A couple years ago, I posed some questions from the SEO world that I couldn't answer. Tonight, I'd like to repeat that process and throw up some dilemmas that, once again, have me in a quandry.
- Does QDD Give an Inherent Boost to Negative Subject Matter?
Some rumors have been passing around the SEO world that QDD, Google's "diversity" algorithm that's intended to show more of a variety of different kinds of content in page one of the SERPs, gives a rankings boost to negative focused content, making it easier to rank for a company/brand/person's name if you "talk smack" about them. True or false? - The 302 Hijack is Back?
From 2005-2007, complaints were flying about Google showing pages that 302 re-directed to other URLs in place of the originals, allowing clever spammers to conditionally 302, then steal the traffic when actual visitors clicked the ranking search result. 2008 was pretty quiet on this topic, but recently, I've seen it flare up again. What happened? Did Google make a misstep? Or is there a new, more sinister version of this loose on the web that gets around the old protections somehow? - Are the Engines Following URL Shortening Services?
Some recent evidence has suggested that engines are using URLs from shortening services like TinyURL, Is.gd and others (even those that don't cleanly 301 as Zi.ma does) for discovery, and possibly flowing link juice through them as well. Has anyone experienced this, tested it or seen results that would prove the case one way or another? - What Was Up with The Hyves Subdomain PR Penalty Checker?
Marcus Tandler leaked the news that using the subdomain "hyves" attached to any root domain on the web would give you a PageRank number indicating whether the domain had suffered a PR penalty for buying/selling links. So many questions on this one - who leaked it? It had to be someone inside Google, right? No one could guess that randomly without at least hearing a whisper from the grapevine. Why would Google create it? Why would they leave it active? Why would they make it publicly accessible in the first place? Every search quality engineer has access to the console to pull up their internal stats on a domain, so they could easily mark it there... So weird. - The Engines All Regularly Follow Many More than 100 Links Per Page?
We work on a lot of pretty authoritative, powerful domains, so I'm wondering if this is just a fluke from being in a link-rich environment, but we see that Google, Yahoo! and Live/MSN are consistently going up above 300 links per page and following and indexing them all. Is this a limited behavior set, or do small sites, newer sites and those of you with test domains see this activity as well? - Is it Really Harder to Get Rankings with a .info, .cc, or .biz Extension?
We've heard from a few sources that these three extensions, along with several other International ccTLDs, might be lowering trust scores or increasing the probability that your site is flagged for exclusion/devaluation/penalties. I haven't done nearly enough testing on domains like this to know. Of course, I don't think they're good for branding, which is a big part of a long term web marketing strategy overall, but that's beside the point. - Does Google Employ Link Buying Moles?
I've now heard tales from 2 different companies, one very prominent in the industry, about all their clients being manually penalized by Google and their link networks and link buying sources identified with immense precision, as though there was an insider leaking the data. I generally have a tough time believing this, since Google usually likes to do things in a very scalable fashion, and hiring moles to spy on link buying activity seems to my mind a very low ROI and bandwidth intensive endeavor. Have you heard/seen anything that would sway you definitively one way or the other? Is Google really conducting corporate espionage with those who would violate its quality guidelines? - Will Anchor Text Value Pass Through Terribly Low Quality Links?
A friend of mine recently hypothesived that while link juice might be compromised or even discounted entirely from spammy, low quality domains and pages, anchor text value could still pass. This phenomenon supposedly is why more aggressive SEOs are buying/acquiring tons of super low quality links from crummy directories, old sites that have lost most of their PR and open comment spam areas. Is there any truth to this? Why would an engine discount query independent metrics like link juice but continue passing anchor text value through links?
Hopefully, you've got more answers (and evidence) than I.
I've been wondering about some of these too, but I can only weigh in on a few:
#3 - Are SE's following URL shortening services
I think so, but for discovery and not link juice. However, I'm still wondering whether enough weighed together can have a direct ranking impact.
#4 - Hyves
My guess is it wasn't for real Google employees, but for the legion of manual reviewers who don't have access to the console. It would be easy to add a button to their toolbar to show if a domain is penalized using the same method as the normal toolbar uses to check PageRank.
