TOPICALAUTHORITY.ORG TAO / ROOT

Methods

TOPICALAUTHORITY.ORG / METHODS / MTH-000 STATUS PROTOCOL SYSTEM ACTIVE
BRANCH / 05 · RESEARCH METHOD OPERATING SYSTEM

Methods make knowledge defensible.

Methods define how a question becomes a traceable result. They control scope, inputs, sampling, procedure, validation, uncertainty and reproducibility so that research can be inspected instead of merely believed.

MTH.OS / PROTOCOL SEQUENCECONTROLLED
MTH
ROOT / RESEARCH PROTOCOLFrom question to defensible decision
01Define the questionOBJECTIVE
02Bound the systemSCOPE
03Acquire the observationsDATA
04Execute a declared procedureMETHOD
05Test errors and alternativesCONTROL
06Report evidence and limitsOUTPUT
CLAIM STATEBOUNDED
PROVENANCEVISIBLE
REPEATABILITYREQUIRED
INPUT / 01Research question
CONTROL / 02Declared procedure
EVIDENCE / 03Traceable observations
OUTPUT / 04Bounded conclusion
MTH / 001 FOUNDATION

A method is not a tool. It is a controlled route.

A tool can return data. A method explains why those inputs were selected, how they were transformed, which errors were tested and what the result is allowed to mean.

A research method is a declared and repeatable procedure for transforming a bounded question and observable inputs into a traceable, qualified conclusion.

In digital research, the method must preserve location, language, device, time, source, sampling rule, transformation logic and known limitations. Without those controls, a number may be precise in format while remaining weak in meaning.

METHOD ≠A tool name
METHOD ≠A screenshot
METHOD ≠An unexplained score
METHOD =Question + controls + procedure + limits
MTH / 002 METHOD CHAIN

Eight gates separate an observation from a defensible claim.

Every gate has one responsibility. Skipping any stage changes what the final result can legitimately support.

01Question

State the exact uncertainty the research must resolve.

WHY
02Boundary

Define what belongs inside and outside the analysis.

WHERE
03Sample

Select observations using an explicit inclusion rule.

WHAT
04Acquisition

Record provider, endpoint, parameters, time and locale.

INPUT
05Transform

Normalize, classify, compare or calculate by declared logic.

PROCESS
06Validate

Check missingness, outliers, conflicts and alternatives.

CONTROL
07Interpret

Separate observed facts from derived or inferred meaning.

MEANING
08Report

Publish result, provenance, limitations and update state.

PROOF
MTH / 003 INTERACTIVE PROTOCOL

Different questions require different controls.

Select an objective. The engine exposes the minimum protocol required before the output should be treated as research rather than interface decoration.

METHOD COMPOSER / MTH-C01AUDIT PROTOCOL ACTIVE
MODE / AUDIT · DIAGNOSTIC SYSTEM

Diagnose a defined system against declared criteria.

QUESTIONWhere does the current system fail its stated coverage and routing requirements?
REQUIRED INPUTCanonical URL set, page roles, query/entity scope, crawl state and dated observations.
PRIMARY CONTROLPublished criteria, exclusions, duplicate-role checks and manually reviewed exceptions.
VALID OUTPUTPrioritized findings tied to observable evidence—not an unexplained universal score.

Boundary: an audit describes the analyzed snapshot. It does not prove causation or guarantee a future search outcome.

MTH / 004 SIX DIMENSIONS

Precision is not more decimals. Precision is controlled meaning.

A robust method makes six dimensions explicit before analysis begins. Together they determine whether another person could inspect, challenge or repeat the work.

D01 / SCOPEBOUNDARY

Semantic boundary

Define the subject, unit of analysis, inclusion criteria, exclusions and adjacent territory that must not contaminate the result.

OUTPUT / DECLARED RESEARCH UNIVERSE
D02 / SAMPLESELECTION

Observation frame

Explain which queries, pages, domains, entities, passages or time points were selected—and why those observations are sufficient for this question.

