fix(seo): a school is publishable on any year's results, not the latest

_school_sitemap_rows tested only the latest year's row, which quietly dropped
every school with results in its history but a null row for the most recent
year — a school that stopped reporting, or whose figures were suppressed for
small-cohort disclosure.

The Mallard Academy (150367) is the case that caught it: real KS2 results for
2015-16 through 2018-19, then null rows from 2022-23 on. Its detail page shows
all four years; the sitemap omitted it. Sampling 40 of the 2,206 excluded
schools found 4 like this, so roughly 220 real pages were being withheld.

Publishable is now a property of the school, computed across every row, while
lastmod still comes from the latest row so the most recent Ofsted date wins.
The field list is a module constant shared with _has_publishable_data so the
two checks cannot drift.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_015mWQnpye9F299NVRCCSRvj

backend/app.py

@@ -7,6 +7,7 @@ Uses real data from UK Government Compare School Performance downloads.
 import hashlib
 import re
 from contextlib import asynccontextmanager
+from datetime import datetime, timezone
 from typing import Optional
 
 import numpy as np
@@ -48,11 +49,14 @@ PHASE_GROUPS: dict[str, set[str]] = {
     "all-through": {"all-through"},
 }
 
-BASE_URL = "https://schoolcompare.co.uk"
+# Must match SITE_URL in nextjs-app/lib/site.ts. The apex 301s to www, and a
+# sitemap <loc> that redirects wastes a crawl on every URL it lists.
+BASE_URL = "https://www.schoolcompare.co.uk"
 MAX_SLUG_LENGTH = 60
 
-# In-memory sitemap cache
-_sitemap_xml: str | None = None
+# In-memory sitemap cache: name -> XML. Populated on startup and by the admin
+# regenerate endpoint after a pipeline run.
+_sitemaps: dict[str, str] | None = None
 
 
 def _slugify(text: str) -> str:
@@ -70,43 +74,149 @@ def _school_url(urn: int, school_name: str) -> str:
     return f"/school/{urn}-{slug}"
 
 
-def build_sitemap() -> str:
-    """Generate sitemap XML from in-memory school data. Returns the XML string."""
-    df = load_school_data()
+# Routes worth submitting that are not a school page. /admissions was missing
+# from the sitemap entirely despite being a static, indexable guide.
+STATIC_SITEMAP_PATHS = ("/", "/rankings", "/compare", "/admissions")
 
-    static_urls = [
-        (BASE_URL + "/",          "daily",   "1.0"),
-        (BASE_URL + "/rankings",  "weekly",  "0.8"),
-        (BASE_URL + "/compare",   "weekly",  "0.8"),
-    ]
 
-    lines = ['<?xml version="1.0" encoding="UTF-8"?>',
-             '<urlset xmlns="http://www.sitemaps.org/schemas/sitemap/0.9">']
+# A page has something a search result could state if any of these is present
+# in any year. Shared by _has_publishable_data and the per-school check in
+# _school_sitemap_rows so the two can never drift.
+_PUBLISHABLE_FIELDS = ("rwm_expected_pct", "attainment_8_score", "ofsted_grade")
 
-    for url, freq, priority in static_urls:
-        lines.append(
-            f"  <url><loc>{url}</loc>"
-            f"<changefreq>{freq}</changefreq>"
-            f"<priority>{priority}</priority></url>"
+
+def _has_publishable_data(row) -> bool:
+    """True when a school page has something a search result could state.
+
+    A school with no results in any year and no Ofsted grade renders an empty
+    page. Submitting it spends crawl budget and drags the corpus-wide quality
+    signal down, so it stays out of the sitemap. The page itself still resolves
+    for anyone who has the URL.
+    """
+    for field in _PUBLISHABLE_FIELDS:
+        value = row.get(field)
+        if value is not None and not pd.isna(value):
+            return True
+    return False
+
+
+def _url_element(loc: str, lastmod: str | None = None) -> str:
+    """One <url> entry. No priority or changefreq — Google ignores both."""
+    body = f"<loc>{loc}</loc>"
+    if lastmod:
+        body += f"<lastmod>{lastmod}</lastmod>"
+    return f"  <url>{body}</url>"
+
+
+def _school_sitemap_rows(df) -> list[str]:
+    """A <url> element per school that has something to show.
+
+    lastmod comes from the school's Ofsted date where there is one and is
+    omitted otherwise. An always-now lastmod is a claim Google learns to
+    distrust; an absent one honestly means "unknown".
+    """
+    if df.empty or "urn" not in df.columns or "school_name" not in df.columns:
+        return []
+
+    rows: list[str] = []
+    seen: set[int] = set()
+
+    # Publishable is a property of the SCHOOL, not of its latest row.
+    #
+    # The first cut tested the latest year's row alone, which quietly dropped
+    # every school that has results in its history but a null row for the most
+    # recent year — a school that stopped reporting, or whose figures were
+    # suppressed for small-cohort disclosure. The Mallard Academy (150367) is
+    # the case that caught it: real KS2 results for 2015-16 through 2018-19,
+    # then null rows for 2022-23 onward. Its page shows all four years; the
+    # sitemap omitted it. Roughly 220 schools were affected.
+    publishable_cols = [c for c in _PUBLISHABLE_FIELDS if c in df.columns]
+    publishable: set[int] = (
+        set(df.loc[df[publishable_cols].notna().any(axis=1), "urn"].astype(int))
+        if publishable_cols else set()
     )
 
-    if not df.empty and "urn" in df.columns and "school_name" in df.columns:
-        seen = set()
-        for _, row in df[["urn", "school_name"]].drop_duplicates(subset="urn").iterrows():
+    # Latest row per URN first, so a school's most recent Ofsted date wins.
+    ordered = df.sort_values("year", ascending=False) if "year" in df.columns else df
+
+    for _, row in ordered.iterrows():
         urn = int(row["urn"])
-            name = str(row["school_name"])
         if urn in seen:
             continue
         seen.add(urn)
-            path = _school_url(urn, name)
-            lines.append(
-                f"  <url><loc>{BASE_URL}{path}</loc>"
-                f"<changefreq>monthly</changefreq>"
-                f"<priority>0.6</priority></url>"
-            )
+        if urn not in publishable:
+            continue
 
-    lines.append("</urlset>")
-    return "\n".join(lines)
+        lastmod = None
+        ofsted_date = row.get("ofsted_date")
+        if ofsted_date is not None and not pd.isna(ofsted_date):
+            lastmod = pd.Timestamp(ofsted_date).date().isoformat()
+
+        rows.append(_url_element(
+            BASE_URL + _school_url(urn, str(row["school_name"])), lastmod))
+
+    return rows
+
+
+# Sitemaps cap at 50,000 URLs per file. 10,000 keeps a child small enough to
+# scan by eye in Search Console, which is the point of splitting at all:
+# coverage is reported per submitted sitemap, so one file per page family is
+# what makes an indexation problem attributable to a family.
+SITEMAP_CHUNK_SIZE = 10_000
+
+# Children are served under /sitemaps/ because Next.js only treats a whole
+# bracketed path segment as dynamic — a route folder named "sitemap-[...parts]"
+# is read as a literal static segment and never matches.
+SITEMAP_CHILD_PREFIX = "/sitemaps"
+
+
+def _urlset(rows: list[str]) -> str:
+    return "\n".join([
+        '<?xml version="1.0" encoding="UTF-8"?>',
+        '<urlset xmlns="http://www.sitemaps.org/schemas/sitemap/0.9">',
+        *rows,
+        "</urlset>",
+    ])
+
+
+def build_sitemaps() -> dict[str, str]:
+    """Build the sitemap index and every child, keyed by name."""
+    df = load_school_data()
+
+    children: dict[str, str] = {
+        "static.xml": _urlset(
+            [_url_element(BASE_URL + path) for path in STATIC_SITEMAP_PATHS]),
+    }
+
+    school_rows = _school_sitemap_rows(df)
+    # Always emit at least one school child, so the index shape is stable even
+    # on an empty database.
+    chunks = [school_rows[i:i + SITEMAP_CHUNK_SIZE]
+              for i in range(0, len(school_rows), SITEMAP_CHUNK_SIZE)] or [[]]
+    for n, chunk in enumerate(chunks, start=1):
+        children[f"schools-{n}.xml"] = _urlset(chunk)
+
+    # On a sitemap index, lastmod means "when this sitemap file last changed",
+    # so generation time is the correct value here — unlike on a <url>, where
+    # it would be a claim about content we cannot support.
+    generated = datetime.now(timezone.utc).date().isoformat()
+    index_rows = [
+        f"  <sitemap><loc>{BASE_URL}{SITEMAP_CHILD_PREFIX}/{name}</loc>"
+        f"<lastmod>{generated}</lastmod></sitemap>"
+        for name in children
+    ]
+    index = "\n".join([
+        '<?xml version="1.0" encoding="UTF-8"?>',
+        '<sitemapindex xmlns="http://www.sitemaps.org/schemas/sitemap/0.9">',
+        *index_rows,
+        "</sitemapindex>",
+    ])
+    return {**children, "sitemap.xml": index}
+
+
+def build_sitemap() -> str:
+    """The sitemap index. Kept for `lifespan` and the admin endpoint."""
+    return build_sitemaps()["sitemap.xml"]
 
 
 def clean_filter_values(series: pd.Series) -> list[str]:
@@ -284,7 +394,7 @@ def validate_postcode(postcode: Optional[str]) -> Optional[str]:
 @asynccontextmanager
 async def lifespan(app: FastAPI):
     """Application lifespan - startup and shutdown events."""
-    global _sitemap_xml
+    global _sitemaps
     print("Loading school data from marts...")
     df = load_school_data()
     if df.empty:
@@ -294,9 +404,9 @@ async def lifespan(app: FastAPI):
         # Pre-compute the latest-year snapshot so the first search request is fast
         await asyncio.to_thread(load_latest_school_data)
     try:
-        _sitemap_xml = build_sitemap()
-        n = _sitemap_xml.count("<url>")
-        print(f"Sitemap built: {n} URLs.")
+        _sitemaps = build_sitemaps()
+        n = sum(x.count("<url>") for x in _sitemaps.values())
+        print(f"Sitemaps built: {len(_sitemaps)} files, {n} URLs.")
     except Exception as e:
         print(f"Warning: sitemap build failed on startup: {e}")
 
