← Issues index

32. Article and listing pages give Google almost no internal links to other stories — High (PV), Medium (DH)

Added 16 Nov 2026. Source audit (PV tpml upstream/master 70d7bebf, DH tpml-dh upstream/main 47a7e85e) + live checks 16 Nov 2026.

What and evidence

Why it matters: deeper articles lack a crawlable path through listing pagination, and automatic recommendation links are absent from the supplied article captures. Sitemaps, footer links, embedded “Also read” cards and other pages can still expose older stories. The defect weakens internal linking; its effect on discovery speed and traffic has not been measured. Low crawl per new URL (#3) and never-shown articles (#31) are context, not proof of causality.

Fix

  1. 🛠️ Server-render a "more from this section" block (5–10 links from the story's section collection, fetched in the page loader) on every article, both sites. Keep the scroll-gated widgets for users.
  2. 🛠️ Real pagination: PV add /:section/:collection/page/:n (or ?page=n) routes that server-render; both sites give react-paginate an hrefBuilder so page links have hrefs. Keep page canonicals self-referencing (see #23).
  3. 🛠️ Raise lazyAfter on the home page to cover the main news rows, or server-render the headline links of lazy rows.
  4. 🛠️ PV breadcrumbs: use the currentPath prop as DH does.

Verify: view-source of an article shows ≥5 same-section story links; a section page links to page 2 with an href that returns 200 with different stories; URL Inspection "View tested page" shows the block.

Search Console crawled HTML supplied on 16 Nov 2026: the Muhurat Trading article (…-4342287) has 17 actual article anchors to 13 distinct stories, all in the footer; zero article anchors outside the footer. Of the footer anchors, 16 (12 distinct stories) are in groups marked hide-mobile; the other is an old Ugadi horoscope link. Taboola containers have no anchors. The article's own paragraphs are populated, so body presence does not ensure recommendation links are present. Method and counts: evidence/gsc_issue32_crawled_2026-11-16_notes.md. The original working HTML was subsequently replaced by the user with the second capture below; its first raw contents are no longer retained. This confirms the contextual-link weakness for the first analysed copy, not an absence of all internal links or a site-wide frequency. The actual Last crawl date and crawler type are still needed from the Search Console panel; 16 Nov is the capture/analysis date. Pagination was not tested by this article capture.

Second capture, 16 Nov 2026 — important qualification: the Prince William article (…-4398310) does contain an embedded “ಇದನ್ನೂ ಓದಿ” (Also read) story card linking to one related story (…-4397730), with two correct hrefs in desktop/mobile markup. A third desktop wrapper href is malformed (/https://www.prajavani.net/…) and nests an anchor; correct links still exist. There are 17 footer article anchors; no automatic latest/most-read block was found, and Taboola containers have no anchors. The article body is populated. Preserved source: evidence/gsc_issue32_prince-william-kate-middleton-mumbai-visit-earthshot-prize-4398310_captured_2026-11-16.html; counts and limits are in the notes file above. Therefore embedded related-story cards can be present in Google's copy: the concern is inconsistent contextual linking and gated automatic widgets, not a universal absence of article-to-article links. No destination fetch or pagination check was run.

Section capture supplied on 16 Nov 2026: Google's saved HTML for PV /district has 10 main-listing story links, but all six pagination anchors lack href, including Page 2 and Next. Their rel="next" attributes supply no destination. No linked page-number route/query was found. Source: evidence/gsc_issue32_crawled_2026-11-16 - non story.html; analysis in evidence/gsc_issue32_crawled_2026-11-16_notes.md. This directly verifies the pagination-link defect for this saved copy, not that older articles are undiscoverable by all other means. Capture date is known; Last crawl and crawler type still need recording. No live page-2 route check was run.

Third article capture, 16 Nov 2026: AI supply-chain story …-4398311 has two valid embedded “Also read” article links, plus 17 footer article anchors; automatic latest/most-read blocks were not found and Taboola containers are empty. Body paragraphs are populated. Source: evidence/gsc_issue32_india-us-ten-nations-ai-supply-chain-agreement-4398311_captured_2026-11-16.html; counts in the notes file above. Embedded related links are demonstrably available on some crawled articles.

Live Chrome DevTools pagination check, 16 Nov 2026: clicking Page 2 on /district loads ten different stories through the district collection API (limit=10&offset=10), but the URL stays /district and pagination anchors still lack href. Reloading returns to page 1 and its original ten stories. Source: evidence/gsc_issue32_pagination_devtools_2026-11-16.json. Thus pagination works for an interactive visitor but supplies no persistent page-2 URL through its controls. Implement distinct, directly loadable page URLs linked by real anchors; changing the address bar alone is insufficient. Hypothetical query/path routes were not manually probed, and no traffic-loss estimate is established.

Direct query-URL check, 16 Nov 2026: opened https://www.prajavani.net/district?page=2 in Chrome DevTools. The query remains in the address bar, but the selected page is 1, the listing contains the same ten stories as the earlier page-1 check, and the canonical is /district. Pagination anchors still lack href. Therefore this query does not currently load the second story batch. Source: evidence/gsc_issue32_direct_page2_2026-11-16.json. This checks the query-string example; other path patterns were not tested.

Correction (16 Nov 2026): Crawl stats JSON samples (crawl_stats/2026-11-16/pv-json/Table.csv) show Googlebot fetching /api/v1/collections/most-read (115 of 257 sampled JSON fetches) and brandspot-collection (85): the calls the scroll-gated widgets make. These samples show Google sometimes requests widget data, while the supplied crawled article copies lack the automatic blocks. The body-phrase results in #2 do not directly measure rendering frequency or widget-link availability. A server-rendered "more from section" block removes this dependency; site-wide link availability remains unquantified.

DH section pages send no story list in server HTML (19 Nov 2026, #20): tpml-dh ab4a2e58 pages/section-page/index.js returns null until a client-side useEffect runs, so the server HTML of /india has 0 story headings (24 after JavaScript) and no section title. Google's first copy of DH section pages therefore carries no listing links. Evidence: evidence/listing_h1_devtools_2026-11-19.json; details in #20. Not yet checked across other DH sections or in Search Console's crawled page.