P4 · Listing pages and internal links
Sites: Prajavani (PV, www.prajavani.net, repo tpml) and Deccan Herald (DH, www.deccanherald.com, repo tpml-dh) · Priority: 4 of 6 · Prepared: 18 Nov 2026 by the Product, SEO and Data team · Internal refs: #32, #19, #20, #23, #27
Summary
Listing, section, tag, author, home and horoscope hub pages give Google weak signals on both sites. Pagination links have no href, so deeper stories have no crawlable path through listings. PV has no page-2 URL at all. On PV, the same stories are listed at several URLs (a short URL, a section URL and tag URLs), and many short URLs have no canonical tag. Listing pages carry about 11 <h1> elements on PV and 10 or 11 on DH tag and author pages, while many DH section pages have none. The horoscope hubs, which have the highest click-through rate of any topic, suffer from all of this at once. This package fixes the listing templates and the PV duplicate-URL problem. The redirect rule for PV short URLs must ship together with the redirect cleanup in P2.
Impact
All figures are Search Console unless stated. Nothing here has a measured traffic gain; these are defects and exposure, not forecasts.
- Duplicate listing URLs (PV). Search Console, 12 Oct – 8 Nov 2026, URLs with 300+ impressions (260 PV non-article URLs). Examples:
/cartoons4.69M impressions (no canonical) beside/news/cartoons3.13M,/tags/cartoons1.88M and/tags/chinakurali1.41M. The search "chinakurali" (821k impressions) is split across 4 URLs. Short URLs usually get more impressions at a worse position (for example/tumakuru15.4k impressions at position 8.9 vs/district/tumakuruat position 5.7). - Traffic on URLs a redirect rule would move (same period, 28 days): kind B 42 URLs, 8,029 clicks; kind A except cartoons and churumuri 10 URLs, about 1,700 clicks;
/cartoons,/cartoons?page=and/churumuri-op-ed3 URLs, 2,264,513 clicks; hand-made collections that stay (kind C) 13 URLs, 47,523 clicks. - Horoscope. Search Console, 12 Oct – 8 Nov 2026: PV
/horoscopehad 267,134 impressions and 26,000 clicks at position 5.7. "dina bhavishya" had 42,822 impressions at position 6.1. DH "horoscope today" is at position 11 (30,410 impressions, 4,401 clicks). Click-through rate on horoscope is 8–18% on both sites. Moving generic searches into the top 3 could be worth several times the current clicks. That is an opportunity estimate, not a measured result. - Headings. Live check, 11 Nov 2026: of 228 DH listing URLs, 59 have no H1 and 27 have 11 H1s.
- Not proven. No traffic loss has been tied to missing pagination hrefs or the lazy home rows. Stories are also found through sitemaps, footer links and embedded "Also read" cards. The effect of each change on discovery speed and rankings has not been measured. Page-1 canonicals on paginated pages (item 2) are a low-impact item.
Items
1. Pagination links with real href and directly loadable page URLs — High (PV), Medium (DH) · PV, DH · Fix proposed
Issue: Both sites use react-paginate without hrefBuilder, so page links render as <a role="button"> with no href. PV has no page-N route. ?page=2 is accepted in the address bar but still shows page 1.
Evidence:
- Code (PV
tpmlupstream/master70d7bebf, DHtpml-dhupstream/main47a7e85e, audit 16 Nov 2026):app/isomorphic/components/molecules/pagination/index.js:24-34has nohrefBuilder. PV server HTML shows about 10 stories per section or tag page (collection-templates/listing-page-template/index.js:27,33). PV author pages load 7 stories (data-loaders/author-page-data.js:5). DH has/<section>/<n>routes (app/server/routes.js:118,200,205) but never links to them. - Search Console crawled HTML for PV
/district, captured 16 Nov 2026: 10 main-listing story links; all six pagination anchors lackhref, including Page 2 and Next. Theirrel="next"attributes give no destination. - Live PV
/news/karnataka-news, 16 Nov 2026: 23 story links, 7role="button"pagers, no?page=links. - Chrome DevTools, 16 Nov 2026: clicking Page 2 on PV
/districtloads ten different stories from the collection API (limit=10&offset=10), but the URL stays/district. A reload returns page 1. - Direct check, 16 Nov 2026:
https://www.prajavani.net/district?page=2keeps the query in the address bar, but shows page 1 with the same ten stories. The canonical is/district. Other URL patterns were not tested.
