AI Visibility Tracking ToolFree checker

Citation share: definition and how to read it

99Visibility Research · Updated October 21, 2026 · 4 min read

Short answer

Citation share is your citations as a fraction of all citations recorded for a set of queries. If AI answers for those queries contained 200 citations and 30 pointed to your pages, your share is 15%. Bing Webmaster Tools reports a Citation Share figure, and Microsoft describes it as an observational metric, not a ranking or a competitive scoreboard. It can fall while your absolute citations rise.

Key takeaways

  • Citation share is a ratio: your citations divided by all citations for the same query set and time window.
  • Your share can drop while your citation count grows, because other sources can gain citations faster than you do.
  • Microsoft frames Bing's Citation Share as observational: it is not a rank, not traffic share and not a quality score.
  • Always read citation share next to the absolute citation count, and per engine, never as one blended number.

Definition

Citation share is the number of citations pointing to your pages, divided by the total number of citations recorded across a defined set of queries in a defined period.

Two things define it, and both matter:

  • The numerator: citations to your domain or pages.
  • The denominator: citations to every source for the same queries, including competitors, publishers, forums and reference sites.

Change either part and the number moves. That makes citation share a relative measure. It tells you how much of the cited space you occupy, not how often you appear.

How it works

Here is a hypothetical. In month one, AI answers for your tracked queries contain 100 citations and 20 are yours: a 20% share. In month two, the same queries produce 200 citations and 30 are yours: a 15% share. You gained 10 citations and lost five points of share.

That pattern is normal. Answers can cite more sources over time, new sources can enter the pool, and a few large sites can absorb most of the growth. Our page on what a falling citation share means works through the common causes.

Where the number comes from

As of October 2026, Bing Webmaster Tools includes Citation Share in its AI Performance report, which covers Microsoft Copilot, Bing's AI summaries and select partner integrations. The report has no API, and it does not show which competitor domains make up the rest of the denominator. See reading Bing's AI Performance report for how to use it, and whether it shows competitor citations.

Third-party trackers compute their own share from prompts they choose. That is a sample, and samples can differ a lot from first-party counts. In Otterly's 2,000 AI blogs experiment, Microsoft's report credited one test site with 42,817 citations while Otterly's prompt tracker logged 3,934, though the time windows may not match.

Why it matters for AI visibility

Citation share answers a narrow question: of everything the engine cited for these queries, how much was you? That is useful for spotting when a query set gets more crowded.

It is also easy to misread, because the denominator is usually dominated by a few big sources. A 2025 study of AI search citations (arXiv:2507.05301) found that the top 20 most frequent news sources accounted for 67.3% of all citations for OpenAI models. When a handful of large sites hold most citations, your share mostly reflects how much room they leave.

The next move is simple:

  1. Track your absolute citation count alongside share.
  2. Split both by engine. Google says AI Overviews and AI Mode may use different models and techniques, so even two Google surfaces should not be averaged.
  3. Note the query set and window every time you report a share, or comparisons over time mean nothing.

Common misconceptions

  • "Citation share is my rank." It is not. No engine publishes a position inside an AI answer, and share says nothing about where in an answer you appear. Our guide on why position isn't the metric covers this.
  • "It equals my share of AI traffic." Microsoft states its Citation Share does not represent traffic share. A citation is not a click.
  • "More citations means more influence." Citation and use are different things. Mechanistic research on how LLMs cite (arXiv:2606.28358) repaired over 90% of missed citations by amplifying specific attention heads, without changing the sources.
  • "It measures the prompts my customers type." OpenAI says ChatGPT search typically rewrites a user's query into one or more targeted queries. Bing's grounding queries show what the AI actually ran, which is why tracking Copilot grounding queries is worth the effort.
  • "A falling share means I did something wrong." Often it means others gained faster. Check the count before you act.

Frequently asked questions

Can my citation share go up while I lose citations?

Yes. If the total number of citations for your query set shrinks faster than yours do, your share rises even as your count falls. Say answers drop from 200 citations to 80 and yours drop from 20 to 12: your share moves from 10% to 15%. That is why share should never be reported without the count beside it.

Is a higher citation share always better?

Not on its own. A high share on queries nobody asks is worth little, and a share can look strong simply because an engine cites fewer sources per answer. Treat share as context for your absolute citations on queries that matter to buyers, measured per engine and per period, rather than as a target to push up.

Can I see citation share for ChatGPT or Google?

Not from a first-party report as of October 2026. Google's Search Console AI report shows impressions only, and OpenAI publishes no citation report for site owners. Any ChatGPT or Google share you see comes from a third-party tracker's own prompt sample, so it depends on which prompts that tracker chose.

Should I use citation share to compare myself with a competitor?

Carefully. Microsoft says its Citation Share is not a competitive scoreboard and does not expose competitor domains. Third-party trackers can estimate a competitor's share, but only for their chosen prompts. Compare like-sized brands on the same query set and window, and read the gap as an observation, not a verdict.

Sources

  1. 2000 AI Blogs Experiment: Does Google Penalize AI Content?
  2. [2507.05301] News Source Citing Patterns in AI Search Systems
  3. [2606.28358] How Do LLMs Cite? A Mechanistic Interpretation of Attribution in Retrieval-Augmented Generation
  4. Searching the web with ChatGPT
  5. AI Features and Your Website

This page is AI-assisted. It was written from a verified evidence base and checked against our editorial policy. Engine behaviour changes often; facts are as of October 21, 2026.

Keep reading