XEO ontology — 177 terms

Assistants that answer with web search and citations (the AEO surface).

subclass of

xeo:Engine

A named market identity that engines may mention, cite, or recommend. Distinct from Website: a brand can be visible on the whole web, not only its own origin.

closeMatch

schema:Brand

subclass of

xeo:ConversionPoint

A body of work inside a Project against one Topic: the URLs and Terms it covers, and the single Objective it is measured by. Recommendations are issuedFor a campaign, and it is the unit an operator plans a cycle around. Soft-deleted like its Project.

A lever or strength a party can bring to bear (content, technical, PR, brand, ...). The unit of the scope gate: a Tactic is in reach only when every Capability it requires is approved. Specific capabilities are instances, not subclasses.

Multi-turn agents and voice surfaces acting on a user's behalf.

subclass of

xeo:Engine

A submission/click/unlock that happened at a time. Immutable once recorded.

subclass of

prov:Activity

A durable on-page mechanism that turns attention into a pipeline event.

One crawl run over one Website: time-bounded, agent-attributed, immutable once finished.

subclass of

prov:Activity

A named, reusable check or framework that reads observations and reveals gaps (e.g. a waterfall-position assessment). The framework is durable; each finding it produces is a Gap.

A system that mediates between a person's question and web content. The taxonomy below is operational, not doctrinal — subclass as engines evolve.

subclass of

prov:SoftwareAgent

A test of a tactic that ran over a time window: immutable once concluded, method-attributed. The precedent record the decision layer learns from.

subclass of

prov:Activity

A diagnosed deficit between current and desired visibility. First-order, but a claim: asserted by a Diagnostic from evidence, never observed directly. Its kind (gapType) is a slotted-in skos:Concept — e.g. a Citation-Waterfall stage — not a subclass baked into the ontology.

Content exchanged for contact information.

subclass of

xeo:ConversionPoint

Search-integrated generated answers — AI overviews and AI modes (the GEO surface).

subclass of

xeo:Engine

A category of tactic in a controlled vocabulary you bring. A skos:Concept, so a whole intervention taxonomy slots in without touching the ontology.

subclass of

skos:Concept

subclass of

xeo:ConversionPoint

One captured response from one engine/model for one Prompt at a moment: time-bounded, source-attributed, immutable once captured. The visibility sibling of SerpSnapshot.

subclass of

prov:Entity

One measurable target of an Objective: metric X should reach targetValue, read as absolute or as a percent lift from a captured baseline. Named a target and never a key result: xeo:Outcome is 'the measured result an experiment produced', an observation, and a goal must not share a name with a reading. Each target is its own node so the experiment loop can hang observations off it; a target dropped on edit is stamped deletedAt, never destroyed, because its id is what measured history is anchored to.

What a Campaign intends to move: an operator-written statement in prose (carried verbatim — the recommendation agent reads it as the goal), a resultsWindowDays every target is measured over, and one or more MetricTargets of which exactly one isPrimary. An intent set at creation. A campaign has one reporting horizon, distinct from the Project's work window: a two-month engagement can carry a six-month results window.

The measured result an experiment produced (lift, delta, significance). The measurement is an observation; any claim that the tactic CAUSED it is an interpretation, carried on the Recommendation, not asserted here.

subclass of

prov:Entity

A reusable, discipline-scoped bundle of Diagnostics, Tactics, and Practices — an 'SEO playbook', an 'AEO playbook', a 'local playbook'. Durable and named, like a Diagnostic. Its scope is the DISCIPLINE, not the account: a Playbook is authored once (by the community or an agency) and reused across every client. What varies per client or campaign is which Playbooks a tenancy activates and which Practices it overrides — never the Playbook itself.

A standing rule that governs how recommendations are made under stated conditions. Best-practices, SOPs, and lessons-learned are all Practices, distinguished by source, not by class. A Practice does one of three things (practiceKind): veto a candidate (guardrail), reorder it (preference), or attach a caveat (guidance). Revisable and versioned (supersedes); scoped to the concepts it appliesToConcept; audited on the recommendations it fired on (appliedPractice).

