last updated
Why this exists
Digital performance is bigger than SEO, and bigger than your website. It is fragmented across brand, content, UX, engineering, search, AI, authority, distribution, conversion and response. Each of those has its own vendors, its own metrics and its own idea of what good looks like. None of them measures the whole.
The result is organizations that score well on every partial measure and still lose. Brand & Signal measures the whole thing on one scale, with a published method, so that performance can be compared, tracked and argued about on evidence.
The gaps between those slices are widening as AI systems become intermediaries for discovery, because those systems draw on brand clarity, technical delivery, structured data and third-party corroboration simultaneously. The framework is deliberately agnostic: it does not favour a platform, a vendor, a channel or a discipline, and it can be applied to any organization from public evidence.
Brand and Signal
Brand Integrity
7 dimensionsWhat an organization means, and whether that meaning survives contact with a real person. Brand dimensions govern clarity, distinctiveness, credibility and the quality of the experience itself.
Signal Strength
8 dimensionsWhat machines and markets can actually detect. Signal dimensions govern whether the organization is legible to search engines, AI systems, citation networks and distribution channels — and whether attention converts and is measured.
How they work together
The dimensions are not independent. Scoring them separately and reading them together is the point of the instrument.
- Signal dimensions carry Brand dimensions. Meaning that cannot be crawled, rendered or retrieved does not reach anyone, however good it is.
- Brand dimensions give Signal dimensions something to carry. Visibility without a proposition converts once and does not compound.
- Engineering Strength is upstream of almost everything: delivery faults cap the evidence available elsewhere.
- Structured data and governance are evidence and operating practices, not additional scored dimensions.
- Brand Entity Presence establishes who the organization is; Third-Party Verification tests whether external evidence supports that identity.
How to read a benchmark
A benchmark is a structured argument from evidence, not a leaderboard.
- Read the dimension profile before the two Indexes. The shape matters more than either number alone.
- Compare like with like: scores are only comparable within one methodology version and one study.
- Treat low Signal scores with high Brand scores as a distribution problem, and the inverse as a substance problem.
- Check the evidence sources listed for each dimension before disputing a result.
- Use the priority order, which is sequenced by impact — not the lowest score, which is often the cheapest.
The Signal Gap
Brand and Signal are scored separately and reported side by side. Where a result sits across the two determines what the work actually is — and in what order.
Present, but interchangeable
The infrastructure exists and machines can retrieve it. What they retrieve reads like a template: category language, no stated position, nothing a competitor could not also claim.
Diagnosis: Content and positioning gap.
Meaningful, but undetectable
The proposition is clear and the experience is credible to a person who arrives. Search engines, AI systems and citation networks cannot reliably reach, parse or attribute it.
Diagnosis: Findability and distribution gap.
Foundation problem
Neither side supports the other. Isolated fixes tend not to move the score, because the faults capping performance sit upstream of the work being done.
Diagnosis: Rebuild, sequenced by dependency.
Target state
Meaning and detectability are moving together, and the distance between the two sides is small enough that neither is holding the other back.
Diagnosis: Maintain through governance.
The two sides are read side by side and never blended into a single composite. The distance between them — the Signal Gap — is itself the finding.
What is evaluated
Fifteen dimensions: 7 on the Brand side governing meaning, and 8 on the Signal side governing detectability. Every organization in a study is scored on all fifteen, regardless of sector, so that results remain comparable.
- Owned digital properties, including subdomains and campaign sites in scope for the study
- Publicly available content, documentation and disclosure material
- Search visibility against a fixed, published query set
- AI findability against a fixed, published prompt set across multiple assistants
- External corroboration: citations, entity records, directories and reviews
- Technical delivery, accessibility and structured data on a sampled template set
Evidence sources
Only publicly observable evidence is used. Organizations are not contacted for participation and do not supply data. Nothing private — analytics, internal process, unpublished work — enters a score, in either direction.
- Owned properties
- Rendered pages as a visitor and as a crawler receives them, including markup, structured data and delivery characteristics.
- Search results
- Positions and presentation against a fixed query set, captured on defined dates.
- Generative systems
- Responses to a fixed prompt set across multiple assistants, recorded verbatim with dates.
- Third-party record
- Citations, entity and directory records, professional registers, reviews and press.
- Automated testing
- Accessibility, performance and structured-data validation on a sampled template set.
By Sector
Sector specific — I
Same seven dimensions. Different named sources.
The dimensions never change — only the evidence used to populate them. Each cell below is a citation, not a score: the reviewer checks a client's actual keyword copy or narrative claim against that named source, and the distance between them sets the 1–10.
| Mining & Resources | Telecom & Wireless | Real Estate | Retail & E-commerce | Technology & SaaS | |
|---|---|---|---|---|---|
| Audience-Intent Keyword Source | SEMrush Mining Sector Study; Google E-E-A-T | General SEO practice (flagged gap); Google E-E-A-T | NAR trend reports; Zillow Research keyword studies | Google Merchant Center guidelines; Salesforce Shopping Index | G2 category reports; Gartner Peer Insights |
| Narrative / Differentiation Source | IR Magazine Benchmarks; PwC Global Investor Survey | Edelman Trust Barometer; applied by analogy from IR literature | NAR Profile of Home Buyers & Sellers; Edelman Trust Barometer | Baymard Institute UX benchmarks; Edelman Trust Barometer | Demand Gen Report B2B Buyer Survey; Edelman Trust Barometer |
| Trust & Compliance Anchor | NI 51-102 / NI 43-101; TSX-V Continuous Disclosure | ACSI; J.D. Power carrier satisfaction studies | RESO Data Dictionary (MLS); Fair Housing Act (HUD) disclosure | FTC endorsement/review guidelines; PCI-DSS payment trust | SOC 2 / ISO 27001 trust signals; verified-review policies |
| Freshness / Recency Signal | SEDAR+ filing cadence; Google Freshness Update | Google Freshness Update; carrier editorial-hub cadence | MLS listing sync cadence; Google Freshness Update | Google Freshness Update; inventory / price sync cadence | Changelog / release-notes cadence; Google Freshness Update |
Sector specific — II
The framework extends as far as the sector list does.