Fix:
- Both sites: pass an
hrefBuildertoreact-paginateso every page link is a real<a href>to a URL that returns 200. - PV: add routes that server-render page N, for example
/:section/:collection/page/:nor?page=nhandled by the section, collection, tag and author loaders. The loader must read the page number and fetch the matching offset. - DH: keep the existing
/<section>/<n>routes and link to them from the pager. - Each page URL gets a self-referencing canonical and its own description (see item 2).
- Changing the address bar with JavaScript alone is not enough. The page must load correctly when requested directly.
Impact: gives Google a crawlable path to older stories from listings. Effect on discovery and traffic is unmeasured.
Verify: view-source of a section page shows a page-2 link with an href; that URL returns 200 with different stories than page 1. URL Inspection "View tested page" on the page-2 URL shows its own stories and its own canonical.
2. Self-referencing canonicals and descriptions on paginated pages — Low · PV, DH · Fix proposed
Issue: Paginated listings point their canonical at page 1. Google's guidance is a self-referencing canonical per page.
Evidence:
- PV
/district/hasana?page=2and DH/india/2both carry the page-1 canonical. DH page 2 also has an empty description (live check, 11 Nov 2026). - PV
/cartoons?page=9has no canonical at all and had 51,175 impressions (28 days, 12 Oct – 8 Nov 2026). - In practice stories are also found through sitemaps, so impact is small.
Fix: after item 1 gives page URLs real links, set each page's canonical to its own URL and give page 2 and later a non-empty description. Change this only together with item 4 (the PV canonical work), so the two do not conflict.
Impact: hygiene. No numbers.
Verify: view-source on page 2 shows a canonical equal to its own URL.
3. Server-render more of the home and section pages; fix empty PV breadcrumbs — Medium · PV, DH · Fix proposed
Issue: Home and section pages render only the first two collections on the server. PV section and sub-section breadcrumbs are empty in server HTML.
Evidence:
- Code, same commits as item 1:
LazyCollection lazyAfter={2}at PVpages/home.js:88-94andsection-page/index.js:89-104, and DHhome.js:84-90. - PV breadcrumbs take
initialPathfromwindow.location(section-page/index.js:38,sub-section-page/index.js:74), so the server has no path. DH uses thecurrentPathprop.
Fix:
- Raise
lazyAfteron the home page to cover the main news rows, or server-render the headline links of lazy rows. - PV: use the
currentPathprop for breadcrumbs, as DH does.
Impact: more story links in the HTML Google sees on the most-linked pages. Unmeasured.
Verify: view-source of the home page lists headline links for the main rows; PV section view-source contains the breadcrumb trail.
4. Duplicate short listing URLs: one canonical URL per topic — High (PV), Low (DH) · PV, DH · Fix proposed
Issue: The same stories are listed at /<slug>, /<section>/<slug> and tag URLs. Short URLs often have no canonical and carry the site-wide description ("Kannada News - Find Trending Latest News…"), so Google treats each as its own canonical and splits searches. DH has the same short URLs, but with canonicals that point to the wrong place (see "Full URL list" below); low impact.
Evidence:
- Cause (PV
tpmlupstream-syncfc684344, 16 Oct 2026;@quintype/seo1.50.0):app/server/routes.js:135sends any/:collectionSlugthat is not a section to the section page type.section-page-data.jsloads the collection with no section. In@quintype/seoa section page with no section falls back to the collection'scanonical-urlmetadata, else no canonical tag.redirectCollectionHandler(app/server/app.js:91) 301s section-type collections to their section URL, but only on/collection/:slug. - Three kinds of short URL (checked against PV's public API, 11 Nov 2026):
- A. A section's own backing collection, for example
/hasana-district,/cartoons,/churumuri-op-ed. 228 of 273 sections have such a collection with a different name, so most sections have three URLs. - B. A bare section slug, for example
/hasana,/tumakuru,/india-news. The loader finds the section but recurses with an emptysectionObjinstead of the section it found. - C. Collections created by hand in the CMS, for example
/horoscope,/daily-horoscope,/news-landing,/abhimata. These are landing pages with no section and no canonical or summary.