The contractual container for a body of work on one Website. Declares its engagementType (retainer or one-off), an interventionQuota (a scope of work with no ceiling is not a scope), a cadence the quota resets over (retainer) or a durationDays it runs for (one-off), and the Interventions and Tactics inScope under the contract. Never destroyed: its campaigns, objectives and targets are the substrate the end-of-campaign report and the experiment loop read, so it is stamped deletedAt and filtered, not deleted.

closeMatch

schema:Project

A tracked natural-language question posed to engines. Distinct from Term: a prompt is what a person asks; the engine's retrieval fans out into Terms.

The labeled claim the whole chain exists to produce: deploy this Tactic to close this Gap, supported by these Experiments, at a given priority. An interpretation — carries method, confidence, and who/what issued it.

Classic ranked-results retrieval (the SEO surface).

subclass of

xeo:Engine

Why the person is asking. Modeled as named individuals, not free strings.

grounded in

Broder (2002), 'A taxonomy of web search', as extended in common practice

The atomic ground-truth metric row for one (websiteId, address, term, date [, device, country]) tuple. Never instantiated as graph relationships; the graph carries windowed ranksFor rollups whose datasetUri points back to the authoritative table.

closeMatch

schema:Observation

grounded in

Google Search Console Search Analytics API

canonical store

bigquery

One observation of a search results page for a Term at a moment, for a stated engine, locale, and device.

subclass of

prov:Entity

A durable, named approach that can address gaps. Categorized by a slotted-in intervention taxonomy (inIntervention), scored on slotted-in dimensions (assessedAs -> TacticAssessment), and gated by the Capabilities it requires.

A scored estimate of a tactic on ONE dimension (onDimension), carrying score, method, and confidence. The dimension set (impact / risk / time-to-value, or ICE, or your own) and how dimensions are weighted are a strategy-layer concern; this class only records that an assessment was made, on what dimension, and by what method.

A search query string, unique by name. Data sources are attributed on observations that mention the term, never as subtypes of the term.

A named grouping of Terms produced by a clustering or labeling process. Site-scoped. Assignments carry method, model, and confidence.

subclass of

skos:Concept

A resource address within a Website. Durable identity; the state of the resource at a time is recorded by Fetch observations.

closeMatch

schema:WebPage

grounded in

WHATWG URL Standard

A site under observation, identified by its origin. The tenancy scope for URLs, crawls, and topics.

closeMatch

schema:WebSite

grounded in

WHATWG URL Standard (origin)

The topic cluster this campaign works. Topics are site-scoped derivations in core; a campaign binds to one.

domain

xeo:Campaign

range

xeo:Topic

A tenancy context puts a Playbook into effect for a body of work. Domain left open: model your own tenancy node (agency, client, engagement, campaign) and point it here — this module stays agnostic about the account hierarchy, which is expressed through appliesAtScope.

range

xeo:Playbook

A candidate tactic for closing this gap — a claim, made before the scope gate is applied.

domain

xeo:Gap

range

xeo:Tactic

Audit edge: this Practice was evaluated against this recommendation and fired. Carries method and confidence — the application is a computed derivation, not itself a norm. The transparent record of why a candidate was blocked, demoted, or flagged.

domain

xeo:Recommendation

range

xeo:Practice

The tenancy level at which a Practice (or activation or override) applies, as a slotted-in concept. A recommended hierarchy is default < agency < client < engagement < campaign < user. Domain left open so it can annotate any strategy-layer assertion. Resolution is most-specific-wins: a client-level Practice overrides an agency default, which overrides the shipped default. This is the 'customization' axis — not a lens, a specificity ladder.

range

skos:Concept

Binds a Practice to the slotted-in concepts it applies to — a page role (hasRole), an Intervention, a gapType, a Capability. Bring your own schemes; the Practice fires only where a candidate carries a matching concept.

domain

xeo:Practice

range

skos:Concept

Links a tactic to a per-dimension scored estimate.