This is the same table, continued — 6 more sectors, same 4 source rows. Nothing new to learn here beyond the entries themselves; skim for your sector and move on.
| Health & Wellness | Travel | Hospitality & Leisure | Financial Services | Energy & Utilities | Government | |
|---|---|---|---|---|---|---|
| Audience-Intent Keyword Source | Google YMYL/E-E-A-T guidance; Press Ganey patient studies | Phocuswright travel research; Google Hotel Ads guidance | STR hotel benchmarking; Google Business Profile category data | Google YMYL/E-E-A-T guidance; Javelin Strategy & Research | U.S. EIA terminology & data; Google E-E-A-T | Plain Writing Act / plainlanguage.gov; Google Search Central |
| Narrative / Differentiation Source | Press Ganey / NRC Health surveys; Edelman Trust Barometer | Skift / TripAdvisor research; Edelman Trust Barometer | Forbes Travel Guide / AAA Diamond criteria; Edelman Trust Barometer | J.D. Power banking/insurance studies; Edelman Trust Barometer | GRI Standards; TCFD alignment | Digital.gov; U.S. Web Design System (USWDS) content standards |
| Trust & Compliance Anchor | HIPAA; The Joint Commission accreditation signals | ADA / IATA accessibility norms; review-authenticity policy | ADA accessibility; local health & safety licensing disclosure | FINRA / SEC Reg BI disclosure; FDIC / NCUA insurance signals | NERC reliability standards; TCFD / SEC climate disclosure | WCAG 2.1 AA + Section 508; ADA Title II |
| Freshness / Recency Signal | Google Freshness Update; clinical-update cadence (CDC/WHO) | STR data cadence; Google Freshness Update | Live booking-engine availability; Google Freshness Update | Rate / market-data update cadence; Google Freshness Update | Outage / rate update cadence; Google Freshness Update | Agency publication-cadence mandates; Google Freshness Update |
Scoring scale
Each dimension is scored from 1.0 to 10.0 against published criteria. The two Indexes are reported separately; Brand & Signal does not create an Overall, Combined or Total Score.
- 8.5–10.0 Leading
- Externally verifiable category leadership.
- 7.0–8.4 Strong
- Strong, repeatable evidence with isolated gaps.
- 5.0–6.9 Credible
- Credible category performance with material room to improve.
- 3.0–4.9 Developing
- Material gaps that limit performance.
- 1.0–2.9 Weak
- Fundamental evidence or delivery faults.
Index calculation
The Brand Integrity Index is the arithmetic average of the seven Brand Integrity Review dimensions. The Signal Strength Index is the arithmetic average of the eight Signal Architecture Review dimensions.
The Signal Gap is Signal Strength Index minus Brand Integrity Index. A negative gap indicates that organizational strength is not being represented as strongly across the wider digital ecosystem.
The two Indexes are never blended. Sector-specific evidence may change what is inspected, but it does not create a weighted overall score.
Human, automated and AI analysis
- Automated
- Deterministic testing — performance, accessibility, markup validity, structured data, crawl and render behaviour. Repeatable and re-runnable on demand.
- AI-assisted
- Large-scale content and entity analysis, and the measurement of how generative systems describe an organization. Outputs are recorded and treated as evidence, never as a score.
- Human
- Every score is set by a human evaluator against the published criteria. Judgement dimensions — proposition, narrative, differentiation — are human-scored and reviewed by a second evaluator.
No dimension is scored automatically. Machines gather and test; people decide.
Timing
- Every study states its evaluation window, and all organizations in a study are measured within it.
- Search and AI observations are date-stamped, because both change without notice.
- A score is a point-in-time measurement, not a standing rating.
- Standing indices state their re-measurement cadence up front.
Versioning
The methodology is versioned. Current version: v1.0. Scores are comparable within a version and are not comparable across versions.
Minor versions clarify criteria without changing what a score means. Major versions change dimensions or criteria; published studies are not rescored against a new major version, they are re-run or left as historical record with their version stated.
Conflicts of interest
Commercial relationships do not change scoring. No organization can pay to be included, excluded, re-scored or removed. There is no sponsored placement in any study, and no advertising in the research library.
Kumu Agency helped originate the framework and is a founding implementation partner.
Where an evaluator has a relationship with an organization under evaluation, that evaluator does not score it, and the relationship is disclosed on the study page. Studies that include organizations connected to the framework's origin disclose that connection in the same place as the result.
Corrections
Any organization may request a correction. We review the evidence, and where a factual error is confirmed we republish with a dated change note. Corrections address factual accuracy only — scores cannot be purchased, negotiated or removed, and no commercial relationship changes a result.
Limitations
- Public evidence only: what an organization does privately or exceptionally well behind a login cannot be measured here.
- Sampling is used for template-level technical testing rather than exhaustive crawling.
- Generative-system outputs vary between runs and change over time; measurements are recorded with dates and prompt sets rather than presented as stable facts.
- Judgement dimensions carry evaluator variance, mitigated by second review and published criteria but not eliminated.
- A framework score is a measure of measurable digital performance. It is not a measure of an organization's competence, product quality or commercial results.
Version history
- v1.0 — Current
- First published version. Fifteen dimensions across two separately reported Indexes, 1.0–10.0 dimension scoring, public-evidence-only sourcing, and human-set scores with automated and AI-assisted evidence gathering.