OUTPUT / INSPECTABLE SAMPLE RULE
D03 / DATAPROVENANCE

Source integrity

Record origin, endpoint or document, collection time, locale, device, provider frequency, transformations and missing fields.

OUTPUT / TRACEABLE INPUT STATE
D04 / LOGICPROCEDURE

Operational sequence

Specify normalization, classification, thresholds, comparisons, calculations and exception handling in the order actually executed.

OUTPUT / REPEATABLE PROCEDURE
D05 / ERRORVALIDATION

Failure controls

Test ambiguity, duplication, missingness, volatility, measurement bias, alternative explanations and cases requiring human review.

OUTPUT / KNOWN ERROR SURFACE
D06 / CLAIMLIMIT

Inference boundary

State what was observed, what was calculated, what is inferred, what remains uncertain and when the result becomes stale.

OUTPUT / QUALIFIED CONCLUSION
MTH / 005 QUESTION DESIGN

Bad questions manufacture noise. Good questions authorize decisions.

The research question must identify the object, condition, comparison, time and intended decision before any dataset is requested.

REJECTED / UNBOUNDED
“Does this website have topical authority?”
  • No defined subject boundary
  • No unit of analysis
  • No observable criteria
  • No time or comparison state
  • Invites an unsupported binary answer
ACCEPTED / OPERATIONAL
“Across the defined entity and intent set, which material coverage and routing gaps remain in the canonical corpus on the observation date?”
  • Names the analyzed corpus
  • Defines entity and intent coordinates
  • Targets observable gaps
  • Preserves snapshot timing
  • Produces an actionable repair list
MTH / 006 SCOPE CONTROL

Every result belongs to a defined coordinate system.

Search and AI outputs change across geography, language, device, time and source. These are not optional metadata fields; they are part of the observation itself.

CTRL / 01

System boundary

Declare the domain, directory, URL set, topic territory, competitors and excluded regions of the information system.

DOMAIN · CORPUS · TOPIC · EXCLUSIONS
CTRL / 02

Observation context

Preserve location, language, device, engine, collection time and data-refresh frequency with every measured state.

LOCALE · DEVICE · TIME · SOURCE
CTRL / 03

Decision horizon

Specify whether the result supports immediate diagnosis, weekly monitoring, longitudinal comparison or strategic planning.

SNAPSHOT · TREND · REVIEW DATE · DECISION
MTH / 007 EVIDENCE PROTOCOL

Data becomes evidence only when it supports a bounded claim.

A provider response, crawl or retrieved passage is an observation. The evidence chain must show how that observation relates to the question and why the interpretation remains defensible.

E / 01Source

Identify origin and authority.

E / 02Observation

Preserve the returned state.

E / 03Context

Attach locale, time and scope.

E / 04Transform

Declare processing logic.

E / 05Control

Test errors and conflicts.

E / 06Interpret

Separate fact from inference.

E / 07Claim

Report only supported meaning.

MTH / 008 CLAIM STATES

One interface. Four epistemic states.

Outputs must reveal whether a statement comes directly from a source, from deterministic calculation, from interpretation or from an illustrative model.

STATE / OBSERVED

Source observation

A returned field, published statement, crawl response or directly recorded event. Report with source, parameters and time.

STATE / DERIVED

Calculated result

A reproducible transformation of observed inputs. Report formula, rules, missing-data behavior and units.

STATE / INFERRED

Analytical interpretation

A reasoned conclusion supported by observations but not directly contained in them. Report alternatives and uncertainty.

STATE / ILLUSTRATIVE

Conceptual model

A synthetic example used to explain logic. It must never be presented as provider data, a public score or measured reality.

MTH / 009 FAILURE CONTROL

A method earns trust by exposing where it can fail.

The following defects can make a polished result analytically weak. Each requires a declared control rather than a visual confidence signal.

FAIL / 01Selection bias

The chosen sample systematically favors one result or excludes contrary observations.

CONTROL / DECLARE INCLUSION RULE
FAIL / 02Scope drift

The analysis expands into adjacent territory and changes the question during execution.

