TrafficTriageDo Hub Pages Help SEO? Six Hubs, One Click in 29 Days
By Ugur Saritepe · October 11, 2026
Do hub pages help SEO? The links a hub creates can help. The hub page itself earned almost nothing. I built six of them on game-scout.app, logged a prediction with a date on it, and when that date arrived the six hubs had returned 1 click between them in 29 days.
Update, 11 October 2026.Eight weeks in, I ran the index check this post tells you to run and I had not run myself. Three of the six hubs have never been crawled. The readout is in the three sections before "How to check your own hub pages", and one sentence further down is corrected in place.
What I built, and what I predicted
game-scout.app is the games directory I run and write about openly. In August 2026 it had one page per topic and no structure above them. The model I copied was a competitor running eight URLs plus two hub pages for a single topic, against our one. The note I wrote at the time was that our best page is not underperforming, it is alone.
So between 18 August and 9 September 2026 I shipped six hub pages, each one a directory over a cluster of existing lists, each launched with in-context links from member pages that already ranked. This is what I typed when the first one went out:
I am trying to create hubs that are not touched or crawled, so my plan is increase reach with hubs
I would like to prepare perfect cluster before waiting for results
Two predictions went into the log the same day, both with a check date of 15 September 2026. The first: the new vehicle-sims hub gets indexed and earns at least 5 clicks within four weeks. The second was the interesting one, because it tested the structure claim with no writing attached. One member list, best-flight-games, had 2,203 impressions and zero inbound internal links. It got siblings and a breadcrumb, no content change, no title change. Baseline 22 clicks per 90 days, expected 40.
What the hubs actually earned
The check date arrived. Here is every hub page, over the 29 days from 18 August to 15 September 2026:
site: sc-domain:game-scout.app dimensions: ["page"] 2026-08-18 to 2026-09-15
/best/sci-fi-games 1 click 24 impr pos 8.1 /best/vehicle-sim-games 0 clicks 5 impr pos 5.0 /browse/action-combat-games 0 clicks 2 impr pos 9.5 /browse/crime-investigation-games no row returned /browse/horror-survival no row returned /browse/job-life-sim-games no row returned
Six hub pages, 1 click and 31 impressions between them. Three of the six did not appear in the export at all, which means not visible rather than zero.
Over a comparable window, three member lists inside just one of those clusters returned 326 clicks: best-trucking-games 174, best-flight-games 115, best-bus-simulator-games 37. The hub sitting above them returned none.
A hub can rank at position 5 and still get nothing
The vehicle-sims hub averaged position 5 across 5 impressions. It was not buried and it was not unindexed. Google served it near the top of page one, five times in a month.
That is the whole lesson in one row. A hub page is named for the aggregate: vehicle sim games, crime investigation games, horror survival. Almost nobody types the aggregate. They type the specific thing, which is what the member lists are named for, and those lists collect the demand. Ranking well for a phrase with no searches behind it produces exactly what you would expect.
This is worth checking before you build, not after. Our own MCP serverreads a site's topical map, which is the same question asked from the other end: which topic families does the site cover, which are thin, and where is the demand it is not appearing for.
The page that did move, and why I cannot claim it
Prediction two looked like a win. best-flight-games went from a 22 clicks per 90 days baseline to 158 clicks per 90 days, measured 15 September 2026, against an expected 40. Nearly four times the target, with no editorial work on the page at all.
Then I pulled the daily series, and the claim fell apart:
best-flight-games, impressions per day 2026-07-20 38 2026-08-15 257 2026-08-07 87 2026-08-16 283 2026-08-08 171 <- climb starts 2026-08-18 230 <- links ship 2026-08-13 157 2026-09-05 715
The climb starts around 8 August, ten days before the internal links shipped on 18 August. Daily clicks were already running at 4 to 6 before the change.
The whole site was climbing at the same time: 1,353 clicks in the 28 days to 18 August against 2,730 in the 28 days to 15 September, with impressions going 68,411 to 132,237. A page inside a site that doubled cannot be used as evidence that one link change doubled it.
My control group was not a control group
This is the part I got wrong at the design stage, and it is the reason prediction two can never be settled. I picked three clusters to leave alone as controls: cozy, wrestling and city building. Over 90 days to 15 September 2026:
- best-cozy-games: 0 clicks, 17 impressions.
- wrestling: Search Console returned no data for the entire 90 day window.
- best-city-building-games: 11 clicks in 28 days, from a prior period with no row at all.
