← Issues index

29. GA4 landing page "(not set)": page_view sent after other events — Medium (PV, DH), analytics

Added 15 Nov 2026. Full write-up: docs/ga4_not_set_landing_pages_2026-11-15.md. Not an SEO ranking issue: it makes GA4 page-level reports undercount real visits, including organic landings.

What. In GA4 (14 Oct–10 Nov 2026, Data API), landing page "(not set)" covers 828k PV sessions (7.3%) and 1.07M DH sessions (9.1%). App sessions are part of that and are expected (PV 439k, DH 214k). The web part (PV 391k, DH 866k) is a tracking gap: these sessions have session_start, first_visit, user_engagement and ad_impression / user_properties_update, but almost no page_view (DH: 1,752 across about 528k sessions). They are real visits on real pages and carry ad impressions (DH 1.08M, PV 200k). About 3% of each site's organic sessions are affected.

Evidence. GA4 Data API pulls (above period); GTM containers GTM-T46PRBZ (DH) and GTM-MGBFQ9R (PV), parsed from the public gtm.js on 15 Nov 2026; headless-Chrome runs on live articles, 15 Nov 2026 (fresh profiles, desktop and 6x-CPU mobile):

Cause.

  1. Both containers fire the Google tag on Initialization with send_page_view: false; GA4 page_view fires only on the site's own dataLayer page_view push. ad_impression is in no GTM tag (sent automatically by the Google tag; DH hits carry an Ad Manager ep.query_id).
  2. The push waits for the React app, then a header useEffect delay, then on story pages a further 1,500ms for the Coral comment count:
    • DH (tpml-dh, feature/nav-v2-home-top @ c459d2bd): app/isomorphic/components/layouts/header/index.js:746-757 (100ms), app/isomorphic/components/helper/datalayer-utils.js:348-358 (setTimeout(..., 1500)).
    • PV (tpml, fix/ssr-story-paragraphs @ bcd5d909): app/isomorphic/components/layouts/header/index.js:169-183 (150ms), app/isomorphic/components/helper/datalayer-utils.js:705-710 (1,500ms); app/isomorphic/components/helper/GlobalSubscriptionSyncer.js:249 pushes user_properties_update for every anonymous visitor as soon as the app runs.
    • Both live bundles (DH app-0d6a25667336aded2caa.js, PV app-42371a7367d100449339.js) contain the same 1,500ms wait.
  3. A visitor who leaves after the first event but before page_view leaves a session with no landing page. Not reproduced directly (no session without page_view captured in the lab); inferred from the setup, code, timing and the matching event mix.

Also found: DH's views/pages/layout.ejs loads gpt.js (line 88) and adsbygoogle.js (line 92) before the GTM partial (line 111); DH's Google tag G-8ZF3X8Y1X2 differs from its web stream G-F45W975H2J (events sent by override; its script failed once in five loads with ERR_BLOCKED_BY_ORB); the vdo.ai player configures two extra GA4 IDs and a dead Universal Analytics ID (UA-113932176-46) through the sites' dataLayer.

🛠️ Fix (proposed, not built). Render the page_view dataLayer push on the server from the story metadata (app/server/handlers/render-layout.js:41 already has the story) in views/pages/layout.ejs, just before the GTM partial, and remove the app's own page_view push. comment_number moves to a later event or is dropped (check with the GA4 report owners). Smaller step on its own: remove the 1,500ms comment-count wait. Housekeeping: remove UA-113932176-46.

Verify. Fresh browser profile: the first GA4 hit is page_view with _ss=1. After release, web "(not set)" share on DH articles falls from 8.9% toward ≤1%; track daily for 2–4 weeks. App sessions, 30-minute session timeouts on open tabs, and bots will still show "(not set)".