Product
AI visibilityCompetitorsCitationsGEO score and page fixesPrompts and demandAI crawlers and referralsContent StudioAgents and automationsAurora strategistCommerceEarned mediaLocal networkAPI and MCPThe whole platform
Solutions
Brand and marketingSEO and GEO teamsContent teamsPR and communicationsE-commerceMulti-location and localAgenciesExecutives
Free tools
Free AI visibility scanPrompt set generatorAI readiness checkTemplates
Resources
Learn GEOUse casesHoliday AI Index 2026BlogNewsletterMethodologyDevelopersChangelogIntegrationsSecurity and trust
PricingAboutSign in
Check your AI visibilityBook a demo
Half-life: A strong current, never renewed, loosens into scattered marks. So does a citation.
Currents, plate 01Half-lifeA strong current, never renewed, loosens into scattered marks. So does a citation.Download the print, 3200 px
Content

Citation decay: why your best page stops getting quoted

How fast sources lose their hold on AI answers, how to measure it, and what slows it down.

The short answer

Citation decay is the drop in how often an AI engine cites a source as the source ages. How fast depends on the format and the engine. Pages decay more slowly when they are updated with visible dates, contain original data, and remain the most complete answer to their question.

A page that engines cite today will not necessarily be cited in three months, even if nothing about it changes. Newer pages arrive, engines refresh their sources, and old fade. We call that .

Why do AI engines stop citing a page?

How fast depends on the format: buying guides outlast product reviews, news fades first, and the curve differs by engine too.

How do we measure citation decay?

Because Lighthouse sends the same prompts on a schedule, we can watch each source enter the answers, peak and fall away. For each source and engine we fit a decay curve from those repeated runs and estimate a half-life: how long until it is cited half as often as at its peak. The method is in our methodology.

What slows citation decay down?

  • Visible, meaningful updates. A dated “updated” note with real changes, not a new timestamp on old text.
  • Original data. Pages with numbers other pages need to cite stay cited.
  • Completeness. If your page already answers the follow-up questions, newer pages add less, and engines have less reason to switch.

A refresh calendar

Use the curves to schedule refreshes before a page drops out, not after. In Lighthouse, citation decay tracking follows the curves for your cited pages and flags when one starts to fade, with the prompts it is losing and the sources replacing it.

Questions people ask

What is citation decay?

The decline in how often AI engines cite a source as it ages, measured from repeated runs of the same prompts.

How often should I refresh content for AI search?

Let the decay curves decide. Pages in fast-moving categories may need a meaningful update every two to three months; evergreen guides much less often.

Get new posts by email.

Updates on AI search and new Lighthouse features. One click to leave.