P5 · Structured data and head tags
Sites: Prajavani (PV, www.prajavani.net, repo tpml) and Deccan Herald (DH, www.deccanherald.com, repo tpml-dh) · Priority: 5 of 6 · Prepared: 18 Nov 2026 by the Product, SEO and Data team · Internal refs: #22, #34, #35, #13, #24, #33
Summary
This package covers what sits in <head> and in the JSON-LD of each page type. Both sites run on Quintype and use the @quintype/seo library, plus their own overrides. The main gaps are: many non-story pages (static, app, search, some tags) show the home page title and description; story pages carry duplicate robots meta tags; DH live blogs have no LiveBlogPosting; DH stories send og:type=website and no og:url; and the publisher in article JSON-LD has no name or logo on both sites. None of this is a large traffic lever by itself. These are hygiene fixes for the next frontend release. One item (the English keyword prefix in PV titles) is an editorial test, not a code change.
Impact
- Not proven: we have no measured traffic gain from any item here. The fixes make markup valid and unique. They do not guarantee ranking or CTR changes.
- Four PV tags show the home title and description:
/tags/cartoons,/tags/horoscope,/tags/op-ed,/tags/weekly-horoscope. They have 1.9M impressions together (Search Console, head-tag audit of 11 Nov 2026). - DH has 5 search URLs with impressions, 1,303 together (same audit).
- DH
/amp/horoscopeis indexed with a self canonical: 16,221 impressions, average position 1.2 (Search Console; the source does not state the period). - PV titles: 2,052 of 38,692 stories (7 Sep – 5 Nov 2026, Quintype
seo.meta-titleexport of 13 Nov 2026) have a stored title ending in "…". PV web CTR at average position 1–3 is 3.4% for new articles (DH 2.3%). Top Stories carousels depress page-level CTR on both sites, so this does not prove a title problem. - DH piracy-name tag pages (
/tags/movierulzand similar) had 17M impressions together in 28 days. That is an editorial brand and ad-safety question, not a code fault. It is noted for awareness only.
Items
1. Pages with no SEO handler show the home title and description — Medium · PV, DH · Status: fix proposed
Issue: Several page types fall back to the home page title and description. They duplicate the home page.
Evidence:
- One URL per page type was fetched on both sites on 11 Nov 2026, and the library
getSeoDatacode was read. - App pages with no SEO handler: PV
authors-page,cartoon-page,horoscope-page,premium,instagram,bookmark,subscribe,search. DHhoroscope-page,premium,news-shots,newsletters,subscribe,search. All use a self canonical and the home title. - Search (
/search?q=…): canonical is/search, title is the home title, robots isindex, follow. - Bookmark, subscribe and profile-type pages: home title, indexable.
- PV tags
/tags/cartoons,/tags/horoscope,/tags/op-ed,/tags/weekly-horoscopeshow the home title and description. - Static pages (
/about-us): empty<title>and empty description on both sites. - The default tag description is broken English: "Read stories listed under on X" (both sites).
- PV
app.jsaround line 610 already passescustomTagsforhome-pageinisomorphicRoutes. DH passescustomTags: {}.
Fix:
- Give each app route a
customTagsentry inisomorphicRouteswith its own title and description. PV: follow the existinghome-pagepattern. DH: replacecustomTags: {}. - Set
<meta name="robots" content="noindex, follow">on/search, bookmark, subscribe and profile-type pages. - Add meta title and description to static pages in the CMS.
- Fix the four PV tag pages: fill their SEO fields in the CMS, or stop the fallback to the home page.
- Rewrite the default tag description so it reads correctly. The source gives no exact wording and does not say where it is generated.
Impact: Removes duplicate titles on app, static and tag pages. Reduces indexable search URLs. Not measured.
Verify: View-source on one URL per type: title and description are unique and not the home page text. /search?q=test has noindex, follow. URL Inspection on /tags/cartoons shows the new title in the rendered HTML.
2. Duplicate robots meta tags on articles and DH brandspot pages — Low · PV, DH · Status: fix proposed
Issue: Article pages output two robots tags.