- A. A section's own backing collection, for example
/horoscopeis both a hand-made collection (33,688 clicks in 28 days) and the slug of section/astro-vastu/horoscope(491 clicks), so rule order matters.
Full URL list (short-URL inventory, 19 Nov 2026; available as two CSV files, one per site). Every one-segment URL the /:slug route can serve, from the Quintype config (all sections and their backing collections) plus every one-segment URL in Search Console (web + Discover, 20 Jul 2025 – 16 Nov 2026) and the menus. Each row has the kind, the action, the section URL to redirect to, and clicks and impressions.
| Site | Kind | Action | URLs | Known to Google | Clicks (16 months) |
|---|---|---|---|---|---|
| PV | A (section's collection) | 301 to section URL | 238 | 93 | 48,437,192 (48,352,845 are /cartoons + /churumuri-op-ed) |
| PV | B (bare section slug) | 301 to section URL | 212 | 154 | 214,421 |
| PV | C (hand-made collection) | keep; own canonical and description | 50 | 50 | 943,536 |
| DH | A | canonical to section URL (or 301) | 140 | 41 | 31,760 |
| DH | B | canonical to section URL (or 301) | 100 | 8 | 6,280 |
| DH | C | keep (already self-canonical) | 102 | 102 | 1,562,840 |
- PV live check (6 random A/B URLs, Googlebot): all 200 with no canonical.
- DH live check: canonicals exist, but kind A pages point to themselves (
/design-lifestyle) and kind B pages to the kind A URL (/sikkim→/sikkim-india), not to the section URL. For DH, change its generator so kind A and B pages canonicalise (or 301) to the section URL. - PV
/horoscopeis kind C (its collection uses thedefaulttemplate), not a section collection: keep it (item 8). - Review targets before building the map: a few slugs match a section in an unexpected place (DH
/karnataka→/elections/karnataka). - Hand-made collections that no menu links to and Google has never shown are not in the list; ask Quintype for a CMS export of all collections to complete kind C.
Fix:
- CMS, no deploy: fill in canonical URL and summary on each duplicate collection (for example
horoscope→https://www.prajavani.net/horoscope,hasana→/district/hasana). - Code default: when a collection has no canonical, use the page's own URL (
customSeo.canonicalUrlin the section loader, or a TPML SEO generator). Preferred route: port DH's generator (item 5). - Redirect rule on
/:slug, in this order: (1) a collection with that slug exists and its template issection: 301 to its section URL (kind A); (2) a collection exists with any other template: leave it (kind C); (3) no collection, but a section has that slug: 301 to the section URL (kind B). Reuse theredirectCollectionHandlerlogic. Keep the query string (/cartoons?page=9→/news/cartoons?page=9), compare slugs in lower case, and put the rule behind an on/off switch. - Tag pages that duplicate a hub (
/tags/cartoons,/tags/horoscope,/tags/chinakurali): canonical to the hub ornoindex, follow. Decide per tag.
Staged rollout (agreed 11 Nov 2026, recommended over a single release):
| Step | What | Undo |
|---|---|---|
| 0. Baseline | Record clicks, impressions and position for every URL that will change and its section URL, plus the searches "prajavani cartoon", "chinakurali", "churumuri". | — |
| 1. Build | The rule above, with /cartoons and /churumuri-op-ed on an exclusion list. Ship the redirect cleanup (P2) in the same release so no hand-written redirect chains through the rule. |
Switch off |
| 2. Test | Every kind A and B URL reaches its section URL in one 301. Kind C URLs and top-level sections are untouched. No loops. | — |
| 3. Wave 1 | All kind A and B URLs except /cartoons and /churumuri-op-ed: 448 URLs, 245 known to Google (full list above). Straight 301. |
Switch off |
| 4. Watch | 3–4 weeks. Compare section-URL clicks with old-URL clicks combined. A dip of 1–3 weeks is normal. | — |
| 5. Wave 2, canonical trial | Set the canonical of /cartoons to /news/cartoons and of /churumuri-op-ed to /op-ed/churumuri (CMS or code), not a redirect. Watch 2–3 weeks until Google shows the section URLs for the brand searches. |
Remove the canonical |
| 6. Wave 2, 301 | Remove the exclusion so the rule 301s them. | Restore the exclusion |
Wave 1 moves about 299K clicks over 16 months (about 10K in a typical 28 days); wave 2 moves 48.4M over 16 months (2.26M in 28 days). The canonical trial exists because Google currently prefers /cartoons (2.19M clicks vs 459k for /news/cartoons). A lower-risk alternative for wave 2: leave both pages as they are and give each a canonical pointing to itself. That ends the "no canonical" state but keeps the duplicate.
