I'm often searching information on auto parts and Google's results are quite literally 100% webspam. Here's an example - The entire first page is the same provider squatting domains:
Man, it irks me because half the time it's not just these sites. eBay, Amazon, and tons of popular sites will come up with links for one discontinued part that hasn't been sold on their site for years, nor have any information on it. Sometimes it's just some random item listing that isn't even related to what I searched for.
Actually about 2 years ago this is precisely the query I would have chosen too.
You could always find a parts catalog, service manual, that sort of thing scanned and posted online. PDF (used to) select for the scanned documents. Used to be my main usage of google was finding service manuals for auto and electronics.
Nowadays the spam just pretends to be a pdf cause we used to always search for them.
It's pretty common for wiring diagrams, service manuals and fix it guides to be in PDF format. Sometimes you're holding broken parts in your hand and the exploded image in a PDF is linked to a part number (The part may not even have a distinct name, just a manufacturer part number) so you can order replacements.
Anything health related is just as bad. Just pages and pages of blog spam, sites with dozens of referral links and sites that aggressively optimize for SEO.
LeoVegas Indian Online Casino
Online Holland Casino
Start a Blog by Ryan Robinson
The Best VPN
Top10VPN
PureVPN
HostChecka.com
HostingAdvice.com
Web Hosting Secret Revealed
It's just paid links. Apache isn't the only one doing it, a lot are. I'm glad they are getting money, but it's paid links pretty openly on super highly ranked pages.
For others who want to learn more about this, here's an excerpt from https://www.apache.org/foundation/sponsorship.html. They really do sell "follow" link treatment as a standalone sponsorship upgrade, which explains the "sponsors" ohashi pasted.
> BRONZE
> Donation level: US$6k/year
> Text mention on our ASF Thanks page with 'nofollow' link
> Text mention location: lower section of page
> -> Alphabetical placement
> -> Shared row
> -> “Follow” link with multi-year commitment
> Use of the ASF Bronze Sponsorship logo for your site
Is this problematic because of the links? Because FOSS is funded by donations and it's common for non-profits to acknowledge their sponsors. For example EFF does the same thing [1], but FSF displays only the logos with no links [2]. Are you advocating for the FSF version?
That's why FSF's listed sponsors actually care about sponsoring the FSF. The ASF "sponsors" pasted by the first comment and many of the listed EFF "sponsors" aren't donating money because they care about the foundations. They're purchasing links in order to rank higher[1], and a link just happens to require donating.
[1]: EFF uses rel=nofollow on all links to sponsors, so a listing on this page probably doesn't actually do what the buyer intended. Google covers this in https://support.google.com/webmasters/answer/66356?hl=en. In a sense, EFF gets free money from people who think they're buying better ranking but actually aren't. OTOH, ASF's decision to specifically sell rel=follow links - which do pass PageRank - is surprising.
Wow, I'm rather shocked by the Apache Foundation doing this. It seems far too sketchy for a respected open source foundation.
FWIW, though, Google can/does penalise such links. From their list of things that can negatively affect ranking:
> Buying or selling links that pass PageRank. This includes exchanging money for links, or posts that contain links; exchanging goods or services for links; or sending someone a “free” product in exchange for them writing about it and including a link
In the UK, there are thousands of local sports clubs that have links to their sponsors. You can buy these links for small sums of money. This is the same as your apache example in theory but the intention is different, it's a clear attempt to manipulate rankings while the apache links are just a side benefit. I doubt microsoft and such would care either way?
I doubt the ones at the top are doing it for SEO. The bottom row of 'do follow' links filled with hosting, gambling and vpns? Absolutely is SEO to manipulate rankings.
Company X that wants to rank high for a keyword pays Company Y with a highly trusted website (in the above example, Apache) to place a link to their website with a variation of the keyword as the anchor text. They pay for the link, Google's algorithm boosts their position in the search results for the targeted keyword.
How would Google differentiate between a subdomain or folder that I use on my website for a different purpose, and a leased subdomain? Given that they're describing this as "penalisation", and not just "treating the subpart as its own entity"?
It's possible that the description Google used is a bit misleading. Meaning that they might actually be doing what you said - splitting out the subdomains as their own entities and the natural effect of this is to have a separate ranking, thus looking like they got penalized for using the subdomain but really they were just reclassified.
Seems like the misleading description is from the article, not Google.
After following the links I got to this tweet that agrees with your theory. They're trying to detect if the subdomaon content is independent of the main site.
I've had a small stint into SEO in 2012 and Im pretty sure that Google treated Subdomains as their own entity back then. I would be suprised if this changed in the meantime.