Evidence:
- Reported on articles:
<meta name="robots" content="max-image-preview:large"/>and<meta name="robots" content="index, follow, max-image-preview:large">. - Confirmed on live DH on 18 Nov 2026 (server-HTML checklist on stories). The source does not say which site the first report came from, or whether PV was re-checked.
- DH brandspot pages output
noindex,nofollow(fromcustomMetaTags) and the defaultindex, follow. Google applies the stricter one, so it works today.
Fix: Output one robots tag per page, with merged directives. On stories: <meta name="robots" content="index, follow, max-image-preview:large">. On brandspot pages, drop the default tag wherever noindex is added.
Impact: Removes ambiguity. No traffic effect expected.
Verify: View-source: exactly one <meta name="robots"> on a story and on a brandspot page.
3. DH live blogs have no LiveBlogPosting — Medium · DH · Status: fix proposed
Issue: Without LiveBlogPosting, live blogs cannot get the "LIVE" label in Top Stories.
Evidence:
- Live check on 16 Nov 2026 on the BRICS live blog (
brics-summit-2026-live-updates-…-4142598): noLiveBlogPosting. - 17 Nov 2026: AMP copies of 4 of 4 sampled DH live blogs also have none (AMP JSON-LD is NewsArticle and BreadcrumbList only).
- Code (
tpml-dh,upstream/main47a7e85e; AMP check at22f4c0f0):app/server/app.js:683-687setsenableLiveBlog: false. The hand-built schema inapp/server/amp/schema-type.jsis used only by non-AMPrender-layout.js:93, and it did not appear live. - That hand-built builder has invalid fields if switched on:
eventStatus: "Live", relative card image URLs, a plain-string author, and no publisher. - PV builds
LiveBlogPostingitself inseo-hard-gate.js:25-133for both page types, and PV AMP live blogs have it.
Fix: Enable the library live-blog schema (enableLiveBlog: true in app.js), or fix schema-type.js. If keeping the custom builder, use absolute image URLs, an author object, a full publisher, and no invalid eventStatus. This item covers the canonical page. The AMP copy is also covered in package P3.
Impact: Makes the live-blog Top Stories label possible. Not measured.
Verify: Rich Results Test on a live blog URL shows LiveBlogPosting. Repeat on its AMP copy.
4. DH article JSON-LD: duplicate blocks, escaped body, empty publisher — Medium · DH · Status: fix proposed
Issue: Each story has two article blocks, articleBody stored as escaped HTML, and a publisher with no name or logo.
Evidence:
- Live on 16 Nov 2026 (a text story
…-4162872, a live blog, a no-hero story…-4139758): libraryArticleplusNewsArticle(enableNewsArticle: true), each witharticleBodyas escaped HTML (<p>Hyderabad: …). PV's hard gate drops the genericArticle; DH's does not. - Live checklist on 18 Nov 2026 (3 stories): NewsArticle
publisheris{"@type":"Organization","id":""}, so no name and no logo. - Breadcrumb exists only in JSON-LD. There are no visible breadcrumb links (same checklist).
- Horoscope articles, 19 Nov 2026 (
…-4223276, after the server-HTML release):articleBodyis<p>https://www.deccanherald.com/</p>. The 12 sign texts are stored instory.metadata, not in text elements. So the defaultarticleBodytakes the story's only text element, which holds the site URL.
Fix:
- Drop the generic
Article, as PV does. - Set
stripHtmlFromArticleBodysoarticleBodyis plain text. - Horoscope: in
seo-hard-gate.js, which already overridesarticleBodyfor premium stories (line 171), add a case forstory-template === "horoscope". It should build a plain-textarticleBodyfrom the 12 sign texts instory.metadata("Aries: …"). - Write the publisher in full, for example:
"publisher": {
"@type": "NewsMediaOrganization",
"@id": "https://www.deccanherald.com/#organization",
"name": "Deccan Herald",
"logo": {"@type": "ImageObject", "url": "<300x300 logo URL>"}
}
DH seo-hard-gate.js:201-232 already builds a NewsMediaOrganization with that @id and a 300x300 logo, so the story publisher should reuse it. The snippet is illustrative; take the real values from that code.