@@ -1087,16 +1197,28 @@ async def robots_txt():
     return FileResponse(settings.frontend_dir / "robots.txt", media_type="text/plain")
 
 
-@app.get("/sitemap.xml")
-async def sitemap_xml():
-    """Serve sitemap.xml for search engine indexing."""
-    global _sitemap_xml
-    if _sitemap_xml is None:
+def _serve_sitemap(name: str) -> Response:
+    global _sitemaps
+    if _sitemaps is None:
         try:
-            _sitemap_xml = build_sitemap()
+            _sitemaps = build_sitemaps()
         except Exception as e:
             raise HTTPException(status_code=503, detail=f"Sitemap unavailable: {e}")
-    return Response(content=_sitemap_xml, media_type="application/xml")
+    if name not in _sitemaps:
+        raise HTTPException(status_code=404, detail="No such sitemap")
+    return Response(content=_sitemaps[name], media_type="application/xml")
+
+
+@app.get("/sitemap.xml")
+async def sitemap_xml():
+    """Serve the sitemap index."""
+    return _serve_sitemap("sitemap.xml")
+
+
+@app.get("/sitemaps/{name}")
+async def sitemap_child(name: str):
+    """Serve a child sitemap (static.xml, or schools-N.xml)."""
+    return _serve_sitemap(name)
 
 
 @app.post("/api/admin/regenerate-sitemap")
@@ -1106,10 +1228,10 @@ async def regenerate_sitemap(
     _: bool = Depends(verify_admin_api_key),
 ):
     """Rebuild and cache the sitemap from current school data. Called by Airflow after data updates."""
-    global _sitemap_xml
-    _sitemap_xml = build_sitemap()
-    n = _sitemap_xml.count("<url>")
-    return {"status": "ok", "urls": n}
+    global _sitemaps
+    _sitemaps = build_sitemaps()
+    n = sum(x.count("<url>") for x in _sitemaps.values())
+    return {"status": "ok", "urls": n, "sitemaps": len(_sitemaps)}
 
 
 # Mount static files directly (must be after all routes to avoid catching API calls)

backend/tests/test_sitemap.py

@@ -0,0 +1,219 @@
+"""Tests for sitemap generation (spec 2026-08-20, workstream W1).
+
+The sitemap is built from the in-memory school DataFrame, so these inject a
+small frame via monkeypatch rather than touching a database.
+"""
+
+import numpy as np
+import pandas as pd
+import pytest
+
+
+def _schools_df() -> pd.DataFrame:
+    """Two schools: one with results, one with neither results nor Ofsted."""
+    base = {
+        "local_authority": "Testshire",
+        "school_type": "Academy",
+        "phase": "Primary",
+        "year": 202425,
+        "ofsted_date": None,
+    }
+    return pd.DataFrame(
+        [
+            {**base, "urn": 100001, "school_name": "Alpha Primary",
+             "rwm_expected_pct": 62.0, "attainment_8_score": np.nan,
+             "ofsted_grade": 2.0},
+            {**base, "urn": 100002, "school_name": "Ghost Primary",
+             "rwm_expected_pct": np.nan, "attainment_8_score": np.nan,
+             "ofsted_grade": np.nan},
+        ]
+    )
+
+
+@pytest.fixture()
+def sitemap(monkeypatch) -> str:
+    """The sitemap index."""
+    from backend import app as app_module
+
+    monkeypatch.setattr(app_module, "load_school_data", _schools_df)
+    return app_module.build_sitemap()
+
+
+@pytest.fixture()
+def schools_child(monkeypatch) -> str:
+    """The first school child sitemap, where school URLs actually live."""
+    from backend import app as app_module
+
+    monkeypatch.setattr(app_module, "load_school_data", _schools_df)
+    return app_module.build_sitemaps()["schools-1.xml"]
+
+
+@pytest.fixture()
+def static_child(monkeypatch) -> str:
+    from backend import app as app_module
+
+    monkeypatch.setattr(app_module, "load_school_data", _schools_df)
+    return app_module.build_sitemaps()["static.xml"]
+
+
+def test_every_loc_uses_the_www_host(sitemaps):
+    # The apex 301s to www. A <loc> that redirects burns a crawl per URL.
+    # Checked across every file, index included, not just one.
+    for name, xml in sitemaps.items():
+        assert "https://www.schoolcompare.co.uk" in xml, name
+        assert "https://schoolcompare.co.uk" not in xml, name
+
+
+def test_school_with_results_is_listed(schools_child):
+    assert "/school/100001-alpha-primary" in schools_child
+
+
+def test_school_with_no_results_and_no_ofsted_is_omitted(schools_child):
+    # Nothing for a search result to say about it. Submitting it spends crawl
+    # budget and drags the corpus-wide quality signal down.
+    #
+    # Asserted against the child, not the index: the index carries no school
+    # URLs at all, so it would pass this trivially and prove nothing.
+    assert "/school/100002" not in schools_child
+
+
+def test_no_invented_priority_or_changefreq(sitemaps):
+    # Google ignores both. They were noise dressed as signal.
+    for name, xml in sitemaps.items():
+        assert "<priority>" not in xml, name
+        assert "<changefreq>" not in xml, name
+
+
+def test_ofsted_date_becomes_lastmod(monkeypatch):
+    from backend import app as app_module
+    import datetime
+
+    def _df():
+        base = _schools_df()
+        base.loc[base["urn"] == 100001, "ofsted_date"] = datetime.date(2024, 3, 14)
+        return base
+
+    monkeypatch.setattr(app_module, "load_school_data", _df)
+    xml = app_module.build_sitemaps()["schools-1.xml"]
+    assert "<lastmod>2024-03-14</lastmod>" in xml
+
+
+def test_no_lastmod_invented_when_date_unknown(monkeypatch):
+    # An always-now lastmod is a claim Google learns to distrust. Absent
+    # honestly means unknown.
+    from backend import app as app_module
+
+    def _df():
+        df = _schools_df()
+        df["ofsted_date"] = None
+        return df
+
+    monkeypatch.setattr(app_module, "load_school_data", _df)
+    # The child only. The index legitimately carries a lastmod, because there
+    # it means "when this sitemap file changed", which we do know.
+    xml = app_module.build_sitemaps()["schools-1.xml"]
+    assert "<lastmod>" not in xml
+
+
+def test_static_routes_are_listed(static_child):
+    for path in ("/", "/rankings", "/compare", "/admissions"):
+        assert f"<loc>https://www.schoolcompare.co.uk{path}</loc>" in static_child
+
+
+@pytest.fixture()
+def sitemaps(monkeypatch) -> dict:
+    from backend import app as app_module
+
+    monkeypatch.setattr(app_module, "load_school_data", _schools_df)
+    return app_module.build_sitemaps()
+
+
+def test_index_lists_each_child(sitemaps):
+    index = sitemaps["sitemap.xml"]
+    assert "<sitemapindex" in index
+    assert "https://www.schoolcompare.co.uk/sitemaps/static.xml" in index
+    assert "https://www.schoolcompare.co.uk/sitemaps/schools-1.xml" in index
+
+
+def test_index_carries_no_url_elements(sitemaps):
+    # A sitemap index holds <sitemap> entries only; mixing in <url> is invalid.
+    assert "<url>" not in sitemaps["sitemap.xml"]
+
+
+def test_index_does_not_list_itself(sitemaps):
+    assert "<loc>https://www.schoolcompare.co.uk/sitemap.xml</loc>" not in sitemaps["sitemap.xml"]
+
+
+def test_static_child_holds_the_static_routes(sitemaps):
+    static = sitemaps["static.xml"]
+    for path in ("/", "/rankings", "/compare", "/admissions"):
+        assert f"<loc>https://www.schoolcompare.co.uk{path}</loc>" in static
+
+
+def test_school_child_holds_the_schools(sitemaps):
+    assert "/school/100001-alpha-primary" in sitemaps["schools-1.xml"]
+
+
+def test_children_are_chunked_under_the_limit(monkeypatch):
+    # Sitemaps cap at 50,000 URLs per file. Chunk at 10,000 so a child stays
+    # small enough to eyeball in Search Console.
+    from backend import app as app_module
+    import pandas as _pd
+
+    rows = [
+        {"urn": 200000 + i, "school_name": f"School {i}", "year": 202425,
+         "rwm_expected_pct": 60.0, "attainment_8_score": None,
+         "ofsted_grade": 2.0, "ofsted_date": None}
+        for i in range(10_001)
+    ]
+    monkeypatch.setattr(app_module, "load_school_data", lambda: _pd.DataFrame(rows))
+
+    maps = app_module.build_sitemaps()
+    assert maps["schools-1.xml"].count("<url>") == 10_000
+    assert maps["schools-2.xml"].count("<url>") == 1
+
+
+def test_build_sitemap_still_returns_the_index(sitemap):
+    # lifespan and the admin endpoint call build_sitemap(); keep it working.
+    assert "<sitemapindex" in sitemap
+
+
+def test_school_with_results_in_an_earlier_year_is_still_listed(monkeypatch):
+    """Regression: The Mallard Academy (150367).
+
+    Real KS2 results 2015-16 to 2018-19, then null rows from 2022-23 onward
+    because the school stopped reporting. The first cut tested the latest
+    year's row alone and dropped it, along with ~220 others, even though its
+    detail page shows all four years of results.
+    """
+    from backend import app as app_module
+    import pandas as _pd
+
+    base = {"local_authority": "Testshire", "school_type": "Academy",
+            "phase": "Primary", "ofsted_date": None, "ofsted_grade": np.nan,
+            "attainment_8_score": np.nan, "urn": 150367,
+            "school_name": "Mallard Academy"}
+    df = _pd.DataFrame([
+        {**base, "year": 201819, "rwm_expected_pct": 67.0},
+        {**base, "year": 202324, "rwm_expected_pct": np.nan},
+        {**base, "year": 202425, "rwm_expected_pct": np.nan},
+    ])
+    monkeypatch.setattr(app_module, "load_school_data", lambda: df)
+
+    xml = app_module.build_sitemaps()["schools-1.xml"]
+    assert "/school/150367-mallard-academy" in xml
+
+
+def test_school_with_no_results_in_any_year_is_still_omitted(monkeypatch):
+    """The fix must not turn into "list everything"."""
+    from backend import app as app_module
+    import pandas as _pd
+
+    base = {"local_authority": "Testshire", "school_type": "Academy",
+            "phase": "Primary", "ofsted_date": None, "ofsted_grade": np.nan,
+            "attainment_8_score": np.nan, "rwm_expected_pct": np.nan,
+            "urn": 100002, "school_name": "Ghost Primary"}
+    df = _pd.DataFrame([{**base, "year": y} for y in (202324, 202425)])
+    monkeypatch.setattr(app_module, "load_school_data", lambda: df)
+
+    assert "/school/100002" not in app_module.build_sitemaps()["schools-1.xml"]

