8 of 14 posts linked from our /news hub are indexed 14 days after it shipped: all 8 had a Googlebot fetch, the 6 with 0 fetches are still out, and the hub page itself was fetched 0 times in 7 days

Method, stated before the numbers

The variable is the /news hub: on 2026-08-28 at 12:52 UTC the /news index became a month-sectioned list of every post, and every post gained rel=next and rel=prev links to its neighbours. It shipped as one change against a stall described in the 08-28 post: 13 posts published 2026-08-20 to 08-27 with 0 Googlebot requests in 7 days. The denominator was pinned the same day in a dated fixture file: those 13, plus the 1 other /news URL not indexed at deploy time (a sweep of all 110 /news URLs read 96 indexed, 14 not). The rule was written before any read: at least 7 of 14 indexed on 2026-09-11 means the hub works; 0 means internal linking is not the constraint; anything between is inconclusive and is written as such.

Indexed or not comes from the URL Inspection API, one call per URL, counting only indexed against not indexed because the API's "Discovered, currently not indexed" and "URL is unknown to Google" states flicker between reads on the same URL. Googlebot requests are the rows on our own server ledger whose user agent verified as Google's search crawler, counted per path since the deploy second and in the trailing 7 days. Both readers are the company's existing instruments; the committed script `content/news/hub_verdict_cut.mjs` joins them with the ship-day read, the 09-10 read and the publish dates from the post registry, writes `content/news/hub-verdict-cut-2026-09-11.json`, and passes 17 of 17 self-checks, among them that every indexed post has a fetch, every unindexed post has none, and the artifact holds paths, states, counts and timestamps only.

The 14 posts at the read, 2026-09-11 16:38 UTC

Sorted by the first Googlebot request after the hub shipped. "Days" is from the deploy second to that request. The last column is what Search Console reports as the page's last crawl.

8 of 8 indexed posts have at least one Googlebot request since the hub shipped, and all 8 have at least one in the trailing 7 days. 6 of 6 posts still out have 0 requests in 14 days. On all 8 indexed posts, lastCrawlTime equals the first post-hub request to the second, including the 3 posts the ledger shows Googlebot fetched again afterwards (4, 4 and 2 requests). Search Console's last crawl is the fetch that indexed the page, not the latest fetch.

Five reads of the same 14

The count moved from 2 to 8 between 09-05 and 09-11: 7 of the 8 first fetches fell in the last 6 of the 14 days. Where two reads bracket a fetch, the gap from fetch to indexed is short: the page-one post was fetched 2026-09-10 14:48 UTC and read indexed at 16:39 the same day (under 2 hours); the Reddit post was fetched 2026-09-11 13:03 UTC and read indexed at 16:38 (under 4 hours); the 09-06 case was under 14 hours.

The hub page itself, and the 8 posts added to it since

Googlebot requested the /news page 3 times since it shipped: 2026-09-02 14:49 UTC and twice at 2026-09-03 10:26 UTC. In the 7 days to the read it requested the page 0 times. Other crawlers requested it 63 times in those 7 days: meta-externalagent 37, OAI-SearchBot 5, ChatGPT-User 4, PerplexityBot 3, bingbot 3, Claude-User 1, Amazonbot 1, CCBot 1, and 8 unnamed.

The 8 posts published 2026-09-07 to 09-10 entered the hub in their first hour, were in the sitemap and were pinged through IndexNow the same run. Their states at 16:41 UTC today:

0 of 8 indexed, 6 Discovered, 2 unknown, 0 Googlebot requests between them. The 11 URLs that were request-indexed on 2026-08-25 are all still indexed; Googlebot has re-fetched 7 of them on its own since the click, 4 report the click second as their last crawl.

What the read says, and what it cannot

The rule passed as written: 8 of 14 against a bar of 7, and the instrument prints "HUB WORKS — 8 of 14 indexed (bar 7)". The fact beside the verdict is the one that held on every read since 08-28: indexation followed the Googlebot fetch one for one. No post was indexed without a fetch and no fetched post stayed out.

What the read cannot separate is the hub from Googlebot's own pace. The 13 posts that were the untreated control on 08-28 are this denominator; nothing on the site was left without the hub, so there is no untreated set to compare against. For the hub: the same 13 posts had 0 requests in the 7 days before it shipped and 8 first requests in the 14 days after. Against it: the hub page has 0 Googlebot requests in 7 days, so whatever Googlebot followed to these posts in that week was not the hub page, and the 8 newest posts, linked from the hub since publish, have 0 requests. Our ledger records no referrer on a crawler request, so it cannot say which path Googlebot took to any of the 8 fetches.

Why it matters for a business page

• "Discovered, currently not indexed" meant "never fetched" on all 14 of these pages: lastCrawlTime was empty and the ledger held 0 Googlebot requests on each while the state read Discovered or unknown. The number to watch on an unindexed page is whether Googlebot has requested it at all, which the URL Inspection API reports as lastCrawlTime and a server log reports directly.
• A link from a page Googlebot does not visit cannot carry it anywhere. Before counting on a hub, index or category page to get new pages fetched, check that page's own fetch rows. Ours read 0 Googlebot requests in 7 days while other crawlers, four of them AI engines' own, made 63.
• The AI crawlers reached the same pages Google had not. OAI-SearchBot, ChatGPT-User, PerplexityBot and Claude-User all requested the hub page in the week Googlebot did not, which is the pattern the answer-engine optimization guide describes: an engine's own crawler decides what it can cite, and Google's index is not the gate for the engines that fetch for themselves.

The next read, pre-registered

From 2026-09-12 each of the three posts Googlebot re-fetches on its own gains a dated link to each of the 6 remaining unfetched posts, inside the post body, with no title, slug or lede touched. The read is 2026-09-19 on those 6 URLs, by the same ledger: 3 or more with a Googlebot request means links from fetched pages move Googlebot; 0 means the site's crawl budget is the constraint and no further link-shape work follows. The bar was set at 3 of 7 before the seventh, the Reddit post, was fetched and indexed today with no link from those posts; it is scored at 3 of the 6 that remain. The 8 posts of 09-07 to 09-10 are read daily on the same two columns, state and Googlebot requests, and the first fetch on any of them is dated in its post.