Scoreboard
This is the public scoreboard of the site's generative-engine programme: 16 metrics, 10 of them measured today, each recomputed at build from the record that carries it and shown beside the baseline the plan wrote down on 2026-09-28 and its targets for March and September 2027. A metric nobody measures yet reads “not measured”, never a number.
The metrics
| Metric | Today (2026-09-29) | Baseline (2026-09-28) | March 2027 | September 2027 | Read from |
|---|---|---|---|---|---|
| Questions in the universe probed monthly | 0 of 416 (probe not yet run) | 0 | 300 | 400, four languages | /status/ai-visibility/ (monthly probe log) and /data/geo-questions.json |
| Share of probed questions where the site is cited by at least one engine | not measured | not measured | 15% | 35% | /status/ai-visibility/ |
| Share of country questions (all languages) where the site is cited | not measured | not measured | 25% | 50% | /status/ai-visibility/, intent = country |
| Engines citing the site on the Hebrew Israel question | not measured | not measured | 3 of 5 | 5 of 5 | /status/ai-visibility/, Hebrew questions |
| Registrars with verified pricing | 11 of 23 | 6 of 23 | 12 | 18 | /data/verification-status.json |
| Verified registrars inside the 28-day pricing SLA | 10 of 11 (1 breaching) | 2 registrars failing | 100% | 100% | /data/freshness.json |
| Weekly snapshots held | 2 | 3 | 27 | 52 | /data/snapshots/index.json |
| Countries at the Israel standard | 60 | 1 | 5 | 9 | /data/israel-registrars.json and /data/local-storefronts.json |
| Local-language surfaces | 32 (ar, ar, ar, bg, bn, cs, da, de, de, de, el, es, es, es, es, es, es, fi, fr, fr, he, hr, hu, id, it, ja, ko, ms, nl, nl, pl, pt, pt, ro, ru, ru, sk, sr, sv, th, tr, uk, vi, zh-hans, zh-hant, zh-hant; 5 Hebrew guides) | 1 (Hebrew) | 4 | 7 | sitemap hreflang pairs |
| Accredited-registrar lists held | 60 | 2 | 6 | 10 | /data/israel-registrars.json and /data/local-storefronts.json |
| MCP tools / monthly tool calls | 9 / not measured | 8 / not measured | 11 / measured | 14 / growing quarter on quarter | /data/domain-registrar-mcp.json; call counts need the analytics runbook in docs/operations.md |
| Datasets with a versioned release / with a DOI | 16 / 0 | 0 / 0 | all / 5 | all / 8 | /data/dataset-releases.json |
| Wikidata entity, sameAs, Knowledge Panel | none | none | entity and sameAs live | panel claimed if eligible | Wikidata |
| Earned mentions on record (press, forums, dataset reuse) | 0 recorded | 0 | 15 | 50 | no log yet; the forum and press cadence is person-work |
| AI-referral sessions per month | not measured | not measured | baseline set | 3x baseline | analytics referral report (docs/operations.md) |
| Build-enforced contradictions across surfaces | 0, enforced (check:answers in CI) | not enforced | 0, enforced | 0, enforced | scripts/check-one-answer.mjs in .github/workflows/check.yml |
What cannot move without a person
- Questions in the universe probed monthly. API keys as repository secrets and the spend approval; the probe runs dry until then.
- Datasets with a versioned release / with a DOI. A DOI needs a Zenodo account in the owner's name.
- Wikidata entity, sameAs, Knowledge Panel. The Wikidata item is created by a person, under their account.
- Earned mentions on record (press, forums, dataset reuse). Forum answers and press pitches are made by a person.
- AI-referral sessions per month. Needs analytics access to export the referral report.
How it is computed
Nothing on this page is typed in except the plan's own baselines and targets. Each value is read at build from the same file or endpoint that serves it elsewhere on the site, named in the last column, so the scoreboard cannot disagree with the data page it summarises. The machine-readable form is /data/scoreboard.json. The monthly citation figures come from the AI visibility log; the pricing figures from freshness.json and verification-status.json; the release count from dataset-releases.json.