docs/branch-cleanup-2026-08-20.txt

@@ -0,0 +1,85 @@
+# Remote branch cleanup, 2026-08-20
+# Restore any branch with:  git push origin <sha>:refs/heads/<name>
+
+## Deleted: fully merged into main (content is in main)
+75677f4252b759ef895e7d5f7c19f8f1745bdb59 add-contact-form-footer
+fa1abff642683dfd26ba88a295a0a6710d77147a chore/byline-removal-and-audit-figure
+95081d38bdf87764ef5d298676c25fae4cd3b792 chore/remove-parent-view
+6877abedebfc1c7d95f1f6ebe68945c62be528ab chore/staged-prod-promotion
+090d5f7bec824e083d3252e2c6e636686016304d ci/frontend-checks-speedup
+8c3a5cc4e9f551f0190d85357ce7741ad87f3a4c design/cohort-identity
+955659580067ca8b81bd77e31e4e2554103f094d feat/allow-analytics-iframe-embed
+6828f6cd4417284ea3eb6f088fa20945b8b40ed3 feat/compare-chips-two-per-row
+d5cd0abfee226885119665da5d2aa8288b59217f feat/compare-data-foundation
+6dd9b04b50bee146682da87efad8fc8b526251c5 feat/compare-frontend-rebuild
+96d5fcf5b07b6f175b48e9b20fcb765a320a907f feat/detail-header-details-reveal
+f1388ff5bd0af1409823a1e047b7ba84246a0f70 feat/gias-sixth-form-flag
+eddf74745f86c9c6d9eb07d867246ff9cc90dc20 feat/hero-artwork-v2
+3015c37bac6dc28db58c80fdb9235942c25b83d7 feat/hero-byline
+8e763e39d17a4964cf558e51c03f044186371f6d feat/info-popover-tooltip
+88c653215d520ab6e902c9de55bf27d86eeb90c3 feat/last-distance-offered
+a72323874f7aebdb2e64b6d64a5febd61152d09d feat/last-distance-offered-full
+c9a1892bfb0370e0672e5849cba294ddfabe3c65 feat/latest-cutoff-only
+4e8df006d75d8be2a1d8529ddf855c445854cba1 feat/near-me-by-search
+45ab479062c6a1639facad636fc0cc0cf0fd9155 feat/proposed-to-close-schools
+3bf2e8f262cbe058fda6de6f8ea3e050224a51b9 feat/school-detail-visualisations
+1f80571b1ff217dc92b660a936c02b5f49d07f0b feat/umami-heatmap-recorder
+609bb923d96aa5730131463ea35d5efdd404bf96 feature/ingest-independent-schools
+94151c58ea38a9256d66d15161293486a500c7a2 fix/admissions-section-height
+79246edc22961c2beb3520437e9d064d3b809d10 fix/annual-dag-ks4-national-selector
+59ac9c10b97e0ef1143f57fea06b324e72ac3d4a fix/chart-marker-contrast
+9f8dba227c95706ca3527bd48d381e7622cc0a5e fix/compare-chart-refetch-resilience
+e74d3882ce78a141fa1a57daa3102d7a58852dc3 fix/compare-expert-fixes
+80176cac4db4820e76ea2a156c7a2974bee2f203 fix/compare-final-review-mustfix
+f579630fab6c456e26a6984a3e8eebdbe3184202 fix/compare-mockup-drift
+dc85254ad2ddf134b4434065d4762cef374d2220 fix/compare-null-year-blanks-chart
+43a2c4a6bc539b621f31655aec05ef319a25f343 fix/compare-refresh-and-fetch
+d677b5453365b72c81d6df2de62b1fa0d05d684d fix/daily-dag-cache-invalidation
+f6bb037c471553e8195b5a8b147467ce0d07a688 fix/detail-all-through
+17bd4d5a5eb0b12ca79b97db587f14f7d671e85b fix/detail-chart-truthfulness
+4e6be0ce65647410b4ff74f8b763207c28920c26 fix/detail-inclusion-admissions
+fdda52ff0af3fae03a4b059a655973cdbb269f91 fix/detail-ofsted-correctness
+e36125b24aba254a8d15c5c33b4d2a296e691995 fix/detail-provenance-anchoring
+b31e71ac884df6507f567adc046c9fc52d9d310a fix/detail-report-card-render-date
+32f8a02862be6a1d49f4c3928b17fcf15d4c94cc fix/detail-trend-chart-taller
+e65688d600a86818fe21ae4c61ba27e5b6ec8d7c fix/e2e-brand-assertions
+3adea73ee04cdedfab54b0351878b297f72756ad fix/e2e-compare-chips-phase
+06e4898c30feaedc471f97aba28ddb0d61379f4d fix/e2e-compare-samephase
+9abd020967670a855e80fe5a908c8048a3aa9f14 fix/e2e-distance-locator
+acec8135e1ec7c9c3c255e5b23733a0ef862b590 fix/e2e-rankings-year-pick
+5944d88f0b1517ef1ef56af1b62270d2cf28e717 fix/expert-signoff-mustfixes
+2433101fa08be5df6f170d41790512ffe823d33e fix/font-cascade-and-map-palette
+74ca76d150deec6725259d9637ea86d7bb90c683 fix/gias-legacy-fallback
+bdaa05cd542f563ef74c45307cd8f7fc465193c9 fix/hero-fallback-and-sharp
+d52d384cf23d282b44e9251176f8f3402d808600 fix/hero-map-ios-fullscreen
+4043270a77bbe4edb18207fa5f1d94d4747fe12f fix/hero-mobile-and-wording
+22e9eb2d48b0d6623e88fd67cb6ca8e5e4074583 fix/homepage-education-accuracy
+8d50afef1e8a2b621b7344eadf475b0d609ad7a2 fix/leaflet-specificity-and-font-assertion
+b2b2cad5acf534ae7a667d3fb2be15efff37c4a7 fix/list-map-report-card-signal
+dc21e80a5e9eebd13aaab84735642eb77cef35e6 fix/mobile-cell-name-size
+e5f7f4c959f024c073472122d858333a0f24866c fix/mobile-compare-polish
+a00cbe916182d1e04661c750a1ce2e7ad27a68ae fix/mobile-sort-select-overflow
+2fd997bfe640c419a6713e85df008463de7f56e8 fix/modal-keyboard-viewport
+3e7705756776a0c44d27966dfc972023f1f38b69 fix/ofsted-link-text
+ce422e64363e2b03c186ef316832b6de15502d67 fix/promote-status-token
+4522cbf64560db1e1cd519e119aa466b42cd16a1 fix/proposed-to-close-copy
+15da060e4af37fbae919e0edf25f108266de2585 fix/rankings-admissions-accuracy
+6c872ce726f210433354ca38dc5314bc6a467534 fix/rankings-year-validation
+1c1df7796194af3d47f8e5ac0a0fbe6f323700e7 fix/report-card-chip-alignment
+b2dc4d0779ced02709429afaa5985acf4abda794 fix/results-map-ios-fullscreen
+95a5783da1fc994df76cb97238b55596dce4cd8f fix/runtime-api-proxy
+8a9ba30cc24e29653a72c03c0e817684b7db7c07 fix/sats-per-level-national
+536832a524fc4f9ed858f047e00b9a4d9d429c46 fix/school-detail-nan-500
+fef83b3bf244a9bf3cb4afa75dbc433d8725014f fix/schoolbar-sticky-offset
+b0c5b6bb57c879477da12c37b15cff950c29ebd3 fix/secondary-anchors-button-affordance
+261403bcd2b01aa4f26ee212e26302fc0f769bf9 fix/special-note-full-width
+ea5249a2ea6faf6bfa5a1522644387ba59770ec8 fix/standardise-distance-units
+e4565e9f158721d4df6b918f2b065de82851f8d9 fix/trends-chart-height
+3aad5101a842539105022f9059e85c224fc5973a fix/welsh-establishment-leak
+315f1feede70bdf3101d2fdd305b3d06e037fdac perf/batch-supplementary
+d2dc78aeb599e16b7ef5019be2b360b08df463bc perf/compare-loading
+e098ad4bd1130152705788d4773837b7d1e7112e perf/server-client-split
+
+## Deleted: superseded by PR #110 (content preserved on feat/seo-crawl-hygiene-main)
+786ec80dd4de4a3cb674a89e35b3b5e461639289 feat/seo-crawl-hygiene
+a5ac0bcd1b37bc10dcbce88f8601d01bf7b3eaaf feat/england-only-corpus