Kind C pages (/horoscope, /news-landing, /daily-horoscope, /district-landing, /trending, /abhimata; 47,523 clicks) are a separate editorial decision: choose one URL per topic, fill in canonical and summary, and point old redirect entries at the chosen URL. Horoscope is covered in item 8.
Risks: a 1–4 week re-processing dip; the consolidated URL ranking lower than the old one (main risk, for /cartoons); a different title or description on the section URL changing click-through; implementation bugs (wrong pages redirected, dropped query strings, loops), which are the most likely way to lose traffic; reports by URL showing the old URL dropping.
Impact: ends split rankings for the topics in the Impact section. Whether the section URLs rank as well as the short ones is not proven; hence the staging.
Verify: view-source shows one canonical on every short URL; URL Inspection shows Google's canonical as the chosen hub; split searches in Search Console converge on one URL.
5. DH canonical generator as the model for PV — Reference · DH (no defect), PV (port)
Issue: DH has the same /:collectionSlug catch-all and an older @quintype/seo (1.42.2) that always skips the collection canonical. DH avoids the problem with its own SEO generator.
Evidence:
- DH audit of 228 non-article URLs with 300+ impressions: only 2 had no canonical (
/news_sitemap.xml,/grievance-redressal). - The generator is
app/isomorphic/components/utils/custom-collection-tags.js(wired inapp/server/app.js,customizedTags). It always outputs a canonical: the section URL, elsesketches-hostplus the collection slug. Added 6–11 Sep 2023 (PRs #468 and #475). Reachedmainin PR #1416, 15 Sep 2025. - DH
main0bc245a1, 23 Oct 2026. GitHub access returned 403/404 on 11 Nov 2026, so the last fetched refs were used.
Fix: port this generator and its section-page branch in customizedTags to PV instead of patching each collection. Tidy its condition (sectionURL !== null || sectionURL !== ""), which is always true.
Impact: covered by item 4.
Verify: as item 4.
6. One H1 per listing page — Medium (PV and DH) · PV, DH · Fix proposed
Issue: PV listing pages have about 11 <h1> elements (one per story card plus the collection name). DH section pages often have none, while its tag and author pages have 11.
Evidence:
- PV code:
molecules/listing-story-card/index.js:96defaultsheaderLevelto 1.listing-page-template/index.jsrenders 10 cards without overriding it.molecules/collection-nameadds another H1.story-card1andmobile-story-card1default to 1;slider-with-multiple-big-cardsandvideo-sliderpassheaderLevel={1}. - DH code:
listing-story-cardline 192,story-card1,story-card7andstory-card9default to 1;story-card18/21/23,slider-with-multiple-big-cards,video-slider, and the budget and election widgets passheaderLevel={1}. - Validated 19 Nov 2026 (code at PV
0b3f7227, DHab4a2e58; Chrome DevTools): PV listings and DH tag pages have 11<h1>in both server HTML and the rendered page. DH section pages (/india): one H1 after JavaScript but none in server HTML, and no story list either (24 story headings appear only after JavaScript), becausepages/section-page/index.jsreturns nothing until a client-sideuseEffectsetsisInitialized. Fix for DH section pages: render the title and the first page of stories on the server, then apply the card heading change. - DH live result (11 Nov 2026): of 228 listing URLs, 59 have no H1 (including
/india,/india/karnataka,/world,/latest-news,/top-bengaluru-news,/horoscope) and 27 have 11 H1s (for example/tags/bengaluru,/author/dh-web-desk).
Fix: card headlines default to H2 or H3 on both sites. One H1 per listing page: the section, tag, author or collection name. Remove the headerLevel={1} overrides in the listed widgets. Keep CSS classes so the look does not change.
Impact: clearer page topic for Google. Ranking effect unmeasured.
Verify: view-source of a section, tag and author page on each site shows exactly one <h1>.