A control group exists to answer one question: would this page have moved anyway? Pages with no traffic cannot answer it. They have no baseline to hold steady and no signal to compare against, so whatever they do is noise. I chose them because they were untouched, when the requirement was that they be untouched and already measurable.
What the ranked guides say, and what our rows say
I checked the search results for "do hub pages help seo" on 15 September 2026. Nine of the ten ranked pages are definitions and build guides, two of them are about a hosting platform that happens to be called HubPages, and not one of them reports what a hub page actually earned. The IONOS guide lists seven SEO benefits with no measurement behind any of them. A Loganix guide states that your hub page will get indexed faster and rated higher.
Our rows say that second claim is half right in the way that matters least. The hubs did get indexed and did rank, between position 5 and 9.5. They returned 1 click. Indexed and ranked are not the same as read.
Corrected 11 October 2026:"the hubs did get indexed" was true of the three that had a row. I had not inspected the three that had none, and when I did, none of them had ever been crawled. The table is in the update below.
Google's own guidance is narrower than the guides are. It says that if your site's pages are properly linked, Google can usually discover most of it, which is why a sitemap is optional below roughly 500 pages. The requirement is a link from somewhere, not a hub page as a destination. The hub is a convenient place to keep the links. The links are the part with evidence behind them, which is the same conclusion the internal links post reached from the opposite direction, and the one the orphan pages post started from.
What I would do differently
- Ship the links without the hub first.The member-to-member sibling links are cheap and are the part Google's guidance actually asks for. The hub page can wait until something needs a destination.
- Check the aggregate phrase has demandbefore naming a page after it. If nobody searches "vehicle sim games", a page called that has no ceiling to grow into, however well it ranks.
- Pick controls that already have clicks. Three untouched pages at 0, 0 and 11 clicks cost me the ability to grade a prediction I had waited four weeks for.
- Log the prediction before the change, with a date. Both of these came due on a date set four weeks earlier, and that is the only reason I graded them honestly instead of reading the 158 and calling it a win.
One piece of housekeeping I owe this experiment: when I decided in September 2026 that platform-specific lists belong inside their parent cluster rather than in clusters of their own, that decision called for a reciprocal link back from the parent. We have not built that link slot, so a PC variant still links up to its parent list without the parent linking back.
Eight weeks later: 2 clicks, and most impressions on URLs I had redirected
Eight weeks after the first hub shipped, every hub URL on game-scout.app together earned 2 clicks and 154 impressions in 28 days. I found out by accident. On 10 October 2026 I was choosing the next cluster to give a hub, and the assistant's table printed "n/a" in the traffic column for two clusters that already had one. I read that as zero:
why there is no click on vehicle sims or crime investigation ?
The "n/a" was a shortcut, not a data result. The member lists in those two clusters had 329 and 226 clicks in 28 days. The hub pages themselves had 0 clicks and 25 impressions between them, and nobody had checked whether the three hubs with zero impressions were indexed.
That reply counted only the hubs' current addresses. On 11 October I pulled every hub URL, old and new, for 13 September to 10 October 2026 (Search Console held data through 8 October):
site: sc-domain:game-scout.app dimensions: ["page"] 2026-09-13 to 2026-10-10
/best/vehicle-sim-games 1 click 97 impr pos 7.2 (old URL) /best/sci-fi-games 1 click 31 impr pos 6.5 (old URL) /browse/action-combat-games 0 clicks 14 impr pos 8.7 /browse/job-life-sim-games 0 clicks 7 impr pos 12.9 /browse/vehicle-sim-games 0 clicks 5 impr pos 4.8 crime, horror, sci-fi at /browse no row returned
2 clicks and 154 impressions across every hub URL. 128 of the 154 impressions are on two addresses that have redirected elsewhere since 1 September.
Over the same 28 days the member lists in the vehicle-sims cluster alone returned 347 clicks and 21,336 impressions. The gap this post reported in September has not closed. The hubs moved from /best to /browse on 1 September 2026 with permanent redirects, and six weeks later Google is still showing the old addresses for two of them.
Three of the six hubs have never been crawled
Three of the six hub pages have no crawl on record, 40 days after they went live at their current addresses. This post has told you since September that a page with no row in Search Console is not visible rather than zero, and that you should check whether it is indexed. I had not done that for my own three. On 11 October 2026 I inspected all six:
hub Google's state last crawl
/browse/vehicle-sim-games Submitted and indexed 26 Sep
/browse/action-combat-games Submitted and indexed 9 Sep
/browse/job-life-sim-games Duplicate, Google chose a
different canonical 2 Oct/browse/sci-fi-games Discovered, not indexed never /browse/crime-investigation-games Discovered, not indexed never /browse/horror-survival URL is unknown to Google never
Read the last column, not the label. On pages Google has never fetched, "Discovered" and "unknown" describe the same fact, and I made one inspection per URL, so the fact I am carrying is the missing crawl date. All three uncrawled hubs are in the sitemap. If you want the opposite failure, where the tool says not indexed about a page that is, that is the false not-indexed verdict post.
