Stale product facts: audit owned surfaces

By Abiot Y. Derbie - Founder, RankEcho - Updated 2026-09-07

What is a stale product fact?

A stale fact is an inspectable owned-page assertion whose value predates and conflicts with a frozen current record under the same product, scope, and effective time. A missing fact has no value at its predeclared locator. A capture that cannot establish the value remains unverifiable.

Do not infer staleness from an answer alone. First establish the current value from a supported operator record, then capture the exact public surface, version, channel, locator, and observed value.

Which current facts are frozen before inspection?

SYNTHETIC EXAMPLE — Lumenfold Supply Co., Lumenfold TraceDesk, every .example URL, fact, surface, assertion, change, prompt, answer, and observation below are fictional; this is not RankEcho data, customer data, provider evidence, or a benchmark.

The fictional register contains six independently named facts: product name, monthly price, trial length, import formats, service region, and support hours. Each value has one effective time, authority, and synthetic evidence pointer.

Which four owned surfaces are in scope?

The fixed inventory covers exactly the homepage, pricing page, product page, and help page on lumenfold.example. Their captured versions are compared as separate records; a current statement on one page cannot silently repair a stale statement on another.

Visible copy, metadata, and JSON-LD remain separate assertion channels. The method does not inspect third-party pages, private product data, an AI provider's index, or an internal answer representation.

What does the worked synthetic assertion ledger show?

Sixteen assertions partition into nine current, five stale, one missing, and one unverifiable record. The five stale rows plus the one missing row are the six actionable inputs; I10 remains outside the correction set because its value was not established.

How does the inventory arithmetic reconcile?

The computed inventory contains 4 surfaces, 6 current facts, and 16 assertions: 9 current, 5 stale, 1 missing, and 1 unverifiable. Exactly 6 rows are actionable.

These are assertion counts, not a freshness score, provider confidence estimate, search-performance measure, or probability of an answer change.

How do page dates and structured data inform the audit?

Google documents Product fields and requires structured data to match visible page content. Those sources support checks for accurate price, availability, and other represented facts; they do not show that an AI answer system consumed the page.

A sitemap lastmod value should reflect a significant page modification and remain accurate. It is supporting publication metadata, not proof of recrawl, indexing, retrieval, citation, or answer freshness.

Which false positives should be excluded?

Exclude legitimate plan variants, tax-inclusive versus tax-exclusive prices, locale-specific regions, deliberately different support schedules, aliases defined in the current record, and captures taken before the fact's effective time.

Do not label a JavaScript widget stale when the capture cannot establish its rendered value. Preserve the locator and capture limitation as unverifiable, then use a suitable direct observation if the fact is materially important.

What deterministic validation completes Find?

A second implementation of the rules should reproduce every classification from the frozen register, surface version, locator, observed value, and effective time. All sixteen rows must be retained so the six-row correction set cannot hide excluded evidence.

Completion depends on reproducible owned-publication evidence. An impression, citation, recommendation, referral, conversion, or traffic observation is not required.

How does Find hand off to Fix and Prove?

Use the AI Visibility Audit handoff to organize the owned-surface inspection, then pass only the six identified assertion IDs to the bounded correction route. Preserve the nine current and one unverifiable rows unchanged.

After deterministic acceptance, Prove observes the same six facts in a fixed prompt-surface panel. Later current or stale answer text is descriptive evidence from that cell and checkpoint, not proof that the owned-page correction caused it.

What does RankEcho handle and what remains a companion record?

RankEcho can support an AI visibility audit and preserve bounded observations. The frozen six-fact register, sixteen-row assertion inventory, and its four-state arithmetic are a companion method; RankEcho does not currently expose this synthetic projection as a supported product metric or inspect a provider's internal freshness state.

Who maintains this guide?

Written and maintained by Abiot Y. Derbie. Published 2026-09-07; sources checked 2026-09-07; last updated 2026-09-07. The worked records are synthetic.

Recheck the operator's current fact authorities and the linked primary documentation before applying the method. Send material corrections through the contact page.

Sources reviewed

Provider eligibility and measurement claims below were checked against primary documentation. These records do not establish a universal selection formula, causation, or a guaranteed ranking, impression, recommendation, or citation.

Frequently asked questions