7. PV hub head tags — Medium · PV · Fix proposed
Issue: Listing hubs have weak titles and the site-wide description. Live check, 12 Nov 2026: /horoscope title is just "ದಿನ ಭವಿಷ್ಯ"; /weekly-horoscope is "Weekly Horoscope" and /monthly-horoscope is "Monthly Horoscope" (English on a Kannada site); all three use the site-wide description.
Fix: own Kannada titles and descriptions per hub, for example "ಇಂದಿನ ದಿನ ಭವಿಷ್ಯ (12 ನವೆಂಬರ್ 2026): 12 ರಾಶಿಗಳ ರಾಶಿ ಭವಿಷ್ಯ | ಪ್ರಜಾವಾಣಿ", "ವಾರ ಭವಿಷ್ಯ", "ಮಾಸ ಭವಿಷ್ಯ". The CMS summary fixes the description. A dated title needs code (customTags or the item 5 generator). Wider title and meta-tag rules are in P5.
Impact: snippet quality on the highest-click-through topic. Unmeasured.
Verify: view-source of each hub shows the new title and its own description.
8. Horoscope hubs and pages: listing-specific fixes — High · PV, DH · Fix proposed
Overlaps are not repeated here. One topic with several URLs is item 4 (kind C). Heading counts on listings are item 6. DH server-rendered horoscope text belongs to P1. The DH articleBody JSON-LD value and other head tags belong to P5. The DH /amp/horoscope canonical belongs to P5. All evidence below is from live checks on 12 Nov 2026 (raw HTML with a Googlebot user agent, plus headless Chrome).
PV fixes:
- Hub choice.
/horoscopeand/daily-horoscopehave no canonical,/astro-vastu/horoscopeis self-canonical, and all three list the same 10 stories; "prajavani dina bhavishya" ranks with 4 URLs. Same pattern for weekly (/weekly-horoscope,/tags/horoscope-weekly,/tags/weekly-horoscope). Choose/horoscopeas the hub. Stopgap with no deploy: set canonical and summary on thehoroscope,daily-horoscopeandweekly-horoscopecollections in the CMS. Then 301/daily-horoscope→/horoscope, point/astro-vastu/horoscopecanonical (or 301) at/horoscope, and give/tags/horoscope*a canonical to the hub ornoindex, follow. Point the oldredirect-urls.jsentries (/dailyhoroscope.html→/daily-horoscope) straight at/horoscope(P2). - Article headline and signs. Even after JavaScript the headline is a
<div>(story-templates/horoscope/index.js:394) and the 12 signs are plain text cards. Put the headline in<h1>(keep the class). Make each sign an<h2 id="...">with idsmesha,vrushabha,mithuna,kataka,simha,kanya,tula,vrischika,dhanu,makara,kumbha,meena, and add a jump list of the 12 signs at the top. - Related links. Every horoscope article links to
/horoscope/ugadi-2025-new-year-horoscope-3226114and none links to/horoscope. Link to the hub and drop the Ugadi 2025 link. - Search demand (Search Console, 20 Oct – 16 Nov 2026).
/horoscopehas 33,043 of 51,952 PV horoscope web clicks. Generic searches sit at positions 5–8: "ಭವಿಷ್ಯ" 75,388 impressions (5.9), "dina bhavishya" 43,306 (5.6), "today rashi bhavishya in kannada" 15,255 (6.1), "rashi bhavishya kannada" 13,861 (7.5). Brand searches ("prajavani dina bhavishya", 57,508 impressions) show/horoscope,/weekly-horoscopeand/daily-horoscopetogether at position 1. - Editorial, not code. The weekly listing skips 1–7 Nov and the monthly listing jumps Nov → Sep → Jun → Feb 2026. Daily articles are mis-filed too:
/astro-vastu/vaastu/articles earned 2,495 horoscope web clicks in the same 28 days (daily 1,528, weekly 833). Re-file under/astro-vastu/horoscope/. SEO titles are inconsistent (12 Nov dated with an English prefix; 11 Nov undated with a doubled vowel sign in<title>); use one dated pattern and put the date in the slug.
DH fixes:
- Two hubs.