docs/superpowers/specs/2026-08-20-seo-programme-design.md

@@ -0,0 +1,309 @@
+# SEO Programme — Design
+
+Date: 2026-08-20
+Status: awaiting review
+
+## Problem
+
+schoolcompare ranks second for "school compare" — an exact match for the
+brand and the domain. It ranks poorly for "compare schools", "school
+comparison" and "schools near me". The first is a naming artefact and
+transfers to nothing; the rest are the queries that actually carry parent
+demand.
+
+The cause is structural, not editorial. The site publishes five route
+families:
+
+    /  /compare  /rankings  /admissions  /school/[slug]
+
+Location intent has no landing page at all. Every competitor outranking us
+on those queries wins with programmatic location pages:
+
+| Competitor    | URL pattern                                |
+|---------------|--------------------------------------------|
+| School Guide  | `/best-schools-in/manchester`              |
+| Locrating     | `/the-best-primary-schools-in-Manchester_…`|
+| FindMySchool  | `/best-primary-schools/manchester`         |
+| Snobe         | `/best-primary-schools/manchester`         |
+| School Atlas  | `/guides/best-primary-schools-manchester`  |
+
+"Schools near me" is a local-intent query. Google resolves it against the
+user's coordinates and serves pages that are *about a place*. A national
+homepage cannot win it. No title or description change fixes this; only
+pages Google can localise will.
+
+## Baseline (measured 2026-08-20, production API)
+
+| Measure                                    | Value   |
+|--------------------------------------------|---------|
+| Unique schools                              | 27,229  |
+| URLs in sitemap.xml                         | 27,232  |
+| Schools with 2024/25 performance data       | 21,266  |
+| Schools with **no** current performance data| ~5,963 (22%) |
+| Welsh establishments (all metrics null)     | 1,569   |
+| Overseas / offshore establishments           | 467     |
+| Static URLs in sitemap                      | 3       |
+| Routes setting a canonical                  | 1 of 5  |
+
+Three findings from that table drive the plan.
+
+**We submit ~6,000 thin pages to Google.** `build_sitemap()`
+(`backend/app.py:73`) enumerates every URN regardless of whether the school
+has any data. Welsh establishments return `school_type: "Welsh
+establishment"` with every performance metric, Ofsted grade and phase field
+null. Overseas and offshore establishments ("BFPO Overseas Establishments",
+"Gibraltar Overseas Establishments", "Jersey Offshore Establishments") are
+in the local-authority list too. At 22% of the submitted corpus this is a
+site-wide quality signal problem and a crawl-budget waste, not a rounding
+error.
+
+W1 item 4 removes 2,036 of those — every non-England establishment — taking
+the corpus to 25,193. The 3,927 that remain are English schools with no
+current data: mostly newly opened, special, nursery or alternative provision.
+Those are a template problem, not a corpus problem, and item 5 handles them
+separately.
+
+**The homepage is its own competitor.** `app/page.tsx` accepts eleven search
+params (`search`, `local_authority`, `school_type`, `phase`, `page`,
+`postcode`, `radius`, `sort`, `gender`, `admissions_policy`,
+`has_sixth_form`) and sets no canonical. Every filter combination is a
+crawlable near-duplicate of the single page we are asking to rank for
+"compare schools".
+
+**School pages are near-orphans.** Reachable from the sitemap and from site
+search, but almost nothing links to them contextually, so they accrue no
+internal authority.
+
+Also noted: the sitemap emits invented `priority` values and no `lastmod`.
+Google ignores `priority` and `changefreq` entirely; `lastmod` is the field
+it does read, and we omit it.
+
+## Keyword clusters
+
+Ranked by judgement of UK parent search behaviour and by the competitive
+SERP evidence above. Google Search Console is connected, so cluster
+priorities are to be re-derived from measured impressions before build
+starts (see Workstream 0).
+
+**C1 — Head "compare" terms.** compare schools · school comparison · school
+comparison tool · compare school performance · compare primary schools ·
+compare secondary schools · compare two schools
+
+**C2 — League tables and rankings.** primary school league tables ·
+secondary school league tables · school league tables 2026 · SATs results by
+school · GCSE results by school · KS2 league tables · Progress 8 rankings ·
+best primary schools in [town] · top 10 primary schools in [LA]
+
+**C3 — Local / near me.** schools near me · primary schools near me ·
+secondary schools near me · best schools near me · good schools near me ·
+schools in [town] · primary schools in [LA] · schools near [postcode] ·
+[postcode] school catchment
+
+**C4 — Individual school long tail.** [school] ofsted · [school] SATs
+results · [school] catchment area · [school] reviews · [school] URN
+
+**C5 — Admissions.** primary school admissions 2027 · national offer day
+2027 · school application deadline · school admissions appeal ·
+oversubscription criteria · distance criteria school admissions · didn't get
+first choice school · school admissions [LA]
+
+**C6 — Metric explainers.** what is a good SATs score · what is Progress 8 ·
+what is Attainment 8 · expected standard KS2 meaning · scaled score
+explained · Ofsted grades explained · Ofsted report cards · pupil premium
+explained
+
+**C7 — Head to head.** [school A] vs [school B] · academy vs community
+school · grammar school vs comprehensive · faith school vs community school
+
+## Workstreams
+
+### W0 — Measure before touching anything
+
+Export a Google Search Console baseline: impressions, clicks, average
+position and CTR by query and by page, for the trailing 16 months. Bucket
+queries into C1–C7. This sets the counterfactual — without it, no later
+claim about lift is defensible, because school-search traffic is strongly
+seasonal (results day in December, offer day in March/April).
+
+Re-rank C1–C7 against measured impressions and adjust the sequence below if
+the data disagrees with the judgement calls.
+
+### W1 — Crawl hygiene and index sanity
+
+Cheap, and it unblocks everything after it. Adding 5,000 pages on top of a
+corpus that is 22% thin would compound the existing problem.
+
+1. Canonical on every route. `/`, `/rankings`, `/compare` and `/admissions`
+   currently set none.
+2. The homepage canonicalises to `/` regardless of search params.
+3. `/compare?urns=…` gets `noindex, follow` — it is an unbounded parameter
+   space with no standalone value.
+4. **England only — DONE.** Wales, the Crown Dependencies, Gibraltar and the
+   service/overseas schools are removed from the corpus at the mart boundary,
+   not hidden at the view layer. `dim_school` and `dim_location` both exclude
+   `TypeOfEstablishment` in {25, 26, 30, 37} — Offshore schools, Service
+   children's education, Welsh establishment, British schools overseas —
+   listed once as `vars.non_england_school_type_codes` in `dbt_project.yml`.
+   That removes 2,036 establishments and 29 local authorities, and because
+   `build_sitemap()` reads the same marts, it drops them from the sitemap in
+   the same stroke. `assert_england_only_schools` fails the pipeline if a GIAS
+   refresh reintroduces them or if the two models drift apart.
+5. Prune the remaining thin pages: exclude any school with no performance data
+   **and** no Ofsted record. Distinct from item 4 — these are English schools
+   with nothing yet to show, so the fix may be a better template rather than
+   removal.
+6. Rebuild the sitemap as a sitemap **index**: one child per page family,
+   real `lastmod` from the data-load timestamp, `priority` and `changefreq`
+   dropped.
+
+### W2 — The location layer
+
+The dominant lever. `dim_location` already carries `town`, `county`,
+`local_authority_name`, `parliamentary_constituency`, `latitude`,
+`longitude` and `postcode`, so no new ingestion is required.
+
+Routes:
+
+    /schools/[la]                     e.g. /schools/manchester
+    /schools/[la]/primary
+    /schools/[la]/secondary
+    /best-primary-schools/[town]
+    /best-secondary-schools/[town]
+    /schools/near/[outcode]           e.g. /schools/near/m20
+    /schools/near-me                  geolocating hub
+
+**Thin-page threshold: generate a town or outcode page only where at least
+five schools have current performance data.** Below that, 301 to the parent
+LA page. This is the single most important constraint in the workstream —
+it is what separates a location layer from index bloat.
+
+Each page must earn its place with content a parent would actually use, not
+a template shell:
+
+- H1 matching the query intent ("Best primary schools in Manchester")
+- Counts framed usefully: "137 primary schools, 9 rated Outstanding"
+- A ranked table of the top 20 on the headline metric
+- Local average against the England average
+- Ofsted grade distribution
+- Map
+- Links to neighbouring towns and to the parent LA
+- An FAQ block (feeds `FAQPage` in W4)
+- Links to every school page in scope — this is what de-orphans W1's corpus
+
+Sizing estimate: ~150 usable LAs × 3 ≈ 450; towns clearing the threshold
+≈ 1,200 × 2 ≈ 2,400; outcodes ≈ 2,300. Roughly **5,000 new pages**,
+comfortably inside a sitemap index and well under the per-file 50,000 limit.
+
+### W3 — Make rankings indexable
+
+`/rankings` is driven entirely by query params, so Google indexes
+approximately one page where there should be hundreds.
+
+    /rankings/[phase]/[metric]
+    /rankings/[phase]/[metric]/[la]
+
+The interactive filter UI stays; its state moves into real paths. Param
+forms canonicalise to the clean path. This is the direct play for C2.
+
+### W4 — Structured data and internal linking
+
+- Replace the bare `EducationalOrganization` on school pages with `School`,
+  and populate it properly.
+- `BreadcrumbList` site-wide.
+- `ItemList` on every rankings and location page.
+- `FAQPage` on admissions and on location pages.
+- New school-page modules: "Other schools in [town]", "Nearby schools",
+  "Compare with similar schools". Each links out to W2 and W3 pages, which
+  is what circulates authority instead of stranding it.
+
+Explicitly **not** doing `Dataset` or `AggregateRating` — no review corpus
+exists, and fabricating one would be both useless and a policy violation.
+
+### W5 — Admissions expansion
+
+One static page currently carries an entire cluster.
+
+    /admissions/[la]                  per-authority deadlines and offer day
+    /admissions/appeals
+    /admissions/national-offer-day
+
+`school admissions [LA]` is high-intent and highly seasonal; per-authority
+pages are the natural unit.
+
+### W6 — Explainer content
+
+    /guides/progress-8
+    /guides/attainment-8
+    /guides/sats-scaled-scores
+    /guides/ofsted-grades
+    /guides/expected-standard
+
+Each links into the corresponding W3 rankings page. Cheap to build, and it
+is what gives the metric vocabulary enough topical weight to support C1–C4.
+
+### W7 — Head-to-head pages
+
+    /compare/[school-a]-vs-[school-b]
+
+C7 is uncontested and native to the product. It is also the easiest way to
+destroy everything W1 fixes: 27,229 schools generate 370 million pairs.
+**Curated pairs only** — same town, both with current data, both with real
+search demand — capped in the low thousands. Gated behind evidence that W2
+is indexing cleanly.
+
+### W8 — Metadata rewrite for C1
+
+Current homepage title is `schoolcompare | Compare every school in England`,
+which spends the most valuable position on the brand. Rewrite the homepage,
+rankings and compare titles and descriptions around C1 phrasing. Small
+change, and the cheapest item in the programme.
+
+## Sequencing
+
+    W0 → W1 → W2 → W3 → W4 → W5 → W6 → (W7 if W2 indexes cleanly)
+
+W1 before W2 is not negotiable: adding pages to a corpus that is 22% thin
+compounds the problem rather than diluting it.
+
+This spec is a programme, not a single implementation plan. Each workstream
+gets its own plan and its own PR; W2 will likely need several. Only W0 and W1
+are ready to plan against today — the rest should be re-read after W0's
+Search Console baseline lands, because that data may reorder them.
+
+## Testing
+
+Per CLAUDE.md, user-facing behaviour changes extend the `e2e/` journeys in
+the same PR. Each workstream adds:
+
+- W1: canonical present and correct on every route; `/compare?urns=` carries
+  `noindex`; sitemap excludes a known dataless URN.
+- W2: a known LA, town and outcode page renders with the expected school
+  count; a below-threshold town redirects to its LA.
+- W3: a clean rankings path renders; the param form canonicalises to it.
+- W4: JSON-LD parses and validates against the declared types.
+
+## Risks
+
+**Index bloat.** The failure mode of every programmatic SEO programme. The
+five-school threshold, the W1 prune and the W7 gate are the three controls.
+
+**Helpful-content exposure.** Google's stance on templated location pages
+has hardened. The mitigation is that each page carries genuinely local
+computed data — real counts, real distributions, real local-vs-national
+comparison — rather than a name substituted into boilerplate.
+
+**Build cost.** School pages already use ISR with a 7-day revalidate and
+`PRERENDER_SCHOOLS` gating full prerender. 5,000 more routes need the same
+treatment; full static generation of 32,000 pages is likely impractical in
+CI.
+
+**Seasonality.** Results day and offer day dominate the traffic curve.
+Judging the programme on a mid-summer window would misread it in either
+direction. W0's baseline must be year-on-year, not month-on-month.
+
+## Open questions
+
+1. Catchment areas are Locrating's moat and a strong C3 driver
+   (`[postcode] school catchment`). `fact_admissions` carries admission
+   distances. Is deriving approximate catchment a later workstream, or out
+   of scope?