Impact: One valid article entity with a named publisher. Not measured.
Verify: Rich Results Test and view-source: one article block per story, plain-text articleBody, publisher has name and logo. On a daily horoscope, articleBody holds the sign texts, not the site URL.
5. DH stories: og:type=website, no og:url, SVG fallback image, canonical ignores canonical-url — Medium · DH · Status: fix proposed
Issue: Open Graph and canonical tags on DH stories are wrong or incomplete.
Evidence:
- Code:
utils/custom-collection-tags.js:153replaces the libraryTextTagsand outputsog:typewebsitewith noog:url. Still live on 18 Nov 2026 (3 of 3 stories). - On the training site with the Article server HTML fix (package P1),
og:typeisarticle.og:url, publisher and robots are unchanged. - Stories without a hero use
DHcover.svg. Discover and social platforms do not use SVG. Confirmed on…-4139758. Rare: 1 of 12,262 recent stories. custom-collection-tags.js:192: the canonical condition is always true, sostory["canonical-url"]is ignored. It matters only if editors set it on syndicated copies. How often it is set has not been checked.
Fix: Set og:type=article and add og:url (same value as the canonical) on stories. Use a 1200x630 JPG fallback image. Honour story["canonical-url"] when set.
Impact: Correct social and Discover metadata. Not measured.
Verify: View-source on a story: og:type is article and og:url equals the canonical. A no-hero story shows a JPG og:image.
6. PV publisher points to an Organization node that does not exist — Low · PV · Status: fix proposed
Issue: The article publisher is only a reference, so Google has no name or logo for it.
Evidence:
- Confirmed live on 17 Nov 2026 on PV AMP pages: NewsArticle
publisheris{"@id": "https://www.prajavani.net/#organization"}and the page has 0 Organization nodes with that@id. - Code (
tpmlupstream/master70d7bebf):seo-hard-gate.js:264-267sets the reference. No page type defines a node with that@id. The library Organization has no@id(@quintype/seogenerate-common-seo.js:64-68). - The home page defines the organisation three times with different names and
sameAs(views/pages/layout.ejs:163-166,:226-233, plus the library) and has twoWebSiteentities. - The live-blog publisher uses
"id"instead of"@id"(seo-hard-gate.js:66). layout.ejs:194has a corrupted grievance email."@productID"should beproductID(seo-hard-gate.js:278).- Apart from the AMP publisher point, these come from the code audit and were not verified live.
Fix: Either give the layout Organization node "@id": "https://www.prajavani.net/#organization" (and reduce the three definitions to one, and the two WebSite entities to one), or write the publisher in full (name, logo ImageObject) as DH does. Change "id" to "@id" on the live-blog publisher. Fix @productID and the email.
Impact: A named publisher for Google. Not measured.
Verify: Rich Results Test on a PV article and on the home page: one Organization node, and the publisher resolves to a name and logo.
7. PV video and visual-story templates, and logging — Low · PV · Status: fix proposed (code audit, not verified live)
Issue: Smaller hygiene findings in PV server code.
Evidence (tpml 70d7bebf):
- Video and visual-story templates skip the hard-gate fixes (
seo-hard-gate.js:135-150).mainEntityOfPagemay be the home page.VideoObject.embedUrlcan be empty. console.log("Story JSON:", storyJson)runs on every uncached story render (story-page-data.js:98). It costs server time and log volume.- Premium stories (
access: subscription, 658 stories, 1.7%): the server truncates to 115 words for everyone (story-page-data.js:92) and deleteshasPart(seo-hard-gate.js:274,291). Live as Googlebot: 115 of 397 words,isAccessibleForFree: false. This is consistent for users and Google, so it is not cloaking. Premium stories perform well (93% Discover any-impression). Leave it as is; it is tracked separately.
Fix: Apply the hard-gate fixes to the video and visual-story templates (correct mainEntityOfPage, non-empty embedUrl). Remove the console.log.
Impact: Valid video markup and less log volume. Not measured.
Verify: Rich Results Test on a video story and a visual story. Server logs no longer print story JSON.