/astrologyand/horoscopelist the same stories, each self-canonical. "horoscope today" is split between/horoscope(position 17.3) and/astrology(position 60.6). Choose/horoscope(it has the impressions); canonical or 301/astrologyto it. - Empty hubs without JavaScript (top DH horoscope item). Raw HTML of both hubs has about 1,800 characters of text, no links to horoscope articles, no sign names and no H1. Search Console, 20 Oct – 16 Nov 2026: for "horoscope today" (28,772 impressions)
/horoscopeis at position 18.3 while the dated daily articles rank 2.0–4.5; the same holds for "today's horoscope", "horoscope" and "daily horoscope". Render the story list as real<a href>links and one H1 in the server HTML; ideally today's 12 signs. - Article headings. After JavaScript the headline is not an H1 and each sign is plain text. Still the case on 19 Nov 2026: since the 19 Nov release the 12 sign texts are in the server HTML (P1), but that HTML has 0
<h1>and 0<h2>. Instory-templates/horoscope/index.js, render the headline (<div styleName="title">, line 604) as<h1>and each sign name (<div styleName="horoscope-name">, line 380) as<h2 id="aries">etc. Keep the classes so the styling is unchanged. - Sign date ranges. Sagittarius ends "December 21" and Capricorn starts "December 23" (lines 217 and 223), so 22 December has no sign. Virgo and Libra both include 23 September (lines 199 and 205). Fix the ranges, e.g. Capricorn "December 22 - January 20".
- SEO titles. The 12 Nov
<title>is "Daily Horoscope: Today's astrological predictions and sun sign insights" (no date, near-identical to 10 Nov); the visible headline is dated. Use one dated pattern, or leave the SEO title empty so the dated headline is used (editorial). - Sign tag pages (
/tags/aries-horoscopeand 11 more): 200, self-canonical, description "Read stories listed under on Aries Horoscope", 11 H1s, and they list all-signs articles. Aries tagging skips 9, 10 and 12 Nov. Short term: tag meta titles and descriptions in the CMS, and tag every daily article with all 12 signs. Longer term: real sign pages (a content plan item, not a technical fix). - Sign images: no
alt, URL ends?w=undefined, served from adh-sandbox/path. Still the case for all 12 images on 19 Nov 2026.ZodiacIcon(horoscope/index.js:38-42) passes onlywidths={[100]}toResponsiveImage, which fills thesrcsetbut not thesrcwidth. Pass a default width andaltwith the sign name (e.g. "Aries horoscope"), and move the images to the production path. /horoscopedescription says "predictions for 2025". Remove the year or update it each January.
Impact: horoscope is the highest click-through topic (figures in Impact). Do not add this benefit to items 4 or 6; it overlaps them.
Verify: URL Inspection shows /horoscope as Google's canonical for PV's three listing URLs and for DH's /astrology. View-source of one daily article per site shows one H1 and 12 sign H2s. Track "dina bhavishya", "rashi bhavishya kannada" and DH "horoscope today" weekly; "prajavani dina bhavishya" should rank with one URL.
Release checklist
- [ ] Page-N URLs exist on PV and are linked from the pager with real
hrefs on both sites (items 1, 2). - [ ] Home and section pages server-render the main rows; PV breadcrumbs appear in server HTML (item 3).
- [ ] PV short-URL rule behind a switch, tested for one-hop 301s, query string kept, kind C untouched (item 4).
- [ ] Baseline of clicks, impressions and position recorded before wave 1;
/cartoonsand/churumuri-op-edon the exclusion list (item 4). - [ ] Redirect cleanup from P2 ships in the same release as the rule.
- [ ] One H1 per listing page on PV and DH (item 6).
- [ ] PV hub titles and descriptions updated (item 7).
- [ ] Horoscope hub chosen per site, canonicals set, headings added (item 8).
- [ ] Watch period of 3–4 weeks after wave 1 before starting the cartoons canonical trial.
Dependencies and related packages
- P1 Article server HTML: the article "more from this section" block (5–10 same-section links server-rendered on every article) lives there. DH horoscope server-rendered text is also there.
- P2 One URL per story: the redirect cleanup (
redirect-urls.js) must ship with item 4 and the item 8 redirects. - P5 Structured data and head tags also carries DH
/amp/horoscope(indexed with a self canonical; fix: canonical to/horoscope). - P5 Structured data and head tags: the DH
articleBodyJSON-LD junk value, the broken default description on tag pages, and general title rules. - P6 Page speed, ads and analytics: none directly. Raising
lazyAfter(item 3) adds server-rendered rows, so check page weight there.