CONTROL / FREEZE BOUNDARY
FAIL / 03Temporal mismatch

Observations from incompatible dates are compared as if they describe one state.

CONTROL / ALIGN TIME WINDOW
FAIL / 04Provider substitution

Different metrics or sources are treated as interchangeable despite different definitions.

CONTROL / PRESERVE PROVENANCE
FAIL / 05Classification ambiguity

Intent, entities or page roles are assigned without decision rules or exception review.

CONTROL / PUBLISH TAXONOMY
FAIL / 06False precision

A decimal score hides uncertainty that the underlying observations cannot support.

CONTROL / REPORT CLAIM STATE
FAIL / 07Causal overreach

Correlation or sequence is described as proof that one change produced another.

CONTROL / TEST ALTERNATIVES
FAIL / 08Silent missingness

Unavailable data disappears from the analysis instead of being reported as an unknown.

CONTROL / EXPOSE NULL STATE
MTH / 010 · REPRODUCIBILITY RECORD

Every run leaves a research fingerprint.

A result should carry enough context for another analyst—or the same analyst later—to understand exactly what was examined, how it was processed and which version produced the output.

RUN RECORD / MTH-RUN-2026-001COMPLETE
01question_idcoverage_gap_defined_corpus
02observation_timeISO-8601 timestamp
03scopedomain / directories / exclusions
04localelocation / language / device
05sourceprovider / endpoint / document
06sample_ruleinclusion / exclusion / cap
07transform_versionnormalization + classification rules
08missing_statenulls / failures / unavailable fields
09claim_stateobserved / derived / inferred / illustrative
10review_statemachine checked + human reviewed
11method_versionMTH-1.0 / reproducible
MTH / 011 METHODS × LABS

The instrument returns data. The protocol governs meaning.

Each TAO Lab should expose the method that controls its inputs, transformations, validation and interpretation. The same interface can support different questions only when the protocol changes with the objective.

PROVIDERDataForSEO observations remain identified by product and explicit endpoint.
REQUEST STATELocation, language, limits, filters and collection time remain attached.
RETURN STATEProvider fields, nulls, task cost and freshness remain distinguishable.
TAO LAYERTAO classifies and prioritizes; it does not rename interpretation as provider fact.
LAB / 01SERPSERP Lab

Sampling, locale, device, result parsing and feature classification.

METHOD / SERP OBSERVATION →
LAB / 02DOMAINDomain Lab

Corpus boundary, visibility snapshot, page distribution and competitor context.

METHOD / DOMAIN PROFILE →
LAB / 03DEMANDKeyword Lab

Seed selection, expansion, deduplication, locale and demand interpretation.

METHOD / QUERY EXPANSION →
LAB / 04COMPARECompetitor Lab

Comparable-domain criteria, shared territory and asymmetry controls.

METHOD / COMPETITIVE SET →
LAB / 05AIAI Visibility Lab

Prompt set, run context, mention detection, citations and volatility.

METHOD / AI OBSERVATION →
LAB / 06ENTITYEntity Lab

Identity resolution, type assignment, attributes, relations and ambiguity.

METHOD / ENTITY RESOLUTION →
LAB / 07INTENTSearch Intent Lab

Classification taxonomy, result evidence, mixed missions and review.

METHOD / INTENT CLASSIFICATION →
LAB / 08ROUTINGInternal Linking Lab

Crawl boundary, graph construction, anchors, depth and orphan rules.

METHOD / LINK GRAPH →
LAB / 09MAPTopical Map Lab

Territory definition, entity-intent coordinates and cluster formation.

METHOD / TERRITORY MODEL →
LAB / 10GAPContent Gap Lab

Target and rival sets, keyword intersections, caps and gap classification.

METHOD / GAP ANALYSIS →
LAB / 11LINKBacklink Lab

Source identity, link state, anchor evidence, timing and comparison scope.