e2e/tests/journeys.spec.ts

@@ -1473,3 +1473,256 @@ test('no single section dominates the height of a school page', async ({ page })
       + sections.map((s) => `${s.id}=${s.h}`).join(', '),
   ).toBeLessThan(2.5);
 });
+
+/*
+ * England-only corpus.
+ *
+ * GIAS ships the whole UK plus overseas and offshore establishments, none of
+ * which carry comparable DfE performance data. dim_school and dim_location
+ * exclude them (vars.non_england_school_type_codes in dbt_project.yml), which
+ * keeps them out of the site, the filter lists and the sitemap together.
+ *
+ * These assert the published surface, not the warehouse: the dbt test
+ * assert_england_only_schools guards the marts, and these guard what the
+ * environment actually serves once the marts have been rebuilt.
+ */
+
+const WELSH_AUTHORITIES = [
+  'Blaenau Gwent', 'Bridgend', 'Caerphilly', 'Cardiff', 'Carmarthenshire',
+  'Ceredigion', 'Conwy', 'Denbighshire', 'Flintshire', 'Gwynedd',
+  'Isle of Anglesey', 'Merthyr Tydfil', 'Monmouthshire', 'Neath Port Talbot',
+  'Newport', 'Pembrokeshire', 'Powys', 'Rhondda Cynon Taf', 'Swansea',
+  'Torfaen', 'Vale of Glamorgan', 'Wrexham',
+];
+
+const NON_ENGLAND_AUTHORITIES = [
+  'BFPO Overseas Establishments', 'Fieldwork Overseas Establishments',
+  'Gibraltar Overseas Establishments', 'Guernsey Offshore Establishments',
+  'Isle of Man Offshore Establishments', 'Jersey Offshore Establishments',
+  'Scotland Offshore Establishments',
+];
+
+const NON_ENGLAND_TYPES = [
+  'Welsh establishment', 'Offshore schools',
+  "Service children's education", 'British schools overseas',
+];
+
+test('the local authority filter offers no Welsh or overseas authority', async ({ page }) => {
+  const res = await page.request.get('/api/filters');
+  expect(res.ok()).toBeTruthy();
+  const { local_authorities: las } = await res.json();
+
+  expect(Array.isArray(las)).toBeTruthy();
+  // Guards against the list being empty, which would pass the check below
+  // for the wrong reason.
+  expect(las.length).toBeGreaterThan(100);
+
+  const leaked = [...WELSH_AUTHORITIES, ...NON_ENGLAND_AUTHORITIES]
+    .filter((la) => las.includes(la));
+  expect(leaked, `non-England authorities still offered: ${leaked.join(', ')}`)
+    .toEqual([]);
+});
+
+test('the school type filter offers no non-England establishment type', async ({ page }) => {
+  const res = await page.request.get('/api/filters');
+  expect(res.ok()).toBeTruthy();
+  const { school_types: types } = await res.json();
+
+  expect(Array.isArray(types)).toBeTruthy();
+  expect(types.length).toBeGreaterThan(10);
+
+  const leaked = NON_ENGLAND_TYPES.filter((t) => types.includes(t));
+  expect(leaked, `non-England types still offered: ${leaked.join(', ')}`)
+    .toEqual([]);
+});
+
+test('searching a Welsh authority by name returns no schools', async ({ page }) => {
+  // Cardiff held 144 Welsh establishments and nothing else, so the authority
+  // should now be absent from the corpus entirely rather than merely thinned.
+  const res = await page.request.get('/api/schools?local_authority=Cardiff&page_size=1');
+  expect(res.ok()).toBeTruthy();
+  const body = await res.json();
+  expect(body.total ?? (body.schools ?? []).length).toBe(0);
+});
+
+test('a Welsh school GIAS does not type as Welsh is gone too', async ({ page }) => {
+  // Beechwood College, Sully, CF64 5SE. GIAS types it "Special post 16
+  // institution" (32), not "Welsh establishment" (30), so the establishment
+  // type filter alone left it behind — it was the last Welsh school standing.
+  const res = await page.request.get('/api/schools?local_authority=Vale%20of%20Glamorgan&page_size=1');
+  expect(res.ok()).toBeTruthy();
+  const body = await res.json();
+  expect(body.total ?? (body.schools ?? []).length).toBe(0);
+
+  const page404 = await page.goto('/school/142458-beechwood-college');
+  expect(page404?.status()).toBe(404);
+});
+
+test('English schools with Welsh postcodes are kept', async ({ page }) => {
+  // Royal Mail postcode areas straddle the border: these Gloucestershire
+  // schools carry NP16/NP25 postcodes but are English schools with English
+  // data. A postcode-based filter would have wrongly deleted them, so this
+  // guards the fix from being "simplified" into one.
+  for (const urn of [115639, 115655, 137382]) {
+    const res = await page.request.get(`/api/schools/${urn}`);
+    expect(res.ok(), `URN ${urn} should still resolve`).toBeTruthy();
+    const { school_info } = await res.json();
+    expect(school_info.local_authority).toBe('Gloucestershire');
+  }
+});
+
+test('a Welsh school URL 404s while an English one still resolves', async ({ page }) => {
+  // Paired on purpose: the Welsh assertion alone would also pass if the whole
+  // site were down, which is the failure this test most needs to distinguish.
+  const english = await page.request.get('/api/schools?search=primary&per_page=1');
+  expect(english.ok()).toBeTruthy();
+  const [first] = (await english.json()).schools ?? [];
+  expect(first, 'no English school available to compare against').toBeTruthy();
+
+  const good = await page.goto(`/school/${first.urn}-x`);
+  expect(good?.status(), 'an English school should still resolve').toBeLessThan(400);
+
+  // Adamsdown Primary School, Cardiff — a Welsh establishment (URN 401559).
+  const welsh = await page.goto('/school/401559-adamsdown-primary-school');
+  expect(welsh?.status(), 'a Welsh school should no longer resolve').toBe(404);
+});
+
+async function sitemapChildren(page: Page): Promise<string[]> {
+  const res = await page.request.get('/sitemap.xml');
+  expect(res.ok()).toBeTruthy();
+  const index = await res.text();
+  expect(index).toContain('<sitemapindex');
+  return [...index.matchAll(/<loc>([^<]+)<\/loc>/g)].map((m) => m[1]);
+}
+
+test('the sitemap index names children that all resolve', async ({ page }) => {
+  const index = await (await page.request.get('/sitemap.xml')).text();
+  // An index holds <sitemap> entries only; mixing in <url> is invalid.
+  expect(index).not.toContain('<url>');
+
+  const locs = await sitemapChildren(page);
+  expect(locs.length).toBeGreaterThanOrEqual(2);
+
+  for (const loc of locs) {
+    expect(loc.startsWith('https://www.schoolcompare.co.uk/sitemaps/')).toBeTruthy();
+    const child = await page.request.get(new URL(loc).pathname);
+    expect(child.ok(), `${loc} should resolve`).toBeTruthy();
+    expect(await child.text()).toContain('<urlset');
+  }
+});
+
+test('the sitemap submits no Welsh or overseas school', async ({ page }) => {
+  const locs = await sitemapChildren(page);
+
+  let total = 0;
+  for (const loc of locs) {
+    const xml = await (await page.request.get(new URL(loc).pathname)).text();
+    total += (xml.match(/<url>/g) ?? []).length;
+    // 401559 (Adamsdown, Cardiff) and 402426 (ACT Schools, Cardiff) were both
+    // submitted before the England-only filter landed.
+    expect(xml).not.toContain('/school/401559');
+    expect(xml).not.toContain('/school/402426');
+  }
+  expect(total, 'sitemap looks empty or truncated').toBeGreaterThan(1000);
+});
+
+test('the sitemap invents no priority or changefreq', async ({ page }) => {
+  const [first] = await sitemapChildren(page);
+  expect(first).toBeTruthy();
+
+  const xml = await (await page.request.get(new URL(first).pathname)).text();
+  // Google ignores both. They were noise dressed as signal.
+  expect(xml).not.toContain('<priority>');
+  expect(xml).not.toContain('<changefreq>');
+});
+
+/*
+ * Canonical URLs (spec 2026-08-20, W1).
