Skip to content
Alytixx

Navigation

Published English navigation

Methodology

See how every Alytixx conclusion is measured, limited, and supported

Alytixx validates the website source and discovered scope before interpreting results. It combines rule-based, derived, provider, structured AI-assisted, and manually confirmed evidence; calculates scores only for applicable measured signals; records coverage and limitations; applies quality gates; and creates immutable report revisions in which important conclusions remain traceable to evidence or an explicit limitation.

Last reviewed
2026-07-22
Methodology version
1.0

Methodology version: 1.0

Last reviewed: 2026-07-22

Six principles

Evidence before narrative

Important conclusions should remain traceable to a page, query, competitor, provider, structured observation, or explicit limitation.

Quality before score

Alytixx determines whether the source, discovery, and final output are reliable before presenting an authoritative overall result.

Coverage before confidence

A score must show how much of the planned and applicable scope was actually measured.

Missing is not zero

Not configured, not applicable, unavailable, and unmeasured are separate states.

One revision, one truth

Every final report revision is immutable. Approved projections should reference the same revision.

Recommendations must be verifiable

A useful action identifies what should change, why it matters, which evidence supports it, and how completion can be checked.

From URL to final revision

URL accepted
→ source quality checked
→ identity, market, language, and site type discovered
→ pages, competitors, search scenarios, and providers planned
→ selected evidence analyzed
→ scores, findings, recommendations, and limitations assembled
→ report invariants checked
→ immutable revision created

The public URL pre-check is outside this paid lifecycle. It confirms only basic public reachability and metadata.

Execution status and report quality

A process may complete while the result still needs review.

Execution status describes progress.

Quality status describes whether the result can be trusted and published.

Current quality states:

  • valid;
  • needs review;
  • invalid.

Source quality

Alytixx checks whether it is analyzing the intended public website rather than:

  • a challenge page;
  • access-denied response;
  • placeholder;
  • parked domain;
  • insufficient page;
  • unsupported resource.

When source quality fails, authoritative scores should be suppressed.

Discovery and scope

Alytixx may discover:

  • company or brand;
  • aliases;
  • normalized domain;
  • language;
  • market;
  • location;
  • site type;
  • category;
  • important pages;
  • page roles;
  • competitors;
  • search scenarios;
  • available providers.

When identity, topic, market, or location is unreliable, dependent analysis should be blocked rather than manufactured.

Capability states

A capability may be:

  • implemented;
  • provider-dependent;
  • applicable only;
  • available through Monitoring or another path;
  • planned;
  • not applicable.

Planned capabilities do not appear as current measured results.

Signal types

Rule-based

A deterministic condition from captured evidence.

Derived

A value calculated from normalized observations.

Provider-based

Data returned by an external source.

Structured AI-assisted assessment

A constrained evaluation using selected evidence and a defined output contract.

Manually confirmed

A reviewed scope, entity, role, or another fact.

Signal types are not assumed to have equal certainty.

Scores and coverage

Parameter results may roll into scoring groups and then into an overall covered summary.

Numerical denominators include applicable measured signals.

They exclude unavailable and not-applicable values.

The report should show:

  • analyzed count;
  • planned count;
  • percentage;
  • excluded scope;
  • limitation.

Business-critical overrides

Strong technical results should not hide a critical missing CTA, unclear offer, weak trust evidence, or materially incomplete commercial decision.

An override changes interpretation or status. It does not erase measured strengths or invent a zero.

Findings, recommendations, and evidence

A finding describes the measured gap.

A recommendation may include:

  • action;
  • business rationale;
  • page;
  • impact;
  • effort;
  • owner;
  • acceptance;
  • evidence;
  • limitation.

Related signals may be deduplicated around one root problem.

Readiness and visibility

SEO readiness is not search ranking.

AI Search Readiness is not an AI mention or citation.

Measured visibility requires explicit provider snapshots with scenario, surface, freshness, and coverage.

Competitor benchmark

A reliable benchmark preserves:

  • competitor role;
  • comparable page;
  • confidence;
  • self;
  • average;
  • best;
  • ideal;
  • coverage;
  • limitation.

Blocked discovery produces no benchmark score.

Revisions and monitoring

Later reports may classify:

  • new;
  • persistent;
  • resolved;
  • reappeared;
  • changed;
  • provider unavailable;
  • changed coverage.

A coverage difference must not automatically appear as website improvement or decline.

Frequently asked questions

See the methodology inside the report

View the Sample Report