METHOD / LINK EVIDENCE →
LAB / 12RETRIEVALRetrieval Lab

Query set, source selection, passage relevance, grounding and citation trace.

METHOD / RETRIEVAL TEST →
MTH / 012 RESEARCH NETWORK

Build the method layer as a complete research cluster.

Twelve focused nodes separate foundations, question design, data collection, analysis, validation, reproducibility and AI-search research into distinct information responsibilities.

MTH / 000 · ROOT GOVERNANCEMethods

Research question → controlled procedure → traceable evidence → bounded conclusion.

MTH / 01FOUNDATION
What Is Digital Research Methodology?

The role of declared procedures, controls, evidence and limitations in digital research.

COREOPEN →
MTH / 02QUESTION
Research Question Design

Turning broad uncertainty into a bounded, observable and decision-relevant question.

OBJECTIVEOPEN →
MTH / 03SCOPE
Scope & Boundary Definition

Defining the subject universe, unit of analysis, exclusions and adjacent territory.

CONTROLOPEN →
MTH / 04SAMPLE
Sampling & Data Collection

Selecting observations and preserving source, locale, time and acquisition context.

INPUTOPEN →
MTH / 05SERP
SERP Research Methodology

Controlled observation of rankings, features, pages and query-level result states.

SEARCHOPEN →
MTH / 06ENTITY
Entity Research Methodology

Identity resolution, attributes, relations, ambiguity and representation depth.

SEMANTICSOPEN →
MTH / 07COMPARE
Comparative Analysis Methodology

Building valid comparison sets, normalized dimensions and qualified differences.

ANALYSISOPEN →
MTH / 08AUDIT
Content Coverage Audit Method

Diagnosing entity, intent, depth, overlap, evidence and internal-routing gaps.

DIAGNOSTICOPEN →
MTH / 09VALIDATE
Evidence Validation

Testing provenance, consistency, conflict, freshness and claim support.

EVIDENCEOPEN →
MTH / 10REPEAT
Reproducible Digital Research

Recording parameters, versions, transformations and review states for repeatable runs.

PROVENANCEOPEN →
MTH / 11LIMITS
Research Limitations & Uncertainty

Reporting incomplete data, volatility, bias, alternative explanations and confidence.

BOUNDARYOPEN →
MTH / 12AI
AI Search Research Methodology

Prompt sets, retrieval observation, citation tracing, run variance and grounding tests.

FRONTIEROPEN →
MTH / 013 GOVERNANCE

A method is versioned because the environment changes.

Providers change fields, result surfaces evolve, taxonomies become insufficient and new error modes appear. Method governance preserves comparability without pretending that one procedure remains permanently complete.

LOOP / MTH-GOV CONTINUOUS CONTROL
01DEFINE
02EXECUTE
03OBSERVE
04VALIDATE
05REVISE
06VERSION
CHANGE / 01Provider or endpoint state
CHANGE / 02Taxonomy or classification rules
CHANGE / 03Threshold or calculation logic
CHANGE / 04Error controls and review process
RULE / 05Never silently rewrite the method
MTH / PRINCIPLE 000 · TOPICALAUTHORITY.ORG

Tools produce outputs. Methods produce accountable knowledge.

A defensible research system does not ask users to trust a polished interface. It makes the question, data, procedure, evidence, uncertainty and limitations visible enough to inspect.

METHODS / MTH-000 · RESEARCH METHOD OPERATING SYSTEMQUESTION → PROCEDURE → EVIDENCE → QUALIFIED CLAIM
TAO / CONTACT · DIRECT TRANSMISSION Have an asset, domain or market position to investigate? ENTER CONTACT SYSTEM →
DIGITAL ASSET INTELLIGENCE + EXECUTION
EXECUTED BY
BB DIGITALNA AGENCIJA

Investigation, consulting and execution of digital assets, premium-domain strategies, information architecture, semantic systems, websites and agreed digital growth plans.

TOPICALAUTHORITY.ORG / SEMANTIC INTELLIGENCE SYSTEM BB DIGITALNA AGENCIJA / BB.HR