+ *
+ * Every indexable route declares exactly one canonical, on the www host, with
+ * no query string. The homepage's eleven search params filter a result set
+ * rather than making a new document, so they all collapse onto "/".
+ */
+const CANONICAL_ROUTES: Array<[string, string]> = [
+  ['/', 'https://www.schoolcompare.co.uk/'],
+  ['/rankings', 'https://www.schoolcompare.co.uk/rankings'],
+  ['/admissions', 'https://www.schoolcompare.co.uk/admissions'],
+];
+
+for (const [path, expected] of CANONICAL_ROUTES) {
+  test(`${path} declares exactly one canonical, on the www host`, async ({ page }) => {
+    await page.goto(path);
+    const hrefs = await page.locator('link[rel="canonical"]').evaluateAll(
+      (els) => els.map((e) => e.getAttribute('href')));
+    expect(hrefs, `${path} should declare one canonical`).toHaveLength(1);
+    expect(hrefs[0]).toBe(expected);
+  });
+}
+
+test('a filtered homepage still canonicalises to the bare root', async ({ page }) => {
+  await page.goto('/?search=primary&phase=primary&sort=name&page=2');
+  const href = await page.locator('link[rel="canonical"]').first()
+    .getAttribute('href');
+  expect(href).toBe('https://www.schoolcompare.co.uk/');
+});
+
+test('a school page canonicalises to its own slug on the www host', async ({ page }) => {
+  const res = await page.request.get('/api/schools?search=primary&per_page=1');
+  expect(res.ok()).toBeTruthy();
+  const [first] = (await res.json()).schools ?? [];
+  expect(first, 'no school available').toBeTruthy();
+
+  await page.goto(`/school/${first.urn}-x`);
+  const href = await page.locator('link[rel="canonical"]').first()
+    .getAttribute('href');
+  expect(href).toMatch(/^https:\/\/www\.schoolcompare\.co\.uk\/school\/\d+-/);
+});
+
+test('a bare /compare is indexable, a parameterised one is not', async ({ page }) => {
+  await page.goto('/compare');
+  await expect(page.locator('meta[name="robots"]')).toHaveCount(0);
+
+  const [a, b] = await twoPrimaryUrns(page);
+  await page.goto(`/compare?urns=${a},${b}`);
+  const robots = await page.locator('meta[name="robots"]').first()
+    .getAttribute('content');
+  expect(robots).toContain('noindex');
+  expect(robots).toContain('follow');
+
+  // noindex but follow: the links out to each school page still count, so the
+  // canonical must still be present and point at the bare path.
+  const canonical = await page.locator('link[rel="canonical"]').first()
+    .getAttribute('href');
+  expect(canonical).toBe('https://www.schoolcompare.co.uk/compare');
+});
+
+/*
+ * Staging must not be indexable (spec 2026-08-20, W1 hygiene).
+ *
+ * These journeys only ever run against staging — deploy.yml passes
+ * STAGING_BASE_URL, and promote.yml only smoke-polls production without
+ * Playwright — so asserting the noindex header here is safe.
+ */
+test('staging answers noindex, and stays crawlable so the noindex is seen', async ({ page }) => {
+  const res = await page.request.get('/');
+  expect(res.ok()).toBeTruthy();
+
+  const tag = res.headers()['x-robots-tag'];
+  expect(tag, 'staging must send X-Robots-Tag').toBeTruthy();
+  expect(tag).toContain('noindex');
+
+  // The other half, and the reason this is one test rather than two: a
+  // Disallow would stop Google fetching the page at all, so it would never
+  // see the noindex above. The two only work together.
+  const robots = await (await page.request.get('/robots.txt')).text();
+  expect(robots).not.toMatch(/^\s*Disallow:\s*\/\s*$/mi);
+});
+
+test('a school page on staging is noindexed too, not just the homepage', async ({ page }) => {
+  const list = await page.request.get('/api/schools?search=primary&per_page=1');
+  const [first] = (await list.json()).schools ?? [];
+  expect(first, 'no school available').toBeTruthy();
+
+  const res = await page.request.get(`/school/${first.urn}-x`);
+  expect(res.headers()['x-robots-tag']).toContain('noindex');
+});

nextjs-app/__tests__/app/metadata.test.ts

@@ -0,0 +1,53 @@
+import { metadata as homeMetadata } from '@/app/page';
+import { metadata as rankingsMetadata } from '@/app/rankings/page';
+import { metadata as admissionsMetadata } from '@/app/admissions/page';
+import { generateMetadata as compareMetadata } from '@/app/compare/page';
+
+describe('canonical URLs', () => {
+  it('the homepage canonicalises to the bare root', () => {
+    // page.tsx reads eleven search params. Without this, every filter
+    // combination is a crawlable near-duplicate of the one page we want to
+    // rank for "compare schools".
+    expect(homeMetadata.alternates?.canonical)
+      .toBe('https://www.schoolcompare.co.uk/');
+  });
+
+  it('rankings canonicalises to the bare path', () => {
+    expect(rankingsMetadata.alternates?.canonical)
+      .toBe('https://www.schoolcompare.co.uk/rankings');
+  });
+
+  it('admissions canonicalises to the bare path', () => {
+    expect(admissionsMetadata.alternates?.canonical)
+      .toBe('https://www.schoolcompare.co.uk/admissions');
+  });
+});
+
+describe('/compare indexability', () => {
+  it('the bare compare page is indexable and canonical to itself', async () => {
+    // This is the landing page for the "compare schools" head term.
+    const meta = await compareMetadata({ searchParams: Promise.resolve({}) });
+    expect(meta.alternates?.canonical)
+      .toBe('https://www.schoolcompare.co.uk/compare');
+    expect(meta.robots).toBeUndefined();
+  });
+
+  it('a comparison of specific schools is noindex, follow', async () => {
+    // ~317 million pairs before triples. Indexing the parameter space would
+    // swamp everything else in the corpus.
+    const meta = await compareMetadata({
+      searchParams: Promise.resolve({ urns: '100001,100002' }),
+    });
+    expect(meta.robots).toEqual({ index: false, follow: true });
+  });
+
+  it('a parameterised comparison still canonicalises to the bare path', async () => {
+    // follow:true plus a canonical means the outbound links to each school
+    // page still pass value even though this URL is not indexed.
+    const meta = await compareMetadata({
+      searchParams: Promise.resolve({ urns: '100001,100002' }),
+    });
+    expect(meta.alternates?.canonical)
+      .toBe('https://www.schoolcompare.co.uk/compare');
+  });
+});

nextjs-app/__tests__/lib/site.test.ts

@@ -0,0 +1,27 @@
+import { SITE_URL, absoluteUrl } from '@/lib/site';
+
+describe('SITE_URL', () => {
+  it('is the www host, which is the one that serves a 200', () => {
+    // The apex 301s to www at Cloudflare. A canonical pointing at a redirect
+    // is a wasted signal, so every absolute URL we emit must already be www.
+    expect(SITE_URL).toBe('https://www.schoolcompare.co.uk');
+  });
+
+  it('has no trailing slash, so joins never double up', () => {
+    expect(SITE_URL.endsWith('/')).toBe(false);
+  });
+});
+
+describe('absoluteUrl', () => {
+  it('joins a rooted path', () => {
+    expect(absoluteUrl('/rankings')).toBe('https://www.schoolcompare.co.uk/rankings');
+  });
+
+  it('joins a path missing its leading slash', () => {
+    expect(absoluteUrl('rankings')).toBe('https://www.schoolcompare.co.uk/rankings');
+  });
+
+  it('maps the site root to a bare trailing slash', () => {
+    expect(absoluteUrl('/')).toBe('https://www.schoolcompare.co.uk/');
+  });
+});

nextjs-app/app/admissions/page.tsx

@@ -1,3 +1,4 @@
+import { absoluteUrl } from '@/lib/site';
 import type { Metadata } from 'next';
 import { AdmissionsView } from '@/components/AdmissionsView';
 