domain

xeo:Tactic

range

xeo:TacticAssessment

Priority as a slotted-in concept — bring your own scheme (P0/P1/P2, now/next/later, ...).

domain

xeo:Recommendation

range

skos:Concept

The claim that this conversion traces to engine visibility captured in this response episode. Multi-engine attribution is inherently inferential: this edge MUST carry method and confidence, and honest implementations will leave it sparse.

domain

xeo:ConversionEvent

range

xeo:LLMResponse

The party that authored a Practice (or an activation or override). Range left open — a person, an agency, the system, or a prov:Agent in your own graph. One half of the customization axis: with appliesAtScope, it decides whose rule wins.

domain

xeo:Practice

domain

xeo:ConversionEvent

range

xeo:ConversionPoint

Topic childOf Topic: the child points to its broader parent, matching skos:broader.

domain

xeo:Topic

range

xeo:Topic

subproperty of

skos:broader

The response grounded on this URL, per the engine's own citation annotations. Carries citationRank where the engine orders its sources. The visibility sibling of hasResult.

domain

xeo:LLMResponse

range

xeo:URL

grounded in

Engine response citation annotations (per-provider API contract)

The gap this recommendation aims to close.

domain

xeo:Recommendation

range

xeo:Gap

The Practice governs recommendations matching its conditions — vetoing, reordering, or caveating per its practiceKind. Which recommendations 'match' is set by appliesToConcept.

domain

xeo:Practice

range

xeo:Recommendation

A term inside the campaign's cluster — the queries its objective is measured against.

domain

xeo:Campaign

range

xeo:Term

A page inside the campaign's cluster. May carry a priority flag marking the pages the campaign is really about.

domain

xeo:Campaign

range

xeo:URL

How long executing the recommendation is expected to take, as a slotted-in concept (hours, days, weeks, a quarter — see xeo-schemes). A categorical horizon for planning a cycle, not a scored dimension: impact, risk and time-to-value belong on a TacticAssessment via onDimension.

domain

xeo:Recommendation

range

skos:Concept

The observed facts a gap was derived from — a Prompt's visibilityFor rollup, a URL's ranksFor, a SerpSnapshot. The seam tying the interpretation tier back down to the substrate. Range left open: any observation-bearing term.

domain

xeo:Gap

subproperty of

prov:wasDerivedFrom

Who does the work, as a slotted-in concept: ours (spends the project's intervention allowance) or advised (the client executes; not agency capacity). The half of the scope gate that says whose hours a recommendation costs.

domain

xeo:Recommendation

range

skos:Concept

A retrieval query the engine reported issuing while answering this prompt. Fan-out queries ARE Terms — the join between the xEO layer and classic rank data. Carries the reporting engine and capture episode.

domain

xeo:Prompt

range

xeo:Term

grounded in

Engine-reported retrieval queries (per-provider API contract)

An HTTP fetch of a URL within a crawl. In the property graph this relationship carries statusCode, contentType, title, metaDescription, robotsMeta, contentHash, fetchedAt.

domain

xeo:Crawl

range

xeo:URL

grounded in

RFC 9110 (HTTP Semantics)

The kind of gap, as a slotted-in concept (e.g. a Citation-Waterfall stage). Bring your own scheme.

domain

xeo:Gap

range

skos:Concept

domain

xeo:Project

range

xeo:Campaign

domain

xeo:URL

range

xeo:URL

grounded in

RFC 6596 (The Canonical Link Relation)

Observed presence of a conversion point on a page (crawl-attributed, bitemporal like linksTo).

domain

xeo:URL

range

xeo:ConversionPoint

domain

xeo:Website

range

xeo:Crawl

Exactly one per campaign.

domain

xeo:Campaign

range

xeo:Objective

The projects contracted against this website. Tenancy rides on websiteId, as everywhere.

domain

xeo:Website

range

xeo:Project

domain

xeo:Prompt

range

xeo:LLMResponse

A result observed on the SERP. Carries position and resultType.

