How to best handle worrysome backlinks

[fathom]

---

System: The following 5 messages were cut out of thread at: https://www.webmasterworld.com/google/4775291.htm [webmasterworld.com] by goodroi - 8:55 pm on Nov 4, 2015 (utc -5)

---

We should split this thread.

@Anthony

Let us first discuss the rational about disavowing 'whatever' without knowing anything about what you are disavowing... Other than you don’t like it… SUPERB FOOTING!

Disavowing forces Google to drop that link from the link graph regardless of any impact (both positive or negative). If said link doesn't use your keywords there can be no negative impact because there was no unnatural reference to induce a positive impact that could never be misunderstood as manipulation.

Therefore you are disavowing a natural link... Might be nearly worth nothing but if the page it was on was indexed, you killed any passed PageRank by disavowing it... Thus the cautionary warning when using that tool.

[goodroi]

If said link doesn't use your keywords there can be no negative impact

Are you guessing or hypothesizing? Can you please provide the evidence behind this claim or a link to where Google made this statement. This hasn't always been my experience.

Therefore you are disavowing a natural link

Sorry but I don't understand your logic. How is it natural to gain thousands of links overnight from domains that are completely unrelated to yours? If I was trying to artificially inflate my backlinks by buying unnatural links it would probably look like thousands of links overnight from unrelated websites.

[fathom]

Google claims to be able to accurately measure relevancy, quality & quantity but reality is often different. Gaining 100k links does not guarantee a good or bad reaction. It depends on the 100k links. If they are buried on a website with no inbound links and no visitors, Google may never realize those links exist and there will be no reaction.

@goodroi your opening premise aligns to my belief as well. Although I caution suggesting your opening sentence are based on “Google’s claims” more like “our interpretation of Google’s claims”.

If those links are on very bad sites with a notorious spam reputation and have 100% identical anchor text, then the odds of a bad reaction go up. If those links are on a private network of selectively chosen domains that you bought after researching to only pick the ones with pre-exisiting clean link juice, then the odds of a good reaction increase.

How did you determine a site was bad (according to Google) e.g. notorious spam reputation. Just eyeballing some links in your GSC account?

How did you determine said websites were from the same network (according to Google)?

If a domain is manually review it doesn’t pass that manual review onto outbound links automatically nor automatically transfer the manual action to another domain.

If a domain is PENGUINized for specific links (anchors) that may indeed impact on a site move e.g. Like move the website to a different domain… But it has no impact on a 301 redirect to a completely different set of topic in another industry.

If we are going to talk about unnatural vs natural link profiles, you don't want 100% identical anchor text. Some "irrelevant anchors" are nothing to worry about. Most sites have many links with the anchor text of "click here" or "visit website".

No one implied identical, I said 100% irrelevant.

If you only have 50-100 backlinks and you gain 100,000 links overnight, that will look very unnatural regardless of the anchor text.

We are certainly investing in discussion creep.


Googlebot needs to crawl all 100,000 overnight which in nearly impossible to have for most websites. That said, what if you had a 1000 to this page, 10,000 to another, 5000 to that page some over here, some more over there.

I don't care about quantity of links, quality trumps quantity. I would rather have a domain with a couple rel=”nofollow” wikipedia links, a gov or edu link that suggests a domain with a quality history and a few more. So how suspicious is that?

What is quality? In my world that is still PageRank… PageRank alone doesn’t afford ranks but pass that PageRank through a blog post, a blog category and then to the page you desire to rank. After that spend $$$$ on linkbait.

[fathom]

The videos have clues but Matt Cutts comments to MOZ.com

Says it all

[m.youtube.com...]
[m.youtube.com...]
[m.youtube.com...]
[moz.com...]

EXCERPT FROM THE BOTTOM OF THE ARTICLE.

Several months ago, Jen Lopez, Moz's director of community, had an email conversation with Google's Head of Webspam, Matt Cutts. Matt granted us permission to publish portions of that discussion, which you can see below:

Jen Lopez: Hey Matt,

I made the mistake of emailing you while you weren't answering outside emails for 30 days. :D I wanted to bring this up again though because we have a question going on in Q&A right now about the topic. People are worried that they can't guest post on Moz: [moz.com...] because they'll get penalized. I was curious if you'd like to jump in and respond? Or give your thoughts on the topic?

Matt Cutts: Hey, the short answer is that if a site A links to spammy sites, that can affect site A's reputation. That shouldn't be a shock--I think we've talked about the hazards of linking to bad neighborhoods for a decade or so.

That said, with the specific instance of Moz.com, for the most part it's an example of a site that does good due diligence, so on average Moz.com is linking to non-problematic sites. If Moz were to lower its quality standards then that could eventually affect Moz's reputation.