# 6 - Are .info, .cc, and .biz domains harder to rank
I've consistently found new .info domains take weeks to get indexed, where a comparable .com domain would be indexed in a few days at most. My parked .com/.net domains usually show up in Google's index with the parking page's text, but parked .info's generally don't.
I haven't bought many .biz or .cc's, but .cc is a country code (Cocos Islands), so it would make sense for it to be harder to rank in Google US.
# 7 - Does Google employ link buying moles?
I think its a far fetched to think they try to get employees working for the company, but I would be surprised if they don't sign up for accounts at TLA to see who's selling links. That's just too easy to pass up.
# 8 - Do low quality links pass anchor text?
Probably not, but everyone seems to be missing the point in having tons of low quality links - its a diversion. As Y!SE is the only real way to find your competitors links, and it doesn't filter out garbage links, it becomes a lot harder to find their value passing links (paid or otherwise). Even if you know your competitor is buying links, good luck finding them when they've got 30k links in Y!SE (hint: Linkscape helps a lot).
Rand
Google doesn't need to PAY moles to detect paid links. They convince some SEOs who can't compete, want to curry favor with Google or are looking for something to write about to turn in other sites. SEO's out paid links, PR passing affiliate programs and other strategies that reveal Google's "vulnerablities". Jealousy, pettiness and short sighted self interest are low cost and infinitely scalable.
Even without legions of self appointed guardians, detecting paid links is only hard when the link seller take time to do a good job. Spammy footers to Porn, Casino's and Sub Prime Mortgages are really easy to spot, as are old blog posts that are magically updated with links with super targeted anchor text.
Of course, don't read too much into what happens in one verticle (such as porn, financial services or SEO). Google's webspam team has proven over and over again that a little guile goes a long way (hi Matt), so don't be too surprised if these spammy directories blow up in your face elsewhere.
I think point 8 is the one everyone would really like to know for sure however it is impossible to ascertain.
I regularly 'spy' on sites ranking at the top of high competition niches and am sometimes very surprised. I won't mention any names but if you check out the 'make money' top ten there are certain sites within it who have around 50,000 backlinks 85% from low PR pages. However some of the anchor text 'make money' is used around 5000 times.
I am firmly of the opinion that anchor text helps to create the over all keyword theme even if from a low quality page.
I'm in agreement on this. In fact, I thought it was a known fact and was surprised to see it on this list.
To me it makes sense logically.
Imagine you're Google and you're trying to figure out what a site is about.
If you have 1 link from a PR 8 site to seomoz.org with the anchor text "seomoz", but you have 1000 links from PR 0 sites with the anchor text "seo company"... if you were Google wouldn't you think those 1000 (supposedly individual) links would be more "votes" than the 1 link. Even more so if they were variations like "cheap SEO company", "SEO company Seattle", etc.
At least, this has been my experience.
Clearly if you're trying to work this kind of keyword ranking you're better off with the 1000 or so links.
I also firmly believe that even though there might not be any link juice that is passed with "nofollow", the anchor text of nofollowed links is definitely taken into account.
I think a link building strategy needs to be built around authoritative and anchor text. If you can't do both together (and sometimes you can't), then you need to split them up. Get authority through one kind of link, get anchor text through another.
vingold - My experience supports what you are saying 100%. Moreover, I have seen many sites that rank quite well with site wide inbound links. However, I always hear people talking about how site wide links are discounted. Bah! From what I have seen, site wide links work quite well.
Here's what I think: Google likes natural links, or at least natural looking.
So if your links look "natural", then you'll be ok.
For instance, it is OK if you have some inbound site-wide links. After all, if you have a nice site you'll likely end up on a few blog rolls and those are likely to be links present on every page of their site. You just can't have ALL of your inbound links come from site-wide links.
Also, you can have some anchor-text optimized inbound links - you just can't have ALL of your inbound links be anchor-text optimized.
I think most of the "experts" that say site-wide links are bad are really saying they wouldn't buy (or spend time pursuing) site-wide links.
I would find it hard to believe that any expert would turn down free site-wide links or that they would say it was a detriment to their rankings.