domain

xeo:SerpSnapshot

range

xeo:URL

The structural role a URL plays on its site, as a slotted-in concept. A recommended starter scheme is coarse and purpose-based: brand-core (home, about), commercial (services, products, sales), content (articles, marketing), infrastructure (pagination, archives, tag and category indexes). Assignment is a DERIVATION from graph and crawl structure — inbound link_position, index-vs-leaf link patterns, template similarity — not URL-string matching; it carries method and confidence, and 'unknown' is a valid, conservative state. Parallels core's inTopic. Norms bind to these roles via appliesToConcept, so 'never redirect an infrastructure page for ranking' is a rule over a role, not a growing list of regexes.

domain

xeo:URL

range

skos:Concept

domain

xeo:Term

range

xeo:SerpSnapshot

The objective's measurable targets. Non-empty; exactly one target isPrimary — what the recommendation agent optimises for. The rest are the secondary metrics an impact-radius check widens to.

domain

xeo:Objective

range

xeo:MetricTarget

A Diagnostic, Tactic, or Practice bundled into this Playbook. Range left open across the three decision/norm terms it may hold.

domain

xeo:Playbook

The gap the experiment was run against. The precedent-similarity key: 'has this tactic worked on a gap like mine?' is a match on gapType through this edge.

domain

xeo:Experiment

range

xeo:Gap

The intervention category this tactic belongs to. Parallels core's inTopic.

domain

xeo:Tactic

range

xeo:Intervention

What the contract covers: an Intervention (a whole category of work, including tactics added to the catalogue later) or a specific Tactic within one. Range left open across those two terms. The contractual half of the scope gate — requiresCapability + approved asks whether the lever exists; this asks whether it was bought.

domain

xeo:Project

Process-attributed topic assignment. Carries method, model, confidence, assignedAt.

domain

xeo:Term

range

xeo:Topic

closeMatch

schema:about

The campaign a recommendation was issued into. Its execution lifecycle (xeo:status), effortEstimate and execution live here: a recommendation is a decision-tier call until it lands in a campaign, and delivery state from then on.

domain

xeo:Recommendation

range

xeo:Campaign

An observed hyperlink. Carries anchorText, rel, firstSeenAt, lastSeenAt, lastSeenCrawlId.

domain

xeo:URL

range

xeo:URL

grounded in

WHATWG HTML §4.5.1 (the a element)

The unit magnitude is read in, as a slotted-in concept (impressions, clicks, referring_domains, words, ... — see xeo-schemes). A unit that names its own provenance, like sessions_estimated, is how a derived figure is kept from rendering as a measured one.

domain

xeo:Gap

range

skos:Concept

The target an experiment's Outcome is read against. Carries a role — primary or secondary — matching the target's standing in its objective, and a basis saying which pages were measured (the ones the recommendation applied to, the gap's evidence, or the pages actually changed).

domain

xeo:Experiment

range

xeo:MetricTarget

The brand's name appears in the response text. Carries mentionRank (order of first appearance) — high mention counts at poor ranks are the 'named among twelve, championed never' pattern. Extraction method (string match, NER model) must be recorded; model-extracted mentions are derivations, not observations.

domain

xeo:LLMResponse

range

xeo:Brand

The dimension being scored, as a slotted-in concept (impact, risk, time-to-value, ICE, RICE, ...). Bring your own dimension scheme.

domain

xeo:TacticAssessment

range

skos:Concept

Which engine (and, via model/modelVersion properties, which model) produced this response.

domain

xeo:LLMResponse

range

xeo:Engine

subproperty of

prov:wasAttributedTo

domain

xeo:URL

range

xeo:Website

The measured result this experiment generated.

domain

xeo:Experiment

range

xeo:Outcome

subproperty of

prov:generated

Windowed rollup projection of SearchPerformanceFacts for a (page, query) pair. Carries clicks, impressions, ctr, avgPosition, periodStart, periodEnd, source, datasetUri, syncedAt.

domain

xeo:URL

range

xeo:Term

grounded in

Google Search Console Search Analytics API

canonical store

neo4j

The tactic this recommendation puts forward.

