The TRUTH about Linking

[MrFewkes]

Hi all,

I thought I would share some results regarding linking (observations).

Basically a month or so ago I was pointed to a tool which shows 3 types of link graphs for a URL.

Monthly
Cumulative
Normalized

I became interested in the Monthly graph - as this shows me - month by month - approximately how many links a URL has attained in that period.

The results are SHOCKING.

Why? Because I have been scared witless about gathering links too fast, uneavenly, off topic etc.

So I was shocked to see graphs showing massive spikes, and massive quantities of links for a given month.

Basically I looked at my competitors who are ranking solidly above me - I wanted to see what I needed to do to catch up regarding link quantity.

Let me share some thoughts on my three previous fears.

Getting links too fast - RUBBISH
I have seen sites with link spikes of thousands of links in one month - then next to nothing in the months around the spikes. I have seen sites with 80,000 links gained over just 6 months ranking solidy for 1 word terms!

Uneavenly - RUBBISH
The link rate does not have to be a steady rate of aquisition. Ive seen sites rank solidy with a monthly link graph that looks like a profile of the Himalyan mountain range.

Off topic - RUBBISH
Ive seen sites ranking solidly with so many links that they can not POSSIBLY come from on-topic sites - not a cat in hells chance. Links from off topic sites appear to me to be just as good as those from on topic sites.

Another thought - theres no way on this planet that a site can write to 80,000 other site owners in 6 months and get 80,000 non-reciprocal links back. Therefore - its open season on most forms of links spam.

Also - bear in mind that not all of these links are DO-FOLLOW ;)

As I say - in all truth as far as I can see it is open season on linking - and if I were you I would steer well clear of scare mongering know it alls and do some heavy link spamming yourselves on some throw away domains if you must see for yourself.

Given what I have observed - google doesnt do half the checking people seem to think they do.

Im an telling it how it is - no scaremongering from me.

Enjoy.

[tedster]

It's good that you're wading into real data for yourself. I often see many things that cause me to rethink the accepted "Google gospel". It's true - Google spokespeople give webmaster advice that they do not consistently execute the way you'd expect, listening only to their public words.

We have a similar discussion in this thread: Is Google blind to spammy tricks? [webmasterworld.com] where I wrote up a more detailed version of how I see it. In short, Google's mission is not to run a fair and equitable contest for webmasters. Their mission is to put the best results they can in front of their users.

You often can push the envelope without any trouble. However, if a site begins to disrupt the quality of user results, especially for important queries, that's often when the alarm bells go off and something happens to the ranking. Otherwise, if their users seem happy - statistically that is, according to the machine learning they execute - many practices seem to slip by.

Another thing Google will do is just zero out the effect of some spammy practice, but not penalize for it. So the URL is not ranking BECAUSE of the spam links, but in spite of them.

[Whitey]

You often can push the envelope without any trouble

I think that especially in competitive niches this is going on all the time. Then every now and then out comes some serious FUD ( e.g. JCPenney etc ) and actions that self promotes about the ranking loss or penalty that a site has incurred. Defendants vary from folks that depend on link sales, to those that do it modestly and see no evil.

The truth about linking for me is that it's necessary, it shouldn't be the ultimate focus of a site and it's abuse and use is only tempered by the fear of getting dropped or penalised. Over time i think links will continue to regress in favour of alternative signals that are harder to game and therefore more in Google's control.

Trophy terms in popular verticals leave folks badly exposed to being flagged and open to human scrutiny for link building.

The game for those that rely on link building, largely is, when have you pushed the envelope too much, and if you can pre empt it , get out of it or will it cost you in lost business.

In most e-commerce verticals , link building is a struggle and Google probably understands the lack of natural linking out there. Most if it is contrived and artifical by some degree.

[wheel]

It all works fabulously until you get a hand review. How often once gets a hand review is another story. But it's certainly something to still be afraid of.

In fact, it's pretty much the only thing stopping me from buying $100 worth of penny links.

[walrus]

In short, Google's mission is not to run a fair and equitable contest for webmasters. Their mission is to put the best results they can in front of their users.

The best results that they can give while still maintaining that essential CTR on their search results pages ads. That has to be quite a pickle.

[bluntforce]

It's pretty easy to add a couple thousand links, even if you don't really mean to. Find any prolific blog or site with many pages and get a site-wide link.

If we make the assumption that the site-wide is really only counted as one reasonably strong link, there could be ranking improvement while the rest of the graph is just noise.

Personally, I wouldn't throw caution to the wind, but take baby steps over to the darkside.

[tedster]

Along these lines, I recently analyzed backlinks for a business that runs two domains - example.com and exampleblog.com. Google acknowledges only the links between those two domains, but in reality they have about 800,000 paid dofollow links.

All those links are obviously an ad buy when you visit the pages. Very clear example to me of Google just saying "we know they're all ads, we'll just nofollow them in our own webgraph.

When I analyze backlinks, I also prefer looking at linking domains and not total links most of the time. It seems that run-of-site links, even in a blogroll, are counted only as a smidge more than a single in-content link at best.

