Content Decay Signal Auditor

Audit content pages for transparent decay-risk proxies derived from observable evidence. The Actor checks publication and modification dates, old year references, explicitly checked citation statuses, supplied search-result title drift, and supplied competitor freshness dates. It returns a bounded risk score with the exact factors that contributed to it.

Use it to prioritize editorial refresh queues, review stale articles, find outdated references, inspect cited-link health, and compare freshness evidence. The Actor does not have access to analytics or ranking history, so its score is a content-maintenance signal rather than proof of traffic or ranking decay.

Risk signals

• staleDays measures elapsed days from the latest visible or structured publication/update date to the run timestamp;
• outdatedYearReference detects visible years older than the configured current-year window;
• brokenCitationCount counts failed statuses only when outbound checks are explicitly enabled;
• serpMismatch compares the observed page title with a title in supplied structured search-result records;
• competitorFreshnessGap checks whether supplied competitor dates are more than 90 days newer than the accepted page date.

Each contribution is bounded, and the final decayRiskScore is capped at 100. Missing dates add a limited uncertainty signal. A zero score means no configured proxy crossed its threshold; it is not a guarantee that the article is current or performing well.

Input

Field	Purpose	Default / cap
pageUrls	Content pages to retrieve	empty
currentYear	Reference year for visible-year checks	2026
staleDateThresholdDays	Age threshold for stale content	365 days
checkOutboundLinks	Check visible citation/link statuses	false
maxOutboundLinks	Maximum links checked per accepted page	10 default, 25 maximum
serpResults	Supplied URL/title records for title comparison	empty
competitorPages	Supplied competitor publication/update dates	empty
maxSignalsPerPage	Bounded signal-row allowance	20 default, 100 maximum
htmlInputs	Captured records with HTML and optional checked-link results	empty
allowedDomains	Optional hostname allowlist for page retrieval	empty
maxPages	Maximum pages retrieved	1 default, 25 maximum
includeEvidence	Preserve evidence excerpts in rows	true
includeRawSnapshots	Store accepted raw page HTML	false
maxChargeUsd	Stop before the next event exceeds this amount	35 USD maximum

Example:

{
  "pageUrls": [
    "https://publisher.example/guides/queue-reliability"
  ],
  "currentYear": 2026,
  "staleDateThresholdDays": 365,
  "checkOutboundLinks": true,
  "maxOutboundLinks": 10,
  "serpResults": [
    {
      "url": "https://publisher.example/guides/queue-reliability",
      "title": "Updated queue reliability guide"
    }
  ],
  "competitorPages": [
    {"updatedDate": "2026-06-15"}
  ],
  "allowedDomains": ["publisher.example"],
  "maxChargeUsd": 5
}

Outbound checks are off by default to keep the run bounded and intentional. The Store default has no page source and returns a nonempty source_required diagnostic.

Dataset output

Each row includes common source provenance plus:

• pageUrl, publishedDate, and updatedDate from accepted HTML;
• staleDays as an integer when a valid date exists, otherwise null;
• outdatedYearReference as a boolean;
• brokenCitationCount for the explicitly checked link set;
• serpMismatch and competitorFreshnessGap from supplied comparison records;
• decayRiskScore from 0 to 100;
• refreshRecommendation listing the contributing review actions;
• issueCode, evidenceSnippet, httpStatus, and scannedAt for auditability.

The recommendation identifies what to review, not how to rewrite the page. Validate current analytics, search intent, subject-matter accuracy, conversion needs, and editorial ownership before scheduling a refresh.

Reports

When includeReport is enabled, key-value storage receives a Markdown summary, CSV records, issue JSON, bounded evidence excerpts, and run metadata. Raw accepted HTML is stored only when explicitly enabled.

Pay-per-event pricing

Event	Price (USD)	When charged
actor-start	0.0350	after successful startup
content-page-analyzed	0.0115	once for each accepted content page
decay-signal-detected	0.0115	before a paid decay signal is recorded
recommendation-generated	0.0115	before each paid recommendation row
executive-report-generated	0.4000	before report artifacts are stored

Apify platform usage pass-through is off. Higher-tier Store discounts are reserved in the pricing policy. maxChargeUsd is checked before every event; rejected, partial, or over-budget charges stop the related dataset or report write.

Safety and interpretation

Only HTTP and HTTPS page or citation URLs are accepted. Loopback and private IPv4 targets are rejected, optional hostname allowlists narrow primary-page retrieval, and page, link, response-size, and timeout caps remain active. Citation checks occur only after explicit opt-in and never expand into general crawling. Access-denied and human-verification responses are recorded as diagnostics.

Dates can be missing, malformed, or editorially misleading. Search-result and competitor comparisons are only as current as the supplied records. Treat the output as a documented review queue and retain human editorial judgment for prioritization and publication decisions.