IDKMesh

Google Search Console and Bing Webmaster Measurement Runbook

Scope: https://mskazemi.com/idkmesh/

This runbook converts the repository-controlled SEO/AEO/GEO architecture into first-party search and AI-citation evidence. It does not promise ranking and it does not treat crawler health as proof of indexing.

The canonical target map is config/seo-topics-v1.json. The canonical observed evidence ledger is evidence/search-visibility/observations.json.

Google Search Console

Property

Use a URL-prefix property for:

https://mskazemi.com/idkmesh/

Google documents that URL-prefix properties can include a path and contain only URLs that begin with that exact protocol/host/path prefix. This keeps IDKMesh measurement separate from unrelated content on mskazemi.com.

Official property guidance: https://support.google.com/webmasters/answer/10432366 https://support.google.com/webmasters/answer/34592

Use an ownership-verification method available to the maintainer account. Do not commit account cookies, OAuth tokens, passwords, or private account exports.

Sitemap

Submit:

https://mskazemi.com/idkmesh/sitemap.xml

After submission, record the submission state and any processing error in issue #665. The repository’s public discovery monitor already verifies that the sitemap is retrievable and contains the topic architecture; Search Console is the first-party evidence for how Google processes it.

URL inspection acceptance set

Inspect these 11 canonical topic URLs:

1. https://mskazemi.com/idkmesh/topics/
2. https://mskazemi.com/idkmesh/topics/ai-agent-verification.html
3. https://mskazemi.com/idkmesh/topics/multi-agent-orchestration.html
4. https://mskazemi.com/idkmesh/topics/ai-code-review.html
5. https://mskazemi.com/idkmesh/topics/llm-judge-reliability.html
6. https://mskazemi.com/idkmesh/topics/verifier-panels.html
7. https://mskazemi.com/idkmesh/topics/agent-governance.html
8. https://mskazemi.com/idkmesh/topics/provenance-evidence.html
9. https://mskazemi.com/idkmesh/topics/mcp-a2a-interoperability.html
10. https://mskazemi.com/idkmesh/topics/verification-scaling.html
11. https://mskazemi.com/idkmesh/topics/verified-swarm-engineering.html

For each URL record:

• indexed / not indexed;
• Google-selected canonical versus declared canonical;
• crawl/index reason when not indexed;
• last crawl when available;
• whether a live test reports the page as retrievable.

Do not turn a successful live test into an indexing claim.

Search Performance exports

Use at least these time windows when data exists:

• 7 days for recent change detection;
• 28 days for the normal iteration loop;
• 3 months for trend context.

Export query and page data. Preserve Google’s reported population and date window. Do not convert average position into a probability of ranking.

For the 100-intent review:

1. normalize query text for comparison only; retain the original query;
2. map an observed query to the closest existing intent only when the semantic relationship is clear;
3. keep truly new intent classes separate rather than forcing a match;
4. prioritize pages with impressions but weak CTR/position for content/title review;
5. distinguish ‘no impressions observed’ from ‘not indexed’.

Bing Webmaster Tools

Site property

Bing supports verification of an entire domain or a site limited to a branch or directory. The preferred IDKMesh scope is the same public path:

https://mskazemi.com/idkmesh/

If the Google Search Console property above is verified, Bing can import a verified Search Console site and its sitemap metadata. Otherwise use a supported Bing verification method.

Official guidance: https://www2.bing.com/webmasters/help/add-and-verify-site-12184f8b

Sitemap and crawl health

Submit or confirm discovery of:

https://mskazemi.com/idkmesh/sitemap.xml

Record:

• submission/discovery method;
• processing status;
• last-read time;
• URL count reported by Bing;
• processing errors.

Bing documents both robots.txt sitemap discovery and direct Webmaster Tools submission. IDKMesh already uses both the sitemap and IndexNow on the public side.

Official sitemap guidance: https://blogs.bing.com/webmaster/July-2025/Keeping-Content-Discoverable-with-Sitemaps-in-AI-Powered-Search/ https://www2.bing.com/webmasters/help/sitemaps-3b5cf6ed

Search Performance

Export page/query performance for the same 7-day, 28-day, and 3-month windows when available. Bing Search Performance reports impressions, clicks, and query phrases across supported search surfaces.

Official guidance: https://www.bing.com/webmasters/help/search-performance-c680da36

AI Performance

Bing AI Performance is the first-party Microsoft evidence source for IDKMesh citations in supported AI experiences including Copilot, Bing AI-generated answers, and selected partner integrations.

Record/export:

• total citations;
• cited pages;
• grounding queries;
• grounding-query to page mappings;
• time-series citation activity;
• intents/topics when available;
• citation share when available.

Keep Bing’s evidence boundary intact: citation activity is not ranking, authority, importance, or proof of causation.

Official guidance: https://www.bing.com/webmasters/help/ai-performance-9f8e7d6c

Exports are available in CSV/Excel according to Bing’s current documentation. Retain the original export outside the repository if it contains account/private metadata; commit only the minimum normalized observations needed for the public evidence ledger.

Offline AI Performance CSV normalization

Because the authenticated Bing Webmaster REST contract is not fully available from the public documentation surface, IDKMesh does not guess API endpoints or payloads. Owner-downloaded AI Performance CSV exports can instead be normalized offline with:

python scripts/bing_ai_performance_import.py \
  --kind grounding \
  --input /path/to/bing-grounding.csv \
  --window-start 2026-08-25 \
  --window-end 2026-09-23 \
  --query-column "Grounding Query" \
  --citations-column "Citations" \
  --output results/visibility/bing-ai-grounding.json

The export headers are supplied explicitly through CLI arguments. The importer does not assume Bing’s CSV column names.

Supported export shapes:

• grounding: grouped grounding phrase + citation count;
• pages: IDKMesh page URL + citation count;
• mapping: grounding phrase + IDKMesh page URL + citation count;
• timeline: date + citation count.

The importer:

• uses only the Python standard library;
• fingerprints the source CSV with SHA-256;
• rejects page rows outside https://mskazemi.com/idkmesh/;
• validates the declared date window;
• maps a grounding phrase to the canonical 100-query portfolio only when the normalized phrase is an exact match;
• does not infer semantic equivalence, ranking, traffic, endorsement, or causation.

Keep the raw account export outside the public repository when it contains private/account metadata. A normalized output may be retained as evidence only after reviewing that it contains the minimum public aggregate information needed for the observation.

Recording evidence in IDKMesh

Every committed search/answer-engine observation must conform to:

schemas/search-visibility-observation-v0.1.schema.json

The canonical ledger is:

evidence/search-visibility/observations.json

Regenerate:

python tools/search_visibility_report.py > evidence/search-visibility/REPORT.md
python -m pytest -q tests/test_search_visibility_report.py

Evidence classes:

• webmaster_export for Search Console / Bing Webmaster exported metrics;
• index_inspection for URL inspection/index status;
• manual_reproduction for a dated reproducible search/AI-answer observation;
• referral_log only when an actual referral observation exists.

Do not store:

• account email addresses;
• OAuth/access tokens;
• cookies;
• private browsing/search history;
• personal identifiers;
• screenshots containing unrelated account data.

28-day review loop

1. export Google query/page performance;
2. export Bing query/page performance;
3. export Bing AI Performance citation/grounding data;
4. update only reproducible observations in the ledger;
5. regenerate REPORT.md;
6. compare observed queries/grounding phrases with the 100-intent map;
7. improve pages where evidence shows a real gap;
8. avoid creating thin pages for spelling or exact-match variants;
9. record the decision and supporting evidence in issue #665.

Interpretation rules

• Crawlable is not indexed.
• Indexed is not ranked.
• An impression is not a click.
• A citation is not an endorsement.
• A grounding query is not necessarily the user’s exact prompt.
• A short-term change after a content update does not establish causation.
• Different engines and surfaces remain different measurement populations.

Current external actions

Issue #665 is the visibility/indexing/citation measurement tracker.

Issue #769 is the separate PyPI distribution activation tracker.

This site is open source. Improve this page.