I spent a lot of time in the blackhat SEO world prior to 2014 or so... I can assure you that subdomains reaped a bountiful reward from their association to the main domain.
They certainly weren't equal but they got more than a little (trust/authority/ranking - call it whatever you want) help from their parent domain.
To the point that this was the case:
If I bought a brand new, never used .com and built 500 pages of spammy content and linked those pages/content from thousands of other super spammy pages... that domain is de-indexed within 2 weeks and the pages never really ranked/got organic traffic worth the cost of the domain name.
If I built those same pages on a subdomain and linked them in the exact same manner but the parent domain was very strong - the pages would rank and traffic would flow steadily, indefinitely. (As long as the parent domains footprint/linking was large enough to absorb whatever penalty that my spammy pages brought).
It's still happening today where people buy up domains that are established with good link profiles - go to archive.org and download the sites old content, re-host it and keep it alive. Then they build spammy subdomains on the parent domain - profit.
I've read that in multiple docs back then but since you observed this it could've been false. The rationale I had for it was that there are many sites with completely independent subdomains or technical resources like asset being served by subdomains. Buying established Domains and scraping old content does not rely on Subdomains. Affiliate, ads or your own product can be placed on the TLD. I dont doubt you tough and I think I will try it sometimes. Old school fun
If it was completely decoupled back in 2012 then there wouldn't be any advantage to leasing the subdomains or Google being peeved by it enough to take action.
I suspect what is more likely is that the subdomains were separate but not completely unrelated/correlated to the primary domain. Google likely assigned a certain amount of value to the mere fact that it came from the core domain. I think that is what changed here.
Since this also affects leased directories, not just subdomains, there's probably more to it. Afaik, "normal" subdomains haven't been affected the same, only those that were leased out.
Penalise is the wrong concept here. What The Google is doing is getting smarter and that intelligence is being applied. Call it a correction if you need a name for it.
Long to short, Google just wants to be human [1], but it wants to do that automatically and at scale. It seeks to reflect what a human would approve or reject. If you look at the history of SEO / the search algorithm this is the nature of the progression. Obviously.
These sites aren't being penalized. Hiding in a blindspot and being discovered is not a penalty. It's simply reality; a reckogning; a day that only the naive would believe was not coming.
[1] Actually, Google wants to be you. The attempt to personaling search is Google wanting to recommend to you if you could recommend to you.
I think github.io is explicitly registered as a shared second level domain (similar to co.uk), which also ensures browsers don't allow cookies spanning multiple subdomains.
There are two halves to the list. The first ("ICANN") half are suffixes that work like a TLD from the perspective of the DNS registry. Nominet handles registration in .co.uk or .org.uk or .net.uk the same way another country might choose to handle their whole TLD.
The rest is names whose registered owners do the same thing as a registry for the public or some subset of them, like Blogger or a cheap bulk hosting site where you don't pay for a name.
PSL listing has a variety of effects. Don't do it hoping to achieve one of them and then being disappointed by the others. Do it because you have a real public suffix.
For example many browsers won't let Cookies escape a suffix, so company A.some.example and companyB.some.example can't share cookies if some.example is on the PSL. Let's Encrypt name quotas care about the PSL so companyA.some.example certs wouldn't share quota with companyB.some.example. DMARC won't work on a public suffix, but HSTS preloading does.
I assume that previously, Google's ranking algorithm would attach a score to a domain name, and by leasing out subsections of the site, the sub-owner would get an artificial ranking boost.
So now perhaps the Googlebot notices if a subsection of a site is radically disconnected and treats it as an independent publication.
Does this include sites hosted on Google Drive, Google Spreadsheets (yes, you can host a site on Google Spreadsheets, by putting HTML in a cell), Google Docs, etc.?
I see, but I assume Google has custom ranking since I never see them show up in any search result, they clearly aren't getting Google.com's rank score.
Absolutely. Subdomains may get some extra credit, but they still need incoming links. "Leased" subdomains and directories don't come naked, they usually get lots of links from the main site, often site-wide in header/navigation, sidebar or footer.
Since the publishers get paid for leasing out the subdomain and linking to it, you could argue that they are essentially selling links. That's apparently not what Google is saying though, because they'd penalize the sites selling the links, too, and only the leased out subdomains/directories have been affected, the rest of the sites are still ranking as they always did.
https://www.google.com/search?q=scrambler+50cc+parts+format%...
Wiring diagrams in Google images - literally every result is webspam in the same format split across squatted domains.
https://www.google.com/search?q=mazda+miata+wiring+diagram+s...
These aren't even particularly niche searches, nor is it hard to detect that all these sites are the same.