8. Truncated stored titles on PV, and the English keyword prefix — Medium · PV · Status: needs testing
Issue: 2,052 stored SEO titles end with "…". Separately, 79% of titles carry an English prefix.
Evidence: Quintype seo.meta-title export of 13 Nov 2026, 38,692 stories, 7 Sep – 5 Nov 2026.
- 30,574 (79%) titles are an English keyword prefix plus the Kannada headline. Median length is 71 characters; 90th percentile is 80.
- 2,052 titles end with "…" (1,976 are print copies). The ellipsis is part of the
<title>. - The
<title>is the SEO title with no site name (meta-title || headline, code-confirmed). - Kannada queries are 1.57M of 8.29M query clicks, so most PV searchers type in English. That may be why the prefix exists.
Fix:
- Code and CMS part: find the generator or CMS limit that adds "…", and stop storing truncated titles.
- The English prefix is an editorial test, not a code change. Suggested test: by section, prefix versus no prefix for 4 weeks, compared on CTR at equal position. Editorial decides.
Impact: Not proven. Wide Kannada glyphs may get 70–80 character titles cut in results; this is a hypothesis.
Verify: The count of new stories with a title ending in "…" drops to near zero. For the test, compare Search Console CTR by position band.
9. DH /amp/horoscope indexed with a self canonical — Low · DH · Status: fix proposed
Issue: This AMP listing page is indexed on its own.
Evidence: 16,221 impressions, average position 1.2 (Search Console; the source does not state the period).
Fix: Set the canonical to /horoscope.
Impact: Consolidates signals on the main page. Not measured.
Verify: View-source on /amp/horoscope shows <link rel="canonical" href="https://www.deccanherald.com/horoscope">. URL Inspection reports that as the canonical.
Release checklist
Expected output per page type (from the 11 Nov 2026 head-tag audit):
| Page type | Title / description | Robots | Canonical | Schema |
|---|---|---|---|---|
| Story | Story SEO fields; no site name | index, follow, max-image-preview:large, one tag |
Story URL, or canonical-url if set; AMP points to the HTML story |
One NewsArticle with full publisher; LiveBlogPosting on live blogs (DH) |
| Section | Section SEO | default | Section URL (page 2 points to page 1) | not stated |
PV short-slug collection (/cartoons, /horoscope) |
Collection name / home description | default | Missing today (see P4) | not stated |
| Tag | Tag meta; otherwise name plus a readable default description | default | canonical-slug or self |
not stated |
| Author | Name / bio | default | Self | not stated |
Static (/about-us) |
Set in CMS (empty today) | default | Self | not stated |
| App pages (premium, subscribe, horoscope-page…) | Own customTags |
default | Self | not stated |
| Search, bookmark, subscribe, profile | Own title | noindex, follow |
Not stated | not stated |
| 404 | Page Not Found, HTTP 404 | not stated | None | none |
| PWA shell | Skipped by design |
- [ ] One robots tag on stories and brandspot pages (both sites).
- [ ] Unique title and description on app, static and the four PV tag pages.
- [ ]
noindex, followon search and user-state pages. - [ ] DH:
LiveBlogPostingon live blogs (canonical page here, AMP in P3), one article block, plainarticleBody, full publisher,og:type=article,og:url, JPG fallback image. - [ ] PV: Organization node with
@id(or full publisher); one Organization and oneWebSiteon the home page. - [ ] No new stored titles ending in "…".
- [ ] Rich Results Test and URL Inspection run on one URL of each changed type.
Dependencies and related packages
- P1 Article server HTML: the
og:type=articlechange was seen on the training site with that fix. Check both together. - P2 One URL per story: the PV
/<story-slug>fallback with a wrong canonical, and/zzz/<real-slug>returning 200 with the correct canonical, are handled there. They are not part of this package. - P3 AMP: AMP copies of DH live blogs also need
LiveBlogPosting. Ship one code change and test both copies. - P4 Listing pages and internal links: PV short-slug collections have no canonical.
- P6 Page speed, ads and analytics: removing the PV
console.loglowers server work. It is a server-side hygiene item, not a page-speed fix.