I think it's all a matter of propotions. 1 against 1000 is one thing, but what about 1 from PR8 and 50 from .info domains? Or 5xPR8 agains 1000 from low PR pages. It's all proportions for me :)
Rand - nice thought provoking list. Re #7, I have first hand experience with a Google mole. Not exactly what you describe, but I have talked to an individual who was applying for a job at our company who claimed he was being paid part time to investigate sites that were showing up in the Google SERPs.
Obviously I can't prove he was telling me the truth, but he had no reason to make this up.
Re link buying moles: while I agree that Google likes to do things in a scalable fashion, it doesn't necessarily follow that they would hire the said moles in order to discover people who buy links.
An alternative purpose could simply be a human benchmark for an algo change, a statistical quality check.
Mind you, if they find link buyers this way, they might as well penalize them; but that doesn't mean that is the goal of the process.
Rand,
For number 3 how are you suggesting that the engines are discovering these links, through there use on twitter or through the shortening services themselves?
Reply from Rand via email,
"I think there's a lot of them on the web, many on Twitter and plenty in RSS feeds and re-published tweets, too."
Good to know and I'm excited to see what happens.
the playing field is not perfectly level across domain extensions - I discovered a serious problem with .US and made this blog post https://www.internetinc.com/the-dot-US-domain-handicap
about a week after making that post and posting on Matt Cutts blog, sharing with my AdSense rep etc., the disadvantage was seemingly corrected. I still suspect that some extensions rank more easily than others. however, all of the extensions that I'm using are getting ranked whereas initially, .us was the kiss of death.
in my mind, it does go to show that all domain extensions are not created equal... and that Google can be surprisingly responsive for such a large company.
-- Eric
Some points on number 4: It doesn't have to be leaked by google: in the Netherlands, directory pages with subdomains on different topics are very common (startpagina.nl,startkabel.nl and about 100 smaller ones). Some of these directories noticed a very high PR (7) on the 'hyves' subdomain of their directory that could not be explained by their link profile. I guess this got the ball rolling, some dutch seo-experts picked it up and discovered the trick to check bans. I first read about the strange pageranks of hyves subdomains in august '08 on this Dutch forum (https://www.tbforum.com/thread/114700.html)
Still, it was pretty strange, too bad it doesn't work anymore.
Regards,
Toon
Hey Randfish, Point 8 on your list is something that I've been mulling over in the last week. I was investigating why a certain handful of sites seem to dominate some of the most competitive keywords in the UK finance space.
Certain sites are ranking in the top 3 of google.co.uk (you'll be able to work out which ones by doing a back-link check and spotting the 25 million+ links!) for big terms like 'credit cards' and 'remortgages'. From what I can see the bulk of these links are being driven by the Digital Point Co-op Advertising Network. I'll let you all be the judges on whether the anchor text links (mostly in the footer of area of these sites) are worthy votes to allow this site to rank so highly. Either way seems to point to evidence that 'anchor text value' from low quality sites is a factor.
Does anybody else have any solid examples of where this seems to be the case?
Gary Preston
In any analysis, I would take the UK financial industry out of it at the moment. There are too many things going on there for analysis on a single factor. Knowing which sites rank in the top 5 for these sort of terms, the sheer volume of links they are getting naturally and through SEO efforts is overwhelming.
Also bare in mind the fate of Go compare last year and their subsiquent re-inclusion, the amount of money involved in these businesses makes it well worth their while (in many cases) taking huge risks, but they have the ability to re-invest and recover relatively quickly.
I'm not saying that they aren't buying links, but that they shouldn't be used to draw conclusions to be applied to other industries, and other websites.
Hey Yoshimi,
I agree that there's a lot more going on than just 'high volumes' of low quality links in this case. However, one of those sites (I'm trying not to name names here..) has 26 million+ back-links.
The others have approx 2+ million links. Now that's quite a difference and worth taking a closer look at I think when addressing Rand's question "Will Anchor Text Value pass through low quality links".
I'm certainly not drawing any universal conclusions from this one example - its just one that came to mind when reading the article.
Cheers
Gary
I would assume that the site you are liooking at is one of the aggregator sites, and probably one of the really big ones which covers every financial and travel product known to man kind, probably a bit like a supermarket?
