Bing Average Cited Pages: citation-breadth measurement and portfolio governance

Short answer: Treat Bing's Average Cited Pages metric as a portfolio-level breadth signal, not a ranking, authority or answer-position score. Microsoft says the metric is the average number of unique pages from your site displayed as sources in AI-generated answers per day over the selected time range. Because it is aggregated across supported AI surfaces, it reflects citation patterns rather than the role, prominence or ranking of any individual page. Use it to monitor breadth, then reconcile changes with page-level citation activity, grounding queries, freshness and ordinary search health.

What Bing currently documents

Bing Webmaster Tools AI Performance is in public preview and reports citation activity across Microsoft Copilot, AI-generated summaries in Bing and select partner integrations.

Microsoft defines Average Cited Pages as the average number of unique pages from a site shown as sources in AI-generated answers per day during the selected period.

Microsoft explicitly says the metric does not indicate ranking, authority or a page's role in an individual answer.

Step 1: define the reporting window

For every export or review, capture:

• site/property;
• date range;
• extraction date;
• supported-surface scope as documented;
• filters;
• analyst;
• report version/public-preview status.

Do not compare periods with different filters or product definitions without annotation.

Step 2: interpret breadth, not importance

If Average Cited Pages rises, a broader set of pages may be appearing as sources across supported AI experiences.

That does not establish:

• higher ranking;
• more prominent placement;
• stronger authority;
• more clicks;
• more conversions;
• better answer quality.

Keep the metric name and definition intact in reporting.

Step 3: create stable page groups

Segment the portfolio into groups such as:

• evergreen guides;
• product/service pages;
• category hubs;
• local pages;
• news/editorial;
• support/documentation;
• comparison content;
• recently updated pages.

Stable groups help identify whether breadth change comes from one cluster or the whole site.

Step 4: reconcile with page-level citation activity

Bing also reports citation counts for specific URLs.

When portfolio breadth changes, ask:

• Which pages newly appear?
• Which pages disappeared?
• Did one cluster expand?
• Did many pages receive one citation each?
• Did a smaller set receive more repeated citations?

Average Cited Pages alone cannot answer those questions.

Step 5: use grounding queries as context

Bing's grounding-query report shows sampled phrases used when retrieving content that was cited.

Use the sample to understand topic context, not to estimate total query demand.

Connect:

• page group;
• cited URL;
• grounding-query theme;
• source freshness;
• content intent.

Do not turn a sampled phrase into a keyword-volume estimate.

Step 6: monitor freshness and indexing

A drop in citation breadth may coincide with:

• stale content;
• indexing changes;
• redirects;
• canonical changes;
• robots controls;
• URL migrations;
• publishing mix;
• ordinary demand shifts.

Check crawl/index health before attributing the change to AI systems.

Step 7: avoid denominator confusion

Because the metric is an average over days in a selected range, interpretation depends on the reporting window.

Document whether a comparison uses:

• equal-length periods;
• seasonal periods;
• launch windows;
• migration windows;
• partial periods.

Do not compare a short event window with a long stable period as if their volatility were equivalent.

Step 8: define portfolio thresholds internally

You may define internal alerts such as:

• material decline from baseline;
• new cluster appearing;
• expected cluster disappearing;
• breadth rising while total citations fall;
• total citations rising while breadth narrows.

These are internal diagnostic thresholds, not Bing ranking thresholds.

Step 9: keep external outcomes separate

Maintain distinct evidence layers:

• Average Cited Pages;
• Total Citations;
• page-level citations;
• grounding-query sample;
• ordinary Bing search metrics;
• site sessions;
• leads/revenue.

Citation breadth should not be promoted to traffic or business value without downstream evidence.

Step 10: version the metric contract

AI Performance is a public-preview product and Microsoft says metrics will continue evolving.

Store:

• source date;
• definition used;
• product status;
• extraction logic;
• known limitations;
• revalidation date.

Update internal dashboards when Microsoft changes definitions rather than silently stitching incompatible periods together.

Governance states

Use states such as:

• BREADTH_STABLE;
• BREADTH_INCREASED;
• BREADTH_DECLINED;
• PAGE_MIX_CHANGED;
• GROUNDING_CONTEXT_REVIEW;
• INDEXING_REVIEW_REQUIRED;
• METRIC_VERSION_CHANGED;
• BUSINESS_IMPACT_UNKNOWN.

The governance rule

Average Cited Pages is most useful as a portfolio-breadth diagnostic that must be reconciled with URL-level evidence and reporting scope.

Use it to see whether citation participation is broadening or narrowing, then investigate the pages and topics behind the change. Microsoft's documentation explicitly limits the metric: it does not indicate ranking, authority or the role of a page inside an individual AI answer.

Sources reviewed

• https://blogs.bing.com/webmaster/February-2026/Introducing-AI-Performance-in-Bing-Webmaster-Tools-Public-Preview