domain

xeo:Recommendation

range

xeo:Tactic

domain

xeo:URL

range

xeo:URL

grounded in

RFC 9110 §15.4 (Redirection 3xx)

A capability that must be approved for this tactic to be in reach. The scope-gate edge: a tactic whose required capabilities are all approved is deployable; otherwise it stays advisory.

domain

xeo:Tactic

range

xeo:Capability

A diagnostic asserts this gap. Carries the assessment method.

domain

xeo:Diagnostic

range

xeo:Gap

Topics are scoped to a Website rather than a free-standing ConceptScheme; skos:inScheme is the analogous grouping.

domain

xeo:Topic

range

xeo:Website

seeAlso

skos:inScheme

This Practice replaces an earlier version. Editing a rule mints a new Practice that supersedes the old (retired to status 'deprecated', never deleted) — the version chain that makes the norms layer revisable and auditable.

domain

xeo:Practice

range

xeo:Practice

The precedent experiments backing this recommendation — the earned trust behind the pick.

domain

xeo:Recommendation

range

xeo:Experiment

subproperty of

prov:wasDerivedFrom

A URL or Prompt is aimed at this intent. A strategist's or model's claim — carries method and confidence.

range

xeo:SearchIntent

The tactic this experiment put to the test.

domain

xeo:Experiment

range

xeo:Tactic

This tenant tracks this prompt. Tenancy is data, not schema.

domain

xeo:Website

range

xeo:Prompt

The operator's disposition on a diagnosed gap, as a slotted-in concept (attack, fix_first, defend, hold, ignore — see xeo-schemes). Distinct from gapType, which says what KIND of deficit it is; the verdict says what we intend to do about it. Carried by one finding per topic per analysis: it is a property of the topic's scoreboard row, not of every finding beneath it.

domain

xeo:Gap

range

skos:Concept

Windowed rollup of response-level observations for a (prompt, brand) pair: mentionRate, citationRate, responses, periodStart, periodEnd, source, datasetUri. Mirrors the ranksFor rollup pattern — raw responses are the authoritative record; this is the traversal-friendly projection.

domain

xeo:Prompt

range

xeo:Brand

canonical store

neo4j

The serialized absolute URL is a schema:url value.

domain

xeo:URL

range

xsd:string

subproperty of

schema:url

grounded in

WHATWG URL Standard (serialized absolute URL)

range

xsd:string

grounded in

WHATWG HTML §4.5.1

Whether this capability is approved for the engagement — the live state of the scope gate. Tenancy-scoped like everything else.

domain

xeo:Capability

range

xsd:boolean

When a derivation edge was written (e.g. a topic assignment). Carried on inTopic alongside method, model and confidence.

range

xsd:dateTime

range

xsd:decimal

grounded in

GSC Search Analytics API

The metric's value when the target was set — the anchor 'how far beyond target' is measured from. An observation captured onto an interpretation; baselineCapturedAt says when, because it is not the same instant as the target's creation.

domain

xeo:MetricTarget

range

xsd:decimal

domain

xeo:MetricTarget

range

xsd:dateTime

'monthly' or 'quarterly' — the period a retainer's quota resets over. Set for retainer, absent for one-off.

domain

xeo:Project

range

xsd:string

The response record was generated at capture time.

domain

xeo:LLMResponse

range

xsd:dateTime

subproperty of

prov:generatedAtTime

1-based order of a cited source where the engine orders its citations. Carried on cites; absent where the engine reports an unordered set.

range

xsd:integer

Share of captured responses in the window that cite the brand's owned URLs. Cited and mentioned diverge — a page can ground an answer that names someone else.

range

xsd:decimal

range

xsd:integer

grounded in

GSC Search Analytics API

When the experiment concluded and its outcome became immutable.

domain

xeo:Experiment

range

xsd:dateTime

subproperty of

prov:endedAtTime

0-1 confidence attached to a derivation edge.

range

xsd:decimal

Identity-key pointer to the CRM object (form id, contact id, deal id) in whatever system runs the pipeline. The join contract to the post-click world — the graph never mirrors the CRM's schema.