The point wasn't whether the effect exists or not, just that looking at the UK financial sector to see it will be like banging your head against a brick wall for an hour befor groping around in the dark to find a ping pong ball. If you want to look at trends and optimisation techniques, this isn't going to be the industry to examine...unless you have a spare decade?
Hey Yoshimi,
No - not them :) I've just PM'd you...
Yep - UK finance niche is a complete wild west. Part of the reason they all get away with it is that they all follow the same tactics, so if Google targets those tactics they end up without sites in their index that people expect. Now and again Google rattles the cage though.
It's a great niche to follow, and if you want a good blog that covers it then take a look at Insiders View (cant rem the URL, but shouldn't be hard...)
Honestly, I think that if Google thinks that these websites employ tactics beyond the Guidelines, they won't take actions because the sites are valuable and relevant in other factors, not because they won't have material for the SERPs :)
Link Buying Moles? If I were Google I would hire a battalion.
Link buying - unless it's an established, respected directory - is pretty sad
David
Back To UK Soon :)
#8 - i have seen some fairly decent results using spammy links. I think it;s like owt, if you throw enough muck, some will stick. Obvisouly, higher powered links pass more juice but if you get enough small time links they can gather a fair bit of power.
I try to mix and match but build a solid layer of quliaty links to insulate against algo changes.
I've been asking this as well. But, if Google were doing this, don't you think someone would leak this and we'd have rumors galore about 'Google is in bed with link brokers'.
I just think either they are using Linkscape (great data...you can find paid links easily) or that other expense link-analysis or they made their own in-house search engine or just had someone thumb through Yahoo's link data.
That's my 2 cents. The rest. I have an opinion but no data that I want to share (yet...). :)
1) I think you may be confusing cause and effect
2) I saw a classic 302 'jack a few months ago - because the issue never went away
3) Dunno
4) Remember -link:, and _link: ? Same reasons all round, I suppose
5) Um, I don't recall there ever being a hard limit on the number of links a spider would follow. You used to be able to pull some nice strokes by placing links below the overall *indexing* limit of a page, and G have recomended 100 links / page as a reasonable maximum, but some DMOZ cats are HUGE
6) Yes and no. Some TLDs get a "conditional flag", like a suspended jail term. If they play nice, it never affects them. If they get other flags, they conditional ones count too
7) >> Is Google really conducting corporate espionage with those who would violate its quality guidelines?
Thats an invalid question; think it through. If you don't accept that manipulating G's results is a problem, you can hardly complain about ANY action they take to defend those results. If you do see a problem with manipulating SERPs, then you aren't buying links, and the issue is irrelevant to you
8) I've always been of the opinion that "juice" and relevancy are calculated separately, so high volume / low quality links will provide value for specific terms, but not for general rankings across all terms
Regarding no 5, I've read on the official Google blog and on Matt Cutts' (I think) that the 100-links-per-page rule of thumb was actually a technical difficulty for the earlier versions of Googlebot, that they had since fixed. Still, the rule was kept in the guidelines because it was considered good practice to keep PageRank clusters as tight as possible.
I think there must be some truth to #8. I've seen some evidence of this in a small market where some sites dominate the rankings through a reciprocal link building network (and they use optimized anchor text for the links).
Point 8 is actually worrying, because I've noticed behavior in the same line you suggest.
From a search engine's point of view, wouldn't it be more logical to tie in relevancy to importance when it comes to the anchor text/link juice passed relation? I mean, if nytimes.com links to me using the keywords "SEO junkie", passing a lot of importance, wouldn't it also be a more relevant link than having the same anchor text on a crap link-to-us-page-link?
On my (romanian) blog, I keep preaching about how importance and relevance should be isolated and understood as separate concepts, but I firmly believe at some point Google and the search engines need to tie the two together, otherwise the situation you described can easily be taken advantage of.
Nice info there !
As far as point 6 is concerned,I would surely agree with you. It is somewhat difficult when you are promoting subdomains and domains with .cc, .biz or .info. Infact you will observe few of these domains ranking on top of Google these days.