@@ -7,6 +8,7 @@ export const metadata: Metadata = {
   title: 'School Admissions Guide',
   description:
     'Understand the Primary and Secondary school admissions process in England, with live countdowns to every key deadline and National Offer Day.',
+  alternates: { canonical: absoluteUrl('/admissions') },
 };
 
 export default function AdmissionsPage() {

nextjs-app/app/compare/page.tsx

@@ -5,6 +5,7 @@
 
 import { fetchComparison, fetchMetrics } from '@/lib/api';
 import { ComparisonView } from '@/components/ComparisonView';
+import { absoluteUrl } from '@/lib/site';
 import type { Metadata } from 'next';
 
 interface ComparePageProps {
@@ -14,14 +15,34 @@ interface ComparePageProps {
   }>;
 }
 
-export const metadata: Metadata = {
+/**
+ * Indexability depends on the query string, so this cannot be a static export.
+ *
+ * Bare /compare is the landing page for the "compare schools" head term and
+ * stays indexable. /compare?urns=… is an unbounded parameter space — 25,193
+ * schools make ~317 million pairs — so it goes noindex. It stays `follow` and
+ * keeps a canonical to the bare path, so the links out to each school page
+ * still count.
+ */
+export async function generateMetadata(
+  { searchParams }: ComparePageProps,
+): Promise<Metadata> {
+  const { urns } = await searchParams;
+
+  const base: Metadata = {
     title: 'Compare Schools',
     description:
       'Compare schools in England side by side — Ofsted inspections, KS2 and GCSE results against the England average, admissions odds and school community.',
     keywords:
       'school comparison, compare schools, Ofsted comparison, school admissions, KS2 comparison, primary school performance',
+    alternates: { canonical: absoluteUrl('/compare') },
   };
 
+  if (!urns) return base;
+
+  return { ...base, robots: { index: false, follow: true } };
+}
+
 // Dynamic via searchParams; remove force-dynamic so internal data fetches
 // can still use Next.js's per-call revalidate cache.
 

nextjs-app/app/layout.tsx

@@ -5,6 +5,7 @@ import { Navigation } from '@/components/Navigation';
 import { Footer } from '@/components/Footer';
 import { ComparisonToast } from '@/components/ComparisonToast';
 import { ComparisonProvider } from '@/context/ComparisonProvider';
+import { SITE_URL } from '@/lib/site';
 import './globals.css';
 
 // Manrope carries headings and key messaging — the guideline's "friendly,
@@ -57,12 +58,12 @@ export const metadata: Metadata = {
   // No `icons` key on purpose: setting it here would override the file
   // conventions. app/icon.svg and app/apple-icon.tsx are the source, and
   // app/opengraph-image.tsx supplies og:image and twitter:image.
-  metadataBase: new URL('https://schoolcompare.co.uk'),
+  metadataBase: new URL(SITE_URL),
   openGraph: {
     type: 'website',
     title: 'schoolcompare | Compare School Performance',
     description: 'Compare primary and secondary school SATs and GCSE performance across England',
-    url: 'https://schoolcompare.co.uk',
+    url: SITE_URL,
     siteName: 'schoolcompare',
   },
   twitter: {

nextjs-app/app/page.tsx

@@ -3,6 +3,7 @@
  * Main landing page with school search and browsing
  */
 
+import { absoluteUrl } from '@/lib/site';
 import type { Metadata } from 'next';
 import { fetchSchools, fetchFilters, fetchDataInfo } from '@/lib/api';
 import { formatAcademicYear } from '@/lib/utils';
@@ -35,6 +36,10 @@ interface HomePageProps {
 export const metadata: Metadata = {
   title: { absolute: 'schoolcompare | Compare every school in England' },
   description: 'Search and compare school performance across England',
+  // This page reads eleven search params. They filter a result set; they do
+  // not make a new document. Collapsing every combination onto "/" stops the
+  // homepage competing with itself for its own head terms.
+  alternates: { canonical: absoluteUrl('/') },
 };
 
 // The page reads searchParams, which makes rendering dynamic by default.

nextjs-app/app/rankings/page.tsx

@@ -5,6 +5,7 @@
 
 import { fetchRankings, fetchFilters, fetchMetrics } from '@/lib/api';
 import { RankingsView } from '@/components/RankingsView';
+import { absoluteUrl } from '@/lib/site';
 import type { Metadata } from 'next';
 
 interface RankingsPageProps {
@@ -20,6 +21,9 @@ export const metadata: Metadata = {
   title: 'School Rankings',
   description: 'Top-ranked schools by SATs and GCSE performance across England',
   keywords: 'school rankings, top schools, best schools, KS2 rankings, KS4 rankings, school league tables',
+  // Param forms (?metric=&local_authority=&year=&phase=) collapse here for
+  // now. W3 replaces them with real indexable paths.
+  alternates: { canonical: absoluteUrl('/rankings') },
 };
 
 // Dynamic via searchParams; remove force-dynamic so internal data fetches

nextjs-app/app/robots.ts

@@ -4,6 +4,7 @@
  */
 
 import { MetadataRoute } from 'next';
+import { absoluteUrl } from '@/lib/site';
 
 export default function robots(): MetadataRoute.Robots {
   return {
@@ -14,6 +15,6 @@ export default function robots(): MetadataRoute.Robots {
         disallow: ['/api/', '/_next/'],
       },
     ],
-    sitemap: 'https://schoolcompare.co.uk/sitemap.xml',
+    sitemap: absoluteUrl('/sitemap.xml'),
   };
 }

nextjs-app/app/school/[slug]/page.tsx

@@ -15,6 +15,7 @@ import {
 } from '@/lib/schoolSections';
 import { parseSchoolSlug, schoolUrl } from '@/lib/utils';
 import type { NationalAverages } from '@/lib/types';
+import { absoluteUrl } from '@/lib/site';
 import type { Metadata } from 'next';
 
 /**
@@ -97,7 +98,7 @@ export async function generateMetadata({ params }: SchoolPageProps): Promise<Met
         title,
         description,
         type: 'website',
-        url: `https://schoolcompare.co.uk${canonicalPath}`,
+        url: absoluteUrl(canonicalPath),
         siteName: 'schoolcompare',
       },
       twitter: {
@@ -106,7 +107,7 @@ export async function generateMetadata({ params }: SchoolPageProps): Promise<Met
         description,
       },
       alternates: {
-        canonical: `https://schoolcompare.co.uk${canonicalPath}`,
+        canonical: absoluteUrl(canonicalPath),
       },
     };
   } catch {

nextjs-app/app/sitemap.xml/route.ts

@@ -1,32 +1,8 @@
-/**
- * Runtime proxy for /sitemap.xml → the FastAPI backend's generated sitemap.
- *
- * Like the /api/* proxy, this reads FASTAPI_URL at request time rather than
- * baking the backend host into the build, so one image works in every
- * environment. robots.ts points crawlers here.
- */
-
-import { NextResponse } from 'next/server';
+import { proxySitemap } from '@/lib/sitemapProxy';
 
 export const dynamic = 'force-dynamic';
 export const runtime = 'nodejs';
 
-function backendOrigin(): string {
-  const base = process.env.FASTAPI_URL || process.env.NEXT_PUBLIC_API_URL || 'http://localhost:8000/api';
-  return base.replace(/\/api$/, '');
-}
-
 export async function GET() {
-  let upstream: Response;
-  try {
-    upstream = await fetch(`${backendOrigin()}/sitemap.xml`, { cache: 'no-store' });
-  } catch {
-    return new NextResponse('Sitemap temporarily unavailable', { status: 502 });
-  }
-
-  const body = await upstream.text();
-  return new NextResponse(body, {
-    status: upstream.status,
-    headers: { 'content-type': upstream.headers.get('content-type') || 'application/xml' },
-  });
+  return proxySitemap('/sitemap.xml');
 }

nextjs-app/app/sitemaps/[...parts]/route.ts

@@ -0,0 +1,24 @@
+import { NextResponse } from 'next/server';
+import { proxySitemap } from '@/lib/sitemapProxy';
+
+export const dynamic = 'force-dynamic';
+export const runtime = 'nodejs';
+
+/**
+ * Children are /sitemaps/static.xml and /sitemaps/schools-{n}.xml. The name is
+ * validated here rather than passed through, so this route cannot be used to
+ * reach arbitrary backend paths.
+ */
+const CHILD = /^(static|schools-\d+)\.xml$/;
+
+export async function GET(
+  _request: Request,
+  { params }: { params: Promise<{ parts: string[] }> },
+) {
+  const { parts } = await params;
+  const name = parts.join('/');
+  if (!CHILD.test(name)) {
+    return new NextResponse('Not found', { status: 404 });
+  }
+  return proxySitemap(`/sitemaps/${name}`);
+}