range

xsd:string

canonical store

crm

Clickthrough rate over the rollup window. Carried on ranksFor.

range

xsd:decimal

grounded in

GSC Search Analytics API

Pointer to the authoritative table a windowed rollup was projected from (e.g. a BigQuery table URI). The provenance contract for ranksFor and visibilityFor: the graph keeps the traversable projection, this says where the ground truth lives.

range

xsd:string

Soft-delete stamp. XEO never destroys a node something else was derived from — xeo-strategy states the rule for a Practice (destroying it destroys the record of what it vetoed) and it holds here for a Project, a Campaign, or a MetricTarget that measured history hangs off. Absent on a live node; domain left open so any revisable term can carry it.

range

xsd:dateTime

'up' or 'down' — which way is better for this metric. DERIVED from metric, never client input: position and rank are 1-based, so a shared '+20%' would silently ask for a rank 20% worse.

domain

xeo:MetricTarget

range

xsd:string

Fixed run length of a one-off project. Set for oneoff, absent for retainer.

domain

xeo:Project

range

xsd:integer

domain

xeo:Term

range

xsd:string

Pointer (collection + id) to the vector in the external store. The geometry lives where ANN search happens; the graph keeps the claim and its provenance.

domain

xeo:Term

range

xsd:string

canonical store

vector-store

'retainer' or 'oneoff' — the discriminator for how scope is expressed. A retainer caps interventions per cadence period; a one-off caps them over durationDays. Sending both cadence and durationDays is a contradiction and is rejected at the edge.

domain

xeo:Project

range

xsd:string

When the fetch observation was made. Carried on fetched.

range

xsd:dateTime

When an observed edge (a link, typically) was first seen. Half of the bitemporal pair that makes link liveness answerable.

range

xsd:dateTime

Optional decay window for a Practice whose source is 'claimed': how long the assertion is trusted absent supporting Outcomes. Encodes the reality that today's answer-engine tactics may not survive the year — a claim nobody's experiments confirm ages out on its own.

domain

xeo:Practice

range

xsd:duration

range

xsd:integer

grounded in

GSC Search Analytics API

Ceiling on interventions this project may deploy per period (retainer) or in total (one-off). Always present. Recommendations in an open status — pending, in_progress, blocked — consume it; advised work does not.

domain

xeo:Project

range

xsd:integer

Exactly one target per objective. The one the agent optimises for.

domain

xeo:MetricTarget

range

xsd:boolean

When an observed edge was most recently seen. With lastSeenCrawlId this is how a link is known to be live rather than merely once-observed.

range

xsd:dateTime

The measured change the experiment recorded on its target metric.

domain

xeo:Outcome

range

xsd:decimal

How big the gap is, where the finding IS a size. Required together with magnitudeUnit and never alone: a magnitude with no unit is an uninterpretable number, a unit with no magnitude is noise.

domain

xeo:Gap

range

xsd:decimal

1-based order of the brand's first appearance in the response text. Carried on mentions edges.

range

xsd:integer

Share of captured responses in the window that mention the brand. Carried on visibilityFor rollups.

range

xsd:decimal

The meta description observed on a fetch. Carried on the fetched edge.

range

xsd:string

grounded in

WHATWG HTML (meta name=description)

How an interpretation or derivation was produced (model, rule, human analyst). Expected on reveals, assessedAs, and every Recommendation.

range

xsd:string

Which metric the target is set on — clicks, impressions, ctr, position, rank_absolute, etv (see xeo-schemes objectiveMetric). The value vocabulary is the consumer's; the reference scheme records each metric's direction and unit.

domain

xeo:MetricTarget

range

xsd:string

Model identifier + version as reported at capture time. Without this the episode is not comparable across time.

domain

xeo:LLMResponse

range

xsd:string

The model build reported at capture time, where the engine distinguishes it from the model name. Without model + modelVersion an episode is not comparable across time.