[kd454]

Yep all is good until the hand review comes.. Like tedster stated if you come up in certain areas it will happen sooner or later, may take a long time but it will happen, just enjoy it while your there and don't be shocked when it happens.

Had a site I was playing with come up for a keyterms I did not want it to, was an awesome key term but way to much exposure, took about 2 years now at -60.

In my experience link penalty's are usually a -60, had to find out the hard way :)

There are also other filters that you will get slapped with as well, basically you just get stuck in the serp and no matter what you do your just stuck. Have had a site stuck at #14 - #17 for almost a year now, never seen this one before until it happened.

[bluntforce]

For a single keyword I follow two domains show in first page results. Somewhat unusual as they are both parked pages.

There are things we can eliminate from ranking factors, they don't have strong internal links and the pages aren't text optimized.

One apparently had a recip linking program going, then dumped the outbound page(s) making all the links one-way inbounds.

The other is forum profile driven.

Of course I could "report" them, but there's so much to be learned from the 4 months or so they've been ranking.

The intelligence is worth far more, IMHO, than the somewhat momentary monetary loss.

@kd454
If you're playing with paid links, it's best to stay within a given vertical. I can explain any link, and the reasons for it's existence to any competitor or a search engine.
Advertising isn't against anyone's guidelines, it's when you move to manipulation that a problem could develop.

All just my opinion, everyone's got one, just like an (nevermind, it'll get deleted).

[TheMadScientist]

All just my opinion, everyone's got one, just like an (nevermind, it'll get deleted).

That's so not right ... Don't you know bait-n-switch is strictly against WebmasterWorld's TOS? Please put the text back in and let the mods delete it like you're supposed to so those of us who make it by before they get to it get to see the full thought. Thanks! (lol)

[bluntforce]

I've got to build my post count under this name before I could even attempt to push limits.

For those who push the TOS, I salute you.

[tedster]

Sorry you don't know that joke, Mad. The actual word is on our filter - it wouldn't ever be live. But you can use Google Suggestions for "opinions are like" and that will point you in the right direction. Funny thing, even Google Suggestions censors the American spelling of the actual word, too. Suggestions are pretty heavily filtered for adult words.

[TheMadScientist]

N0ZES? [Didn't want it to get Filtered]

Actually, I've heard it before, I was just feeling like being a smart[one] ;)

[Shaddows]

WRT the OP, I think the thing to remember is that the rules are different for each player. Once you are established, you can do rather more than a newcomer.

Also, I find more useful metics to be "rate of change" rather than absolute. Getting 10000 backlinks in a month is more likely to be an issue for a site that only got 100 in its first year, compared to one that already has a million.

Then, to echo tedster, there is the fact that most dodgy links are merely discounted. The penalisation of links absolutely does occur- without the FUD, link spam would be even more prolific than it already is.

[BenFox]

As it happens MrFewkes I've recently been playing around with that same tool (I'm guessing we all know which one it is) and I found a few interesting things that are worth bearing in mind. Just personal opinion I got from playing with the data in Excel but I thought you might find it interesting.

First off, like Tedster said, you've really got to look at linking domains and not at raw link volumes - it can tell a totally different story .

Secondly I found that the tool seems to have followed links from Google Display Network or Adsense or something because I found a period (08-09) where my site's aquired links shot up with links from MFA sites (no we hadn't been buying links).

Thirdly I noticed that using article marketing you can get content syndicated onto 500 new domains a month but I know from personal experience that although this does creates a big spike it doesn't actually move the needle when it comes to improving rankings (please feel free to correct me if I'm wrong on this one). So I started thinking that if Google applies a filter to some links it might be worth trying to do something similar with this data so I stripped out links from low quality pages (using that tools quality metrics I removed anything with a score of less than 3) and it totally changed my appreciation of the data.

Fourthly I used a SEARCH function in Excel to determine whether the anchor text of a link was branded or not (whether it contained the site's brand name). Plotting the aquisition of branded and none branded anchor text next to each other was amazing. In some cases you could see where sites had hired their SEOs just because the pattern changed so much. Unfortunately before I could run an analysis on some sites with statistically significant link profiles (i.e. national brands) my monthly download allowance using that tool ran out.

[MrFewkes]

Theres some seriously good comments in here in response to this. I am surprised at how "naturally" my post has been taken on-board - this shows me that I am / was a bit behind the times when it came to link spam.

Its as though you guys are saying - yeah - we know its fine.

:)

Dam and Ive been scared about some of my domains after buying links and doing the profile and blog spam. Nothing to worry about after all.......

Or - is there? Above it seems the human review topic crops up a few times. Well - I suppose its true.... it could be something to worry about.

Then again - what happens when the best site there is embarks on link spam? Google must want to retain the best site in the top ten across the board (so sites not site).

They must want to hit it down due to link spam - but boost it up as its great. A quandry for them.