Point 6: .info domains. Im an avid fan of the extension and have invested a lot of money as it bodes perfectly with the platform i have produced for tourist information. Surely Google cannot simply apply a blanket rule over an entire extension. My content is of a high quality and receiving on average 8 page views per visit, it seems like my users think so too. How is Google going to manage when the new .absolutelyanythingyouwant start appearing in 2010. Penalties depending on extension are in my eyes are a complete myth. My site ranks well for important terms on the citys we run the platform on especially in Yahoo and MSN and they are .info domains. For me personally they also convert well on PPC campaigns
They've stated that don't penlaized specific CC's, but there are several cases were this has been shown to be false (like the .cn fiasco).
However a key point with .info's is that they are only cheap for the first year - after that they cost as much as a .com and lose their appeal for spammers. Its always a good signal to have your domains registered more than a year into the future, but it may have special importance with .info's.
Regarding the new TLD's 2010: We've already got .jobs, .travel, etc and I've never seen one in a Google SERP. That probably has more to do with interest level than the algo, but I wouldn't expect a drastic change with the new TLDs.
Hello Rand,
As for #6, I have a couple sites that are .info, and I must say they are doing well in the search engines. However, I have yet to see them in a #1 position even though they would match the search query word for word.I'll keep adding content and try to get more inbound links to see if I can change that.
To point #5 Rand, we have seen the search engines follow over 150 links on a page and index the pages those links went to. This is one of the reasons we increased the previous 100 links per page to 150 in our site index. Although I will say that I do think this depends on the real page rank of the page that those links are on, because I have also seen much fewer links not get followed on a page that had a low, to no PR. So the tests we have run show that the number of links per page that a spider will follow depends (among other smaller factors) on the page rank of the page. If the page can support more than 100 links based on its PR than its worth moving those links up in the layout of the site. This is on a per site basis so I can't say this is true as a mass generalization of all sites.
It's probably safe to say Google indexes an entire page when a crawler first encounters it (I never see a partially cached page in Google's results) and possible they might only follow a certain number of links on the page until a certain criteria is met (which some people might have experienced as the "limit").
I'm sure Google wants to increase their shelf space for ads, so indexing new pages is a major goal for Google. However you wouldn't want to follow one million links if a page contained that many, even if the page wasn't spam, for the purposes of making efficient use of resources.
This all tells me there isn't a real limit on the number of followed links per page, just filters in place with respect to the page's perceived authority, extending or retracting that initial number to a certain amount.
Regarding the aftermath of #4, Google cache brings up hyves nl for all hyves prefixes, example https://74.125.95.132/search?q=cache:https://hyves.seomoz.org/. What's interesting here is that the Hyves Netherlands ccTLD has a pr of 7.
can anyone rate .com .net .org .biz and .info
How much is the difference? i know .com and .org rocks but why do others loose ground?
The QDD pisses me the hell off. I keep removing this bloody page only for it to pop back up again on page one, with the resulting pushed up pages dumped on page 5! Screw Google.
Rand - on #5, they absolutely do follow more than 100 links, and you do not need to be Yahoo or MSN. We use more than 100 links wherever it makes editorial sense to do so, and have never had a problem.
Of course, you should really only do it if there is a solid editorial reason to do so. They will follow them as long as you have some decent amount of authority on your site. For example, don't try this on sites of PR5 or lower. I can't say that PR 6 is enough either, or that PR is the right metric, but lower PR sites will definitely have more of a problem.
This is a little off topic, but I thought someone may have good feeback. I have been optimizing www.koleabeachvillas.com to an extent the last two years. My main search term is Kolea. The second largest search term to go with that is Hawaii. I own the domain name, www.koleahi.com. Would it be worth my while to switch domain names or is that going to take me way back?
I can answer Number 3.
Yes engines DO follow URL shortening services by passing on link juice. I'm sure it must by with the service setup to store and publish urls as I tested this theory months ago and gained backlinks/juice...having told no one my shortened url and pinged it only.
Topic #1 is probably coincidence, though I lack any sort of evidence. Since complaining is a popular pastime online, it is logical that a few popular complaint sites will rank well for branded terms. There's also probably a lot of legitimate news or review sites that could rank well for a brand, while featuring negative content.
It's not that I doubt Google's ability to create a robot that understands negative context, it's just that I don't think they'd use it to rank content. :-P
One thing that can help with #6 is logging in to Google Webmaster Tools and setting the Geographic Target to the desired location. That way a non-US specific TLD, for example, could be set to target the US.