The other three are not clean either. The action hub was last fetched on 9 September, the day it was published. The job and life sims hub is fetched and set aside: Google chose /best/simulator-games as its canonical, which is an old filter page I retired into that hub by redirect on 3 September. And the two old hub addresses were last crawled on 19 and 28 August, before the redirects existed, which is why Google still serves them.
We have not fixed any of these three problems yet; the inspection is one day old.
A page got indexed under a hub Google has never seen
One member page went from not indexed to 23 clicks while the hub above it stayed unknown to Google, so the hub cannot have been its crawl path. best-vampire-games sits in the horror cluster. It was published on 28 July 2026 and sat at "Discovered, currently not indexed" into September.
url https://game-scout.app/best/best-vampire-games
coverageState Submitted and indexed lastCrawlTime 2026-10-07T21:27:46Z
url https://game-scout.app/browse/horror-survival
coverageState URL is unknown to Google lastCrawlTime none
The list is indexed and was crawled four days ago. The hub it reports to has no crawl on record.
The vampire list earned its first impression on 15 September and had 23 clicks and 1,133 impressions by 10 October. Two things happened to it at the start of September: the page was rewritten on the 1st, and the cluster's member-to-member links shipped on the 2nd. I cannot separate those two. I can rule out the third, because Google cannot follow links on a page it has never fetched.
That is the first item in the list above, measured instead of argued. On 11 October 2026 a seventh hub shipped, for the family and cozy cluster, in the same pull request as that cluster's member-to-member links. It is one day old and will not be readable before November.
How to check your own hub pages
Open Search Console, filter pages by the hub's URL, and set the window to at least 28 days after it was published. Then read the two columns together rather than one at a time:
- Impressions near zero, position good: the page ranks for a phrase with no demand. Renaming it toward a phrase people search is the fix, not more links into it.
- Impressions healthy, clicks near zero: a different problem, and a snippet one. That is the impressions but no clicks case.
- No row at all: not visible, which is not the same as zero. Check whether it is indexed before concluding anything about its performance.
If the hub is a crawl path rather than a destination, that is a fine thing for it to be. Just do not count it as a page that will earn traffic, and do not let it stand in for giving your real pages links. More of the AI-assistant SEO series works through our own data this way, and the free triage report checks whether a URL is reachable and indexable in the first place.
FAQ
Do hub pages help SEO?
The links a hub creates can help. The hub page itself usually earns almost nothing. We built six hub pages on game-scout.app between 18 August and 9 September 2026, and in the 29 days to 15 September 2026 all six together returned 1 click and 31 impressions, while three member lists in one of those clusters returned 326 clicks over a comparable window.
Why does my hub page rank well but get no traffic?
Because ranking and demand are different things. Our vehicle-sims hub averaged position 5 and collected 5 impressions in 29 days. Position 5 on a phrase nobody searches produces nothing. Hub pages target the aggregate name for a topic, and people search the specific one: not 'vehicle sim games' but 'best trucking games'.
Should I build a hub page or just add internal links?
Add the links first, because they are the part with evidence behind them. Google's own guidance says a sitemap is optional if your pages are properly linked, and says nothing about needing a hub page as a destination. A hub is a convenient place to put those links, not the thing that earns the traffic. When we inspected our six hubs on 11 October 2026, three had never been crawled by Google, and a member list had been indexed under one of those three anyway.
How long should I wait before judging a site structure change?
Longer than four weeks, and only against a control group that has measurable traffic. We logged a 4-week prediction on 18 August 2026 and it came due on 15 September. The number cleared, but our control pages had 0, 0 and 11 clicks between them, so they could not show whether the site would have moved anyway. A control with no traffic is not a control.
Check the demand before you build the structure. Create a free account, connect Search Console, and the MCP serverlets your assistant read your site's topical map: which topic families you cover, which are thin, and which demand you are not appearing for. That is the pull that would have told me "vehicle sim games" had no ceiling before I built a page for it.
In your AI assistant
Ask your AI about your own site
TrafficTriage plugs into Claude, Cursor, and ChatGPT. Your assistant runs the 8 checks, reads your Search Console data, and can even check your site every morning. Free account, no API key, nothing to install.