I look at these penalties issued against big companies over the years (Like BMW) as just a way to strike fear into us smaller guys. Showing us what can happen - like the teacher in the classroom cracking down by picking on the biggest guy.

I dont trust google as you know - but now I see them in an even worse light.

That said - I think the next 6 months are going to be some very happy spamming days for me as I buy up all the latest spamming tools :)

[Whitey]

Had a site I was playing with come up for a keyterms I did not want it to, was an awesome key term but way to much exposure, took about 2 years now at -60.

That's accurate. Exposure , subsequent penalty and time lapse. Makes me think that Google has problems algorithmically applying penalties on some link profiles and graphs, and relies more on flagging and hand reviews for back up.

[tedster]

I agree with that assessment, Whitey. And unless there is a lot of exposure for the URL, the spam promoted results don't affect Google users all that much.

Its as though you guys are saying - yeah - we know its fine.

We're saying sites can rank even when they don't follow Google's linking guidelines. But the ticklish part is how long that might be in any particular case.

Make no mistake, Google can and does penalize for spammy backlinks. Recently this tends to happen in big sweeps through major parts of their web graph. Then you'll see a Google blog post say something like "We changed the way we are scoring backlinks" or something cryptic like that.

[Ummon]

Could anyone point me in the direction of this tool. I am interested

[tedster]

We normally do not discuss specific tools because our members here create and market so many tools that we can start a marketing war. I'm making an exception in this case because there are only two major providers that do independent crawling of the web for links.

One is the SEOMoz Linkscape tool, and the other is MajesticSEO. That description is the kind of report MajesticSEO offers.

[aristotle]

It should be noted that the sudden acquisition of a lot of spammy backlinks might not be the fault of the website's owner. Last year one of my sites suddenly picked up more than 2000 backlinks from spammy .cn (China) domains. I believe that they were auto-generated by some program or script. I was worried that they might hurt the site's rankings, but luckily that didn't happen.

[CainIV]

The biggest thing to remember is that actions on Google's part are not often immediate, especially in the case of off topic / irrelevant links.

However once the website hits top ten for any money terms, there are now eyeballs on the website - and backlinks - and this is where a spam report might be viewed, if initiated from a competitor.

It's much more difficult in my opinion trying to assess and learn what links demoted a website than it is to simply build a strong, quality link program from the beginning.

And if you are doing this with only high quality links, you will find you need less links than you might think to achieve favorable results.

[tedster]

However once the website hits top ten for any money terms, there are now eyeballs on the website...

What do you think about a more intensive automated analysis as well? It might make sense to conserve computing cycles for URLs that are not getting a lot of SERP impressions - but to kick in something more intensive once those impression numbers start to jump.

[limoshawn]

and this is where a spam report might be viewed

so, when/if google does a hand review, are they looking at the quality of the site or the quality of the signals that google used to put the site atop the results?

[TheMadScientist]

Both?

What would you do? You have nearly unlimited resources ... You used algorithms to order your results ... Would you review the site, the signals or both?

You have to check the site to see if the report is true or false ... You have to look at the signals used to adjust and improve your algo ... I'm almost certain they look at both.

[Whitey]

on Google's part are not often immediate

Absolutely . Some site's seem to sit there for months or even years from the day they rank , and then bang , gone. It's really like Google editors are sitting though a backlog of flagged sites, taking on priorities to investigate and clear.

For the ranking / linking gamers putting a trusted site into this vulnerable position needs to be a very calculated risk / benefit consideration from a commerical point of view.

[tedster]

Speaking at SMX this past week, Matt Cutts confirmed a good bit of what this thread is saying. Vanessa Fox reports:

Matt reiterated that competitors generally can't do anything to hurt another site. The algorithms are built to simply not value links that violate search engine guidelines. Demotions generally only occur when a larger pattern of violations is found.

[searchengineland.com...]

[CainIV]

I would agree Tedster - there might likely be a hybrid process, which might take into consideration the total number of spam results submitted - which then triggers a 'semi automated' process which focuses more cycles on any particular website.

If more flags are raised, the website is then processed into another queue - the manual review.

As well, particular verticals would have more workload focused upon them due to the nature of the vertical itself - health being one example. More sensitive genres might see a website then reaching a manual review faster than others.

The algorithms are built to simply not value links that violate search engine guidelines.

Demotions generally only occur when a larger pattern of violations is found.

The problem with this statement is that in "million dollar" verticals both statements conflict.

Competitors ARE able to spend money to create a "larger pattern" of inbound links to a competitor site especially in the case where the competing site is in those very lucrative top 3 positions. I am assuming this is what we are referring to in 'violations'

[tedster]

Completely the case. Some of those million dollar verticals seem to be ranked by a completely different algorithm. But even so, the business of taking out your competition has certainly been thriving for many years.

In some verticals, I'd guess that more resources go into "disruptive SEO" than into straight-up marketing. I wouldn't want to be the one responsible for sorting all that out.

[justawriter]

However once the website hits top ten for any money terms

That may be something of a generalisation - I'm not seeing that in the money terms that we're working in for our clients.