Geo targetting in Webmaster Tools never worked for me. Honestly, I never consider it when I play it local.
6. - .info sites.
About a year ago I started buying up .info sites (as I wante to make small informational sites, and have a domain buying problem.
I am ranking fairly competatively on a few of them, in some cases outdoing the older, slightly less SEO'd .coms.
That said, If I was using any form of social link building, I would stick to a .com. The stigma (although I hate it!) still exits behind non org,edu,coms. However, for creating a personal network (as long as its legit, I don't see a problem with the .info.
#8 - frustrating as it is i am working against a website at the minute with lots of backlinks, albiet with varied anchor text, from crummy directories. Unfortunatley, i think they do still pass enough power to start to make an impact on the SERP's However, one small algo change and it's the Folrida update all over again :)
does the hyves subdomain still work, I tried checking some other sites that have "implemented" this and it says address could not be found from server
has it been discontinued
I hate it when I'm late to the dance....
First time I've heard of this and from a technical standpoint it makes zero sense. Subdomains are controlled by the DNS, not Google. We have wildcard domains set up so for any of our sites I can type in gibberish.sitenamehere.com and end up at the home page.
Has anyone actually seen it work or taken screenshots? Because it seems to me that instant buzz which immediately doesn't work anymore because "big brother Google" killed it is very suspect.
Can we submit it to Mythbusters yet?
Exactly my thoughts. I hadn't heard anything about it until today and now it's too late. I wish I had came across it when it was still working!
I'm beginning to see .biz sites show up on Google.co.uk, in one instance for quite a competitive keyword phrase.
This particular phrase was not even considered by many worth optimising for just 5 years ago but now in business its almost a mandatory activity for those who wish to survive without any or ineffective online marketing activities.
Yep putting anchor text on a low ranking page will make a difference and for many long tail keywords you only need a few instances of anchor text to score an organic page one result, very useful for smaller clients on tight budgets.
Its about time search engines came down really hard on who take the p*ss with their links - if a skilled SEO can spot it then so should a search engine, surely it cannot be that difficult?
'd rather work with someone to get a link from a relevant, ranked site (which might take time researching and then PHONING them to personally discuss) than just drop a link from some "crummy directory" (which seem to be on the comeback only this time listing "events in your area", like anyone cares! Rant over)
i've had experience with #1. it's absolutely true and rampant in niche industries like investment advice and tropical fish. angry consumer sites have a ton of in-links, so i'm not sure that its a matter of preferential treatment or just a product of page rank. i've found the negative serps are typically higher in yahoo than in google.
there are ways to get rid of them... creating public profiles on social media sites is a way that's worked for me.
8 is true too. seen it a million times. i am personally too paranoid to give it a try...
I dont think Wikipedia is included in the diversity algorithm. It is the one constant in our diverse search. In terms of inbound links and #8. If google search is still based on links and most of the authority sites stop giving link juice; what else is there but spam links. IMO that is why spam links work so well because the authority sites stopped linking out.
My hunch is that #7 would be pretty easy to track algorithmically via Google's massive data sets. All it takes is a few sites to enter a link buying network before Google sees a pattern emerge that is way askew of ordinary, and recognizes the link network. As more sites get added, the more it's reinforced.
Thanks for covering the "less scientific" SEO topics (topics that are not so easy to test and get answers about). You are truly an expert I enjoy learning from and it's nice to see you scratching your head once in a while!
Jason Nadaf
Internet Developer
Great list of questions Rand, unfortunately I doubt you get any quality or reliable answers back unless it is from Cutts or other engine reps.
Answers from anyone else should be backed up with tangible evidence posted for everyone to review.
"Business networking" www.4networking.biz go check
That wasn't supposed to be link spam, I get enough elsewhere. Sorry, SEO moz mods please remove the post above if you feel its not necessary,I was trying to show a.biz site in first position on Google.co.uk
For point no. 6,
actually there is no difference between using free domain such as .co.cc .biz and soon to get ranked. Google never discriminated free domain, they only discriminated for those who used to buy link to get ranked faster and pay someone to review their website