nextjs-app/lib/site.ts

@@ -0,0 +1,19 @@
+/**
+ * The one place the site's absolute origin is written down.
+ *
+ * The apex domain 301s to www at Cloudflare, so www is the host that actually
+ * serves a 200. Canonicals, og:url, sitemap <loc> entries and the robots.txt
+ * Sitemap: line must all agree with it — a canonical pointing at a redirect
+ * makes Google resolve the hop before it can consolidate the signal.
+ *
+ * backend/app.py holds the same value as BASE_URL for the sitemap. The two are
+ * asserted against each other by the e2e journeys rather than shared at build
+ * time, because the backend and frontend ship as separate images.
+ */
+export const SITE_URL = 'https://www.schoolcompare.co.uk';
+
+/** Absolute URL for a site-relative path. Tolerates a missing leading slash. */
+export function absoluteUrl(path: string): string {
+  const rooted = path.startsWith('/') ? path : `/${path}`;
+  return `${SITE_URL}${rooted}`;
+}

nextjs-app/lib/sitemapProxy.ts

@@ -0,0 +1,29 @@
+/**
+ * Runtime proxy for the sitemap family → the FastAPI backend.
+ *
+ * Like the /api/* proxy, this reads FASTAPI_URL at request time rather than
+ * baking the backend host into the build, so one image works in every
+ * environment. robots.ts points crawlers at /sitemap.xml, which is the index;
+ * the index names children under /sitemaps/, which land on the same proxy.
+ */
+import { NextResponse } from 'next/server';
+
+function backendOrigin(): string {
+  const base = process.env.FASTAPI_URL || process.env.NEXT_PUBLIC_API_URL || 'http://localhost:8000/api';
+  return base.replace(/\/api$/, '');
+}
+
+export async function proxySitemap(path: string): Promise<NextResponse> {
+  let upstream: Response;
+  try {
+    upstream = await fetch(`${backendOrigin()}${path}`, { cache: 'no-store' });
+  } catch {
+    return new NextResponse('Sitemap temporarily unavailable', { status: 502 });
+  }
+
+  const body = await upstream.text();
+  return new NextResponse(body, {
+    status: upstream.status,
+    headers: { 'content-type': upstream.headers.get('content-type') || 'application/xml' },
+  });
+}

nextjs-app/next.config.js

@@ -55,6 +55,37 @@ const nextConfig = {
   // Headers for caching and security
   async headers() {
     return [
+      {
+        /*
+         * Keep non-production hosts out of the index.
+         *
+         * Staging serves the same image as production off stx., so without
+         * this it is a full crawlable duplicate of the site.
+         *
+         * X-Robots-Tag, NOT a robots.txt Disallow. Disallow blocks crawling,
+         * which is not the same as blocking indexing — a disallowed URL can
+         * still be indexed from external links, and worse, blocking the crawl
+         * means Google never fetches the page and never sees a noindex at all.
+         * Staging therefore stays crawlable and answers "noindex" when crawled.
+         *
+         * Matched on the staging host explicitly rather than "any host that is
+         * not production". The inverted form is tempting because it would cover
+         * future environments automatically, but its failure mode is
+         * deindexing production if the Host header ever arrives rewritten by a
+         * proxy. This form's failure mode is a new environment being indexable
+         * until someone adds it here — recoverable, where the other is not.
+         *
+         * Any new non-production hostname must be added to this list.
+         */
+        source: '/:path*',
+        has: [{ type: 'host', value: 'stx.schoolcompare.co.uk' }],
+        headers: [
+          {
+            key: 'X-Robots-Tag',
+            value: 'noindex, nofollow',
+          },
+        ],
+      },
       {
         source: '/:path*',
         headers: [

pipeline/transform/dbt_project.yml

@@ -11,6 +11,32 @@ seed-paths: ["seeds"]
 target-path: "target"
 clean-targets: ["target", "dbt_packages"]
 
+# schoolcompare publishes England only. GIAS ships the whole UK plus overseas
+# and offshore establishments, none of which have comparable DfE performance
+# data — Wales does not publish on the English measures at all. Excluding them
+# at the mart boundary keeps them out of the site, the filter lists and the
+# sitemap together. Codes are TypeOfEstablishment, see seeds/gias_code_names.csv.
+#   25 = Offshore schools            (Jersey, Guernsey, Isle of Man, Gibraltar)
+#   26 = Service children's education (BFPO, Fieldwork Overseas)
+#   30 = Welsh establishment
+#   37 = British schools overseas
+# dim_school, dim_location and assert_england_only_schools all read this list.
+vars:
+  non_england_school_type_codes: [25, 26, 30, 37]
+
+  # Type alone is not enough. Beechwood College (URN 142458, Sully, CF64 5SE)
+  # is a Welsh school that GIAS types as a Special post 16 institution (32)
+  # rather than a Welsh establishment (30), so the type filter let it through.
+  # LA code is the reliable discriminator: GIAS gives the 22 Welsh unitary
+  # authorities the contiguous block 660-681, which English authorities never
+  # use (they occupy 201-384 and 800-938).
+  #
+  # Filtering on postcode would be wrong: Redbrook, Tutshill, Wyedean and two
+  # other Gloucestershire schools carry NP16/NP25 postcodes because Royal Mail
+  # areas straddle the border, and they are English schools with English data.
+  welsh_la_code_min: 660
+  welsh_la_code_max: 681
+
 models:
   school_compare:
     staging:

pipeline/transform/models/marts/dim_location.sql

@@ -31,5 +31,11 @@ select
         else null
     end as longitude
 from {{ ref('stg_gias_establishments') }} s
--- Must match dim_school's status filter exactly (the API inner-joins the two).
+-- Must match dim_school's status and England filters exactly (the API
+-- inner-joins the two).
 where s.status_code in (1, 3)
+-- coalesce, not a bare NOT IN: a null type code would make the predicate
+-- null and drop the row silently. Unknown type is not grounds for exclusion.
+and coalesce(s.school_type_code, -1) not in ({{ var('non_england_school_type_codes') | join(', ') }})
+-- Welsh schools GIAS does not type as Welsh establishments (see vars).
+and coalesce(s.local_authority_code, -1) not between {{ var('welsh_la_code_min') }} and {{ var('welsh_la_code_max') }}

pipeline/transform/models/marts/dim_school.sql

@@ -91,3 +91,10 @@ left join {{ ref('int_ofsted_latest') }} o on s.urn = o.urn
 -- 1 = Open; 3 = Open, but proposed to close (still operating; drops out when
 -- GIAS flips to Closed — marts fully rebuild each run).
 where s.status_code in (1, 3)
+-- England only. dim_location must apply this filter identically (the API
+-- inner-joins the two). See vars in dbt_project.yml for what the codes are.
+-- coalesce, not a bare NOT IN: a null type code would make the predicate
+-- null and drop the row silently. Unknown type is not grounds for exclusion.
+and coalesce(s.school_type_code, -1) not in ({{ var('non_england_school_type_codes') | join(', ') }})
+-- Welsh schools GIAS does not type as Welsh establishments (see vars).
+and coalesce(s.local_authority_code, -1) not between {{ var('welsh_la_code_min') }} and {{ var('welsh_la_code_max') }}

pipeline/transform/tests/assert_england_only_schools.sql

@@ -0,0 +1,54 @@
+-- Custom test: the published corpus is England only.
+--
+-- GIAS ships the whole UK plus overseas and offshore establishments. None of
+-- them carry comparable DfE performance data, so dim_school and dim_location
+-- filter them out (see vars.non_england_school_type_codes in dbt_project.yml).
+-- This test is the guard: a GIAS refresh that reintroduces them, or an edit
+-- that drops the filter from one of the two models, fails the pipeline rather
+-- than quietly republishing ~2,000 dataless pages to the site and the sitemap.
+
+select
+    urn,
+    school_name,
+    school_type_code
+from {{ ref('dim_school') }}
+where school_type_code in ({{ var('non_england_school_type_codes') | join(', ') }})
+
+union all
+
+-- dim_location is inner-joined to dim_school by the API, so it must filter
+-- identically. Anything here that dim_school does not have means the two
+-- models have drifted apart.
+select
+    l.urn,
+    null::text    as school_name,
+    null::integer as school_type_code
+from {{ ref('dim_location') }} l
+left join {{ ref('dim_school') }} s on l.urn = s.urn
+where s.urn is null
+
+union all
+
+-- Belt and braces on the Welsh LA code range.
+--
+-- The range 660-681 is documented GIAS structure, not something this project
+-- can verify from its own data — local_authority_code is not exposed anywhere
+-- the app can assert on. So the range does the filtering and the authority
+-- NAME does the checking: if 660-681 is ever wrong or incomplete, a Welsh
+-- authority reappears here and the pipeline fails loudly rather than the
+-- corpus quietly regrowing a Welsh tail.
+--
+-- Beechwood College (142458) is why this exists: a Welsh school GIAS types as
+-- a Special post 16 institution, which the type filter alone let through.
+select
+    l.urn,
+    l.local_authority_name as school_name,
+    null::integer          as school_type_code
+from {{ ref('dim_location') }} l
+where l.local_authority_name in (
+    'Isle of Anglesey', 'Gwynedd', 'Conwy', 'Denbighshire', 'Flintshire',
+    'Wrexham', 'Powys', 'Ceredigion', 'Pembrokeshire', 'Carmarthenshire',
+    'Swansea', 'Neath Port Talbot', 'Bridgend', 'Vale of Glamorgan',
+    'Rhondda Cynon Taf', 'Merthyr Tydfil', 'Caerphilly', 'Blaenau Gwent',
+    'Torfaen', 'Monmouthshire', 'Newport', 'Cardiff'
+)