The factors that make things safer are the commonsense things you'd expect, e.g. adding a nofollow will eliminate the linking issue completely. Short of that, keyword rich anchortext is higher risk than navigational anchortext like a person or site's name, and so on."

That's to a highly respected and quite authoriative SEO site.

There is tons of evidence out there... And while the source of links should be questioned the anchor phrase is what gets devalued and if you are not targeting it ... NO NEED TO WORRY!

[fathom]

Your latter question about my logic on disavowing a natural link.

Google sees B&W there in no grey area.

In that context a link is either UNNATURAL or it must be seen as NATURAL to Google.

A link either passes PageRank or it doesn't.

That's my logic... So when you error in disavowing you are negatively impacting your domain.

[goodroi]

@fathom I appreciate your enthusiasm for posting but I have a hard time trying to follow your thoughts which seem to be rambling. Your thoughts still do not make 100% sense to me. Some of your ideas I agree with but some other ideas of yours just do not match up with my experiences or research. I guess we will agree to disagree.

[Walt Hartwell]

@fathom
you state:

Disavowing forces Google to drop that link from the link graph regardless of any impact (both positive or negative). If said link doesn't use your keywords there can be no negative impact because there was no unnatural reference to induce a positive impact that could never be misunderstood as manipulation.

What can you possibly not understand about this specific link spam being anchor text related? I've seen multiple instances of it, it is certainly topically related, yet you blather on about it cannot have negative impact. Disavow in effect places a "nofollow" on those inbound links, but I don't believe those links disappear.

I've gotten to the point where I'm believing the "fathom" account is now controlled by a group of people with limited English skills and a propensity for selling mentoring. It wasn't that way before, but now I literally cannot follow what is trying to be conveyed by that account.

Feel free to send more abusive PMs, it doesn't bother me.

[FranticFish]

If said link doesn't use your keywords there can be no negative impact

That assumes the non-existence of 'topical PageRank', which - whilst not a certainty - is a probability.

Alluded to in Matt Cutts video - [youtube.com...]
Discussion by Bill Slawski including paper on the subject from a Google search engineer - [seobythesea.com...]

So there's at least a chance that by not disavowing a massive block of links from a completely irrelevant site that you are harming your own relevancy score.

[edited by: FranticFish at 8:00 am (utc) on Nov 5, 2015]

[tangor]

I take a different approach.

I did not create the link. I don't care if it is follow or no follow. Most spammy links don't hit my site anyway (with traffic). I didn't seek it or pay for it. It is noise in the wind as far as I am concerned.

I will not do g's job of knowing (as they profess) which are good links or bad links, thus have never used the disavow tool(s). In fact, I see that as "admitting to a wrong I never did" and in that regard take comfort of the Fifth (American slang for those not in the USA which says "I ain't talkin' and you can't make me".)

Can there be a deleterious effect? I suppose so. I haven't seen it, but many others have claimed such... and most who do created the problem in the first place.

Until g PAYS me to clean up their link catalog there's no way I will volunteer to do so.

Just one man's opinion on this topic.

[nomis5]

I absolutely agree with Tangor above - 100%.

I feel strongly that all this "disavow" stuff makes not one jot of difference for us. Any disavows are accumulated by G to get an overall impression of which sites they need to review. They are in effect asking us to do their job for them. And from their perspective, why not if it saves them work.

[AnthonysItalianFood]

Thanks guys, this is helping me with my original question I posted here: [webmasterworld.com...]

Checking out the vids now.

[fathom]

My thoughts are rather random trying to cover all the random angles put into a post.


Not hard to fathom…takes your lead from tangor… Why does he avoid PENGUIN by not DISAVOWING? Is he just lucky, or is he acting UNSAFE? Or maybe there is some truth to it.

All links start out being natural links until PENGUIN determines there is a pattern of manipulate. One link from one domain even if that domain has a million pages and the single link is on a file include, is a pattern of one. A pattern of one is too granular for PENGUIN. If you are not manipulating anything then there is nothing to disavow.

You can certainly nofollow (disavow) if you are worried about worrisome links but the only links you need to worry about are the ones that induced ranks directly via anchor text manipulation. Unrelated anchor text (even if a keyword or phrase is useful for another domain but not something your DOMAIN is trying to rank for) isn’t manipulation to your domain because it is of no ranking value to your domain so what plausible harm can it do? If there is no intent to manipulate, guessing a link you don’t recognize has direct manipulation intention to devalue you is just bad research.

Please show where Google has stated that, provided a video on that.

My question … Show me your sources… Or are you simply guessing?

PageRank is indeed useful. But ONLY PAGERANK DOES NOT produce ranks. Google handed everyone a gift with PENGUIN.

All links without the nofollow micro-expression pass PageRank. That means all links start off being NATURAL. They can only become UNNATURAL with a ranking impact being detected.

[fathom]

That assumes the non-existence of 'topical PageRank', which - whilst not a certainty - is a probability.

Alluded to in Matt Cutts video - [youtube.com...]
Discussion by Bill Slawski including paper on the subject from a Google search engineer - [seobythesea.com...]

So there's at least a chance that by not disavowing a massive block of links from a completely irrelevant site that you are harming your own relevancy score.

I can't rule out Topical PageRank (if it exists) to produce better ranks but I can certainly rule out Topical PageRank as a devaluation method today... The latter is being discussed here.

That could be a PENGUIN 5.0 or 6.0 or later.

[aristotle]

Most sites get these types of backlinks. They're just part of the worthless garbage, much of it auto-generated, that's created in abundance everyday. Surely google's algorithm can recognize it for what it is and totally disregard it, and so disavowing it almost certainly has no effect at all on your google rankings. I've never disavowed any of it and my sites are doing just fine. So the best thing to do is ignore it.

[FranticFish]

I can certainly rule out Topical PageRank as a devaluation method

I think this is splitting hairs and I don't think you can do this. In this particular instance, does it matter if something is a valuation or a devaluation method?

A valuation method that looks for a clean signal will return a higher score for a cleaner signal, and a lower score for a less clean signal. One calculation on overall value.
OR
A valuation method looks at all the clean signals and scores a positive subtotal on that, only to be adjusted by the negative scores from the devaluation method to arrive at the total.

These two processes should produce exactly the same result.

I feel strongly that all this "disavow" stuff makes not one jot of difference for us

I've tried very hard to find a reliable record of an algorithmic penalty being reversed via disavow and have found it very difficult. However (and this is important) this could be is due to the fact that algorithmic penalties under Penguin are only reversed at the next update, and most people don't wait that long to study one factor in isolation.

Dixon Jones, manual penalty reversed: [seroundtable.com...]
Manual penalty reversed: [webmasterworld.com...] & [webmasterworld.com...]
Article linked to from the thread above (manual penalty) [searchenginejournal.com...]

But... now, read this: [webmasterworld.com...]
Someone does only a link cleanup & disavow (which we all agree is the same thing) and sees algorithmic recovery.

[fathom]

That's a pattern of at least three. Actually I errored. That's a pattern of four... Missed Nomis5.

This public exposure was my second PBN lost at about $250,000 in lost domain assets [productforums.google.com...] it occurred 2 years before PENGUIN 1.0 launched.

I lost two more in 2011 at almost $500,000 each a year before PENGUIN 1.0 which offered massive insights and by the time Google was ready with PENGUIN I didn't need PBNs, but PENGUIN still has a tough time auto-detecting them.

The pre-launch of all three PENGUIN upgrades took manual actions on PBNs. The cause & effect ... Webspam got more sophisticated, then so must PENGUIN, then so must webspam, then so must PENGUIN, then so must webspam, etc.

IMHO most claims of NEGATIVE SEO campaigns can be chalked up to a novice not actually appreciating what is really going on. A GENUINE NEGATIVE SEO CAMPAIGN must first induce ranks. When you hear anyone claiming NEG SEO that's the first question to ask.

Who builted the ranks?

Answer: those just happened over the last few months. If not, not a NEG SEO CAMPAIGN.

Disavowing is a waste of time, lost PageRank, lost ranks especially if your living isn't made from devaluation recovery.

[edited by: fathom at 8:59 pm (utc) on Nov 5, 2015]

[fathom]

@FranticFish never claimed you can't reverse a PENGUIN devaluation by only disavowing...

The problem is identifying the links that caused the devaluation. The PENGUIN GIFT I mentioned however points to a truth. How precisely do you see a recovery from less links with less PageRank? If you can work out a plausible answer for that, you have also solved this mystery.

If you are attempting to rank for SEARCH ENGINE OPTIMIZATION and you have external link anchors with exact matches of SEARCH ENGINE OPTIMIZATION I would indeed disavow that... Although if you actually created it in some way better to get the rel="nofollow" micro-expression in the link element instead.

If you never created it... best to communicate with the owner to have the rel="nofollow" micro-expression in the link element instead.

Partial matches may be less risky but they are still risky!

In GSC in the tab LINKS TO YOUR WEBSITE below the two columns of link references is a tabular list of the 200 most popular ways websites link to yours. If any of your KEYPHRASES or KEYWORDS appear near the top you are in risk of being devalued. If ONLY brand oriented or call-to-action refererences only appear here... