{"schema_version":1,"research_id":"eoa_inverse_innovation_exp03_external48_20260801","source_assessment_id":"eoa_inverse_innovation_exp03_opportunity320_20260801","cell_id":"layer_decay_and_expiration_management__communication_media_studies","selection_stratum":"DEPLOYABLE_PRIORITY","search_queries":["site:gov.uk guidance content lifecycle withdrawn pages redirects publishing official","news article expiration date archive superseded content newsroom policy","study users mistake old news articles as current date context social media","official newsroom corrections archive outdated article policy search demotion","experiment old news article publication date perceived current social media study","timestamp disclaimer old news experiment","Do Journalists Update Retracted Science News Valinciute Halffman findings","site:about.fb.com notification old news article sharing context 90 days"],"sources":[{"source_id":"S1","title":"Do Journalists Update Retracted Science News?","publisher":"Journalism Practice","url":"https://www.tandfonline.com/doi/full/10.1080/17512786.2025.2540460","source_class":"PRIMARY_RESEARCH","publication_date":"2025-09-24","accessed_at":"2026-08-02","claims_supported":["The study examined 945 news articles mentioning subsequently retracted or formally questioned COVID-19 papers.","Only 8% of sampled articles had an overt post-retraction update; 4% had a general timestamp disclaimer.","Updates rarely explained plainly how the retraction affected the article's claims.","The study is a bounded science-news sample and does not establish prevalence across all publishing or measure audience currentness errors."]},{"source_id":"S2","title":"When old stories go viral","publisher":"Columbia Journalism Review","url":"https://www.cjr.org/news_literacy/old_stories_going_viral.php","source_class":"AUTHORITATIVE_SECONDARY","publication_date":"2014-02-17","accessed_at":"2026-08-02","claims_supported":["Documents several cases in which recirculated reporting was mistaken for current news.","Describes an old weather story being reshared and even promoted by its publisher after traffic increased.","Provides case evidence, not a representative prevalence estimate or controlled causal study."]},{"source_id":"S3","title":"Providing People With Additional Context About Content They Share","publisher":"Meta","url":"https://about.fb.com/news/2020/06/more-context-for-news-articles-and-other-content/","source_class":"COMMERCIAL_FIRST_PARTY","publication_date":"2020-06-25","accessed_at":"2026-08-02","claims_supported":["Facebook deployed a notification before users shared news articles older than 90 days while still allowing sharing.","Meta reported that publishers had expressed concern about older stories circulating as current news.","Meta referred to internal research on timeliness but did not publish methods or comparative outcome estimates on this page."]},{"source_id":"S4","title":"Retire outdated content","publisher":"Government Digital Service, GOV.UK","url":"https://guidance.publishing.service.gov.uk/writing-to-gov-uk-standards/plan-manage-content/retire-content/","source_class":"OFFICIAL_GUIDANCE","publication_date":"n.d.","accessed_at":"2026-08-02","claims_supported":["GOV.UK supports withdrawal that preserves the URL, displays an explanatory banner, and removes content from internal search and other discovery surfaces.","History mode labels content from a previous government and reduces its internal-search prominence.","Withdrawn and unpublished states are reversible; ordinary unpublishing is exceptional.","The guidance distinguishes contextual judgment from automatic retirement based solely on age."]},{"source_id":"S5","title":"Withdraw or unpublish standard content types","publisher":"Government Digital Service, GOV.UK","url":"https://guidance.publishing.service.gov.uk/publish-update-retire-content/standard-content-types/withdraw-unpublish-standard/","source_class":"OFFICIAL_GUIDANCE","publication_date":"n.d.","accessed_at":"2026-08-02","claims_supported":["Managing-editor permission is required for withdrawal or unpublishing.","Public withdrawal explanations can link to replacement content.","The documented workflow supports unwithdrawal, republication, redirects, and retention of unpublished records in the publishing system.","The guidance recognizes that a government web archive may retain snapshots."]},{"source_id":"S6","title":"HHS Website Content Lifecycle Management and Archive Guidance","publisher":"U.S. Department of Health and Human Services","url":"https://www.hhs.gov/digital/policies/web/content-lifecycle-management/index.html","source_class":"OFFICIAL_GUIDANCE","publication_date":"2024-02-15","accessed_at":"2026-08-02","claims_supported":["HHS requires annual review of public-site content and displays ownership and last-reviewed metadata.","Reviewed content is classified as retain, revise, remove, or archive.","Superseded sites and historically useful pages are archived, with links between successor and archived predecessor where applicable.","HHS identifies records-management responsibility and uses regular archive captures."]},{"source_id":"S7","title":"How old is too old? I’m 77 and I don’t know yet. But I will when I get there","publisher":"The Guardian","url":"https://www.theguardian.com/commentisfree/article/2024/may/04/old-age-social-media-young-people","source_class":"COMMERCIAL_FIRST_PARTY","publication_date":"2024-05-04","accessed_at":"2026-08-02","claims_supported":["An accessible Guardian article displays its age warning prominently above and below the headline.","This demonstrates deployed newsroom age labeling but not supersession classification, demotion, or effectiveness."]},{"source_id":"S8","title":"Timing matters when correcting fake news","publisher":"Proceedings of the National Academy of Sciences","url":"https://doi.org/10.1073/pnas.2020043118","source_class":"PRIMARY_RESEARCH","publication_date":"2021-01-25","accessed_at":"2026-08-02","claims_supported":["Two experiments with 2,683 participants found that correction timing affected later headline truth discernment.","Post-exposure correction produced larger one-week improvements than simultaneous or pre-exposure labels.","The experiment concerned true/false labels, not article supersession or distribution demotion, so it is adjacent rather than direct effectiveness evidence."]}],"problem_evidence":{"support":"MODERATE","rationale":"A primary bounded study found that overt updates were rare among news stories tied to later-retracted research, while documented cases show old stories being recirculated and mistaken for current. Meta also reports publisher concern. However, no located study estimates population-wide stale-artifact impressions, directly measures currentness misclassification for superseded articles, or traces consequential decisions to that confusion.","source_ids":["S1","S2","S3"]},"stakeholder_evidence":{"support":"STRONG","rationale":"Publisher concern is reported by Meta, and Meta, GOV.UK, HHS, and the Guardian have implemented age warnings or lifecycle-management mechanisms. This demonstrates operational stakeholder pull and credible adopters, although it does not establish willingness to fund the candidate's full composition.","source_ids":["S3","S4","S5","S6","S7"]},"prior_art":{"proximity":"ESTABLISHED_PRACTICE","closest_analogues":[{"name":"GOV.UK Whitehall retirement workflow","similarity":"Uses explicit withdrawn, unpublished, and history states; visible explanations and successor links; same-URL preservation; internal-search exclusion or demotion; redirects; permissioned editorial action; and reversible unwithdraw or republish procedures.","remaining_difference":"The documentation does not show a controlled comparison against banner-only treatment, systematic inbound-citation checks, legal-hold integration, restore drills with measured success, or the proposed audience-error and workload evaluation.","source_ids":["S4","S5"]},{"name":"HHS content lifecycle and online archive program","similarity":"Uses scheduled review, ownership and last-reviewed metadata, retain/revise/remove/archive classes, supersession-triggered archiving, successor-to-predecessor links, records responsibility, and recurring archive captures.","remaining_difference":"The public guidance does not document recommendation demotion, blinded currentness testing, reference-breakage thresholds, reversible quarantine metrics, or comparative effectiveness against a notice-only baseline.","source_ids":["S6"]},{"name":"Publisher and platform age-context notices","similarity":"The Guardian visibly labels old articles, while Facebook warns before sharing articles older than 90 days and permits users to proceed.","remaining_difference":"These are primarily age/context interventions. They do not establish editor-confirmed supersession, dependency and hold checks, reversible retirement, or superiority of simulated demotion over a prominent successor notice.","source_ids":["S3","S7"]}],"distinctive_claim_remaining":"Lifecycle labeling, demotion, archiving, successor linking, permissions, and reversibility are not distinctive by themselves. The remaining testable claim is that, for one predeclared volatile newsroom class, adding dependency and hold checks plus recoverability verification to editor-confirmed supersession—and using simulated distribution demotion—improves currentness discrimination beyond a prominent date-and-successor notice without unacceptable reference, access, suppression, restoration, or workload costs.","confidence":"HIGH"},"implementation_evidence":{"support":"STRONG","rationale":"GOV.UK and HHS document operating workflows with lifecycle states, review responsibility, banners, search suppression or demotion, redirects, archiving, and reversal. Meta and the Guardian demonstrate deployed age-context interfaces. These sources establish mechanical implementability, not the candidate's comparative effectiveness or sustainable newsroom workload.","source_ids":["S3","S4","S5","S6","S7"]},"scores":{"meaningful_impact":{"score":3,"rationale":"Stale layers are documented in a high-stakes science-news sample and real recirculation cases, but audience-error prevalence and consequential harm remain unmeasured.","source_ids":["S1","S2"]},"stakeholder_pull":{"score":4,"rationale":"Publishers have expressed concern, and several major organizations operate related controls. Evidence of demand for the full integrated composition or a committed budget was not found.","source_ids":["S3","S4","S6","S7"]},"incremental_advantage":{"score":2,"rationale":"The proposed notice, demotion, successor-link, archive, permission, and reversal mechanisms substantially overlap existing GOV.UK and HHS practice. Added dependency checks, holds, restore metrics, and comparative evaluation could help, but no outcome evidence shows advantage over notice-only or incumbent lifecycle systems.","source_ids":["S4","S5","S6"]},"distinctiveness_plausibility":{"score":2,"rationale":"A narrow evaluative and safety-control difference remains, but the core mechanism composition is established practice rather than a clearly novel intervention.","source_ids":["S3","S4","S5","S6","S7"]},"technical_implementability":{"score":4,"rationale":"Large publishers already implement most components. Identity resolution, citation discovery, syndication dependencies, hold integration, and unbiased manual classification remain material integration challenges.","source_ids":["S4","S5","S6"]},"adoption_authority_feasibility":{"score":5,"rationale":"Official systems assign action to managing editors, content owners, and records functions, confirming that a publisher can authorize changes on its own channels while preserving limits on external copies.","source_ids":["S5","S6"]},"evidence_readiness":{"score":5,"rationale":"A safe offline comparison can be constructed from documented superseded artifacts and incumbent interface patterns, with currentness error, findability, reference integrity, restoration, suppression review, and workload as predeclared outcomes.","source_ids":["S1","S3","S4","S6"]},"safety_net_benefit":{"score":4,"rationale":"Stable URLs, public explanations, successor links, reversible states, archives, and records responsibility provide credible safeguards. The additional value of holds, dependency checks, and restore drills remains untested.","source_ids":["S4","S5","S6"]},"scalability":{"score":4,"rationale":"GOV.UK, HHS, Meta, and the Guardian show that several components operate at institutional scale. Scaling contextual supersession decisions and dependency review across heterogeneous newsroom backlogs may still be labor-intensive.","source_ids":["S3","S4","S6","S7"]}},"score_confidence":"MODERATE","costs":{"first_evidence":{"band_2026_usd":"50K_TO_250K","scope":"Six-week, no-public-change shadow study covering 150–300 volatile advisories: inventory and provenance work, editorial supersession review, three prototype conditions, participant recruitment and incentives, blinded currentness tasks, simulated search exposure, citation and accessibility checks, restore rehearsal, records/legal review, analysis, and decision report.","confidence":"LOW","assumptions":["Existing content and exposure logs can be analyzed under publisher authorization.","Prototype pages and simulated rankings are used; production visibility is unchanged.","The team includes fractional editorial, engineering, user-research, data-analysis, records/legal, accessibility, and project-coordination labor.","No direct published cost source was found; official workflows bound the required work components but not their prices."],"source_ids":["S4","S5","S6"]},"initial_deployment_startup":{"band_2026_usd":"250K_TO_1M","scope":"One-publisher preparation across several volatile classes, including lifecycle metadata, CMS and search integration, identity and successor resolution, dependency reports, role permissions, hold handling, reversible quarantine, audit logs, archive validation, security and privacy review, training, accessibility, software, and evaluation instrumentation.","confidence":"LOW","assumptions":["The publisher has modifiable CMS, search, analytics, and archive systems.","Historical backlog remediation is risk-prioritized rather than exhaustive.","Software and storage largely reuse existing infrastructure.","Final classifications and irreversible actions remain human-authorized."],"source_ids":["S4","S5","S6"]},"operational_launch":{"band_2026_usd":"1M_TO_5M","scope":"Organization-wide launch for one medium-to-large publisher across articles, advisories, live-blog entries, social metadata, search, recommendation, and syndication interfaces, including backlog triage, engineering, data migration, archive equipment or services, security, legal and records compliance, accessibility, staff training, partner coordination, monitoring, and independent rollout evaluation.","confidence":"LOW","assumptions":["Launch does not attempt to control third-party copies or archives.","Legacy content identities and inbound references require substantial remediation.","Multiple editorial desks and governance functions participate.","The band is resource-equivalent and includes internal labor even where no cash procurement occurs."],"source_ids":["S4","S5","S6"]},"annual_recurring":{"band_2026_usd":"250K_TO_1M","scope":"Ongoing validation queues, editor and records review, exception and legal-hold management, dependency checking, archive and restore testing, storage and software, analytics monitoring, accessibility and security maintenance, audits, training, partner coordination, and periodic outcome evaluation for one publisher.","confidence":"LOW","assumptions":["Risk-based rules keep review volume bounded.","Automation prioritizes cases but does not decide truth, suppression, or destruction.","Most base storage and publishing infrastructure already exists.","Sustained backlog growth could move recurring cost into a higher band."],"source_ids":["S4","S5","S6"]}},"verified_pipeline_gates":{"externally_supported_problem":{"status":"YES","reason":"Primary research verifies a bounded class of superseded news that was rarely updated, and independent reporting documents currentness confusion during recirculation. The aggregate prevalence and consequential-harm rate remain unknown.","source_ids":["S1","S2"]},"externally_credible_adopter_or_authorizer":{"status":"YES","reason":"GOV.UK and HHS are operating institutional adopters, while managing editors, content owners, and records functions are credible authorizers within their controlled channels.","source_ids":["S4","S5","S6"]},"distinct_testable_incremental_claim":{"status":"YES","reason":"The remaining claim can isolate simulated demotion and added dependency, hold, and recoverability controls from a prominent date-and-successor notice, while measuring currentness errors and preservation costs. The claim is narrow because most base mechanisms are established.","source_ids":["S3","S4","S5","S6","S7"]},"bounded_next_evidence_step":{"status":"YES","reason":"A six-week offline shadow study on a predeclared artifact class can compare interfaces and simulated rankings without changing public content or distribution.","source_ids":["S1","S3","S4","S6"]},"no_unresolved_safety_or_authority_stop":{"status":"YES","reason":"The next step uses copied pages and simulated exposure only, retains all artifacts, requires authorized access, and makes no public classification, deletion, third-party request, or production-ranking change.","source_ids":["S4","S5","S6"]},"credible_cost_scope_and_range":{"status":"YES","reason":"The bands include editorial, engineering, research, data, software, archiving, security, accessibility, legal/records compliance, coordination, and evaluation. Confidence is low because official sources document workflow scope rather than prices.","source_ids":["S4","S5","S6"]}},"next_evidence_step":"Run a six-week, no-production-change shadow study on 150–300 predeclared volatile advisories. Editors independently identify genuinely superseded items and successor links; records/legal staff flag exclusions and holds. In an offline mirror, randomly assign blinded participants to (A) the current date-only presentation, (B) a prominent explicit superseded-state and successor notice, or (C) the same notice plus simulated removal from internal-search results. Compare currentness misclassification and stale-item selection, and audit citation resolution, successor findability, classification disagreement, protected-class or viewpoint disparities, restoration success, and reviewer minutes. Falsify progression if B does not improve currentness judgment over A, if C adds no decision-relevant benefit over B, or if either treatment exceeds predeclared reference, access, bias, restoration, or workload limits. No live deployment or public demotion is part of this step.","blocking_evidence":["No representative estimate of unmarked-supersession prevalence across publication classes or publishers.","No direct causal estimate linking unmarked supersession to currentness misclassification or consequential decisions.","No controlled evidence that distribution demotion adds benefit beyond a prominent supersession and successor notice.","No measured broken-reference, successor-findability, historical-access, suppression-error, or restoration rates for the proposed composition.","No measured reviewer throughput, disagreement, queue growth, exception burden, or legal-hold workload.","No evidence that classification errors are acceptably distributed across political, cultural, demographic, and subject-matter categories.","No direct 2026 cost data for newsroom-wide lifecycle integration and recurring review."],"research_disposition":"KNOWN_PRACTICE_DIFFUSION","world_novelty_boundary":"This bounded open-web search found close operational analogues and establishes that the core lifecycle mechanism is already practiced. It did not comprehensively search patents, procurement records, vendor CMS features, private newsroom systems, non-English practices, or every archival standard. Accordingly, this is not a world-novelty determination; only the narrow comparative safety-and-effectiveness claim remains open."}