range

xsd:string

The identifying name of a Term, Topic, Brand, or Capability. Unique within its scope; see schema.cypher for the per-class identity keys.

range

xsd:string

domain

xeo:ConversionEvent

range

xsd:dateTime

domain

xeo:Website

range

xsd:string

grounded in

WHATWG URL Standard (origin serialization)

Inclusive last day of a rollup window. Carried on ranksFor and visibilityFor.

range

xsd:date

Inclusive first day of a rollup window. Carried on ranksFor and visibilityFor; never pool across windows silently.

range

xsd:date

1-based rank of a result within a SERP snapshot. Carried on hasResult. Distinct from avgPosition, which is a windowed rollup, not a single observation.

range

xsd:integer

How the Practice acts: 'guardrail' (veto a matching candidate), 'preference' (reorder it), or 'guidance' (attach a caveat, change nothing). The mechanism, kept as a small controlled string rather than a subclass.

domain

xeo:Practice

range

xsd:string

Carried on an appliedPractice edge: 'vetoed' (a guardrail removed the candidate), 'reranked' (a preference moved it), or 'caveated' (guidance flagged it and changed nothing). One value per practiceKind. Domain left open because it rides on the edge, not on a node.

range

xsd:string

Why the rule exists — surfaced to a user when the Practice fires, so a veto or demotion is explained, never silent.

domain

xeo:Practice

range

xsd:string

The rel attribute of an observed hyperlink (nofollow, sponsored, ugc, ...). Carried on linksTo.

range

xsd:string

grounded in

WHATWG HTML (link types)

How many captured responses the rollup window covers — the denominator behind mentionRate and citationRate. Carried on visibilityFor; a rate without it cannot be weighted.

range

xsd:integer

'significant', 'no_change' or 'inconclusive'. Answers DID THIS MOVE; lift's sign answers which way. A regression is significant — the value never smuggles a direction in.

domain

xeo:Outcome

range

xsd:string

Why an outcome was inconclusive, from a small controlled set (window_incomplete, co_intervention, confounded_external, baseline_unavailable, ...). The two confounder reasons differ on purpose: a co-intervention keeps its lift and shares attribution; an external confounder records none.

domain

xeo:Outcome

range

xsd:string

The measurement window every target in this objective is read over. One horizon per campaign; per-target windows are deliberately not modelled.

domain

xeo:Objective

range

xsd:integer

The kind of SERP result observed (organic, featured snippet, local pack, ...). Carried on hasResult; the vocabulary is per-engine and belongs to the consumer.

range

xsd:string

The value assigned on the assessed dimension. Scale is defined by the dimension scheme, not fixed here.

domain

xeo:TacticAssessment

range

xsd:decimal

Provenance of an assertion. The value vocabulary is subject-specific. On a Practice: 'authored' (a human SOP or best-practice — high standing, durable), 'learned' (earned from the experiment ledger — confidence ratchets with Outcomes), or 'claimed' (an unproven assertion, e.g. a current answer-engine tactic — ships low-confidence with a halfLife and decays unless experiments promote it); this lets one class hold a firm SOP and a fragile claim without conflating their trust. On a ranksFor or visibilityFor rollup: the data source the window was built from. DOMAIN LEFT OPEN deliberately — rdfs:domain is an entailment rule, not a constraint, so naming xeo:Practice here would silently type every rollup edge carrying a source as a Practice.

range

xsd:string

The rule in plain language, as a human authored or confirmed it — e.g. 'Never 301 a paginated archive URL for ranking purposes.' The canonical human-facing form; the structural fields (practiceKind, appliesToConcept, constrains) make it executable.

domain

xeo:Practice

range

xsd:string

Lifecycle state of a revisable term. The value vocabulary is subject-specific. On a Practice: 'proposed', 'active', or 'deprecated' — a proposed Practice is visible but does not fire, and superseding one deprecates it. DOMAIN LEFT OPEN deliberately, so another revisable term (the execution lifecycle of a Recommendation, say) can carry a lifecycle without being entailed a Practice.

range

xsd:string

range

xsd:integer

grounded in

RFC 9110 §15 (Status Codes)

'absolute' (targetValue is a value of the metric) or 'percent' (a relative change from baseline in the metric's better direction).

domain

xeo:MetricTarget

range

xsd:string

The value the metric should reach, interpreted per targetKind. Bounded per metric and kind at the edge — a percentage, a count, a rank and a currency do not share a range.

domain

xeo:MetricTarget

range

xsd:decimal

The natural-language body of a Prompt as posed to the engine. Part of a Prompt's identity together with websiteId.

range

xsd:string

The document title observed on a fetch. Carried on the fetched edge, not on the URL: the address is durable, the title is what one crawl saw.

range

xsd:string

grounded in

WHATWG HTML (the title element)

Whether the engine could search the web for this response. Citation behavior differs sharply with this flag; never pool across it silently.

domain

xeo:LLMResponse

range

xsd:boolean

The tenancy key. Scope lives in this property on every scoped node and edge, never in a dynamic label — the identity half of 'tenancy is data, not schema'.

range

xsd:string

Where this term's instances authoritatively live: neo4j, bigquery, vector-store, crm. The ontology is storage-agnostic; identity keys (websiteId, address, term name, date) join across stores.

One of: entity, episode, observation, derivation, interpretation, norm. Types every term by the kind of claim it makes. Core terms stay at entity/episode/observation/derivation; extension modules may introduce interpretation- or norm-layer terms and must mark them.

Names the XEO property whose skos:Concept range (or controlled-string value) this scheme is a reference filling for — xeo:gapType, xeo:atPriority, xeo:inIntervention, xeo:status. A consumer reading a slot property can follow this to a published default; nothing in the ontology requires it.

The primary standard or API contract that defines this term's meaning.

type

xeo:SearchIntent · owl:NamedIndividual

type

xeo:SearchIntent · owl:NamedIndividual

type

xeo:SearchIntent · owl:NamedIndividual

type

xeo:SearchIntent · owl:NamedIndividual

Optional crosswalk from XEO terms to foundational web vocabularies. Alignments, not imports: XEO is self-contained without them.

seeAlso

schema: · prov: · skos:

creator

Inbound Found

version

0.5.0

The top stratum of the core sample: where an observed condition becomes a diagnosed Gap, a Gap becomes candidate Tactics, the capability gate filters them to what is in reach, prior Experiments say which tactic has earned trust, and a Recommendation is issued — labeled, honestly, as an interpretation. Ships CLASSES, never VALUES: the dimensions, weights, and governing rules live one layer higher, in xeo-strategy.

creator

Inbound Found

version

0.5.0

The container work is delivered in: a Project (the contract — type, quota, scope), its Campaigns (a body of work against one Topic), the Objective each campaign is measured by, and the MetricTargets an Experiment is read against. Lifted from a production implementation; fills the layer the decision and norms tiers left open.

creator

Inbound Found

version

0.5.0

DRAFT v0. The seam between web-engine visibility and the pipeline: intents, conversion points, conversion events, and identity-key pointers into whatever CRM you run.

creator

Inbound Found

version

0.5.0

SKOS concept schemes filling the slots the XEO ontology leaves open — gap types, verdicts, priorities, the intervention catalogue, lifecycles, metrics. Published reference values, not part of the ontology's class vocabulary; adopt, extend, or replace.

seeAlso

https://xeoontology.org/schemes.json

creator

Inbound Found

version

0.5.0

The layer xeo-decision deliberately left open: the standing rules — best practices, SOPs, lessons learned — that veto, reorder, or caveat a Recommendation before it reaches a human, the reusable Playbooks those rules travel in, and the tenancy axis that decides whose rules win. Introduces the norm stratum: a claim about what SHOULD govern a choice, as opposed to what is or is inferred to be.

creator

Inbound Found

version

0.5.0

Extends XEO Core with the terms needed once the engine stops being a results page and starts being an answer: engines, prompts, captured responses, citations, mentions, and the visibility rollups over them.

creator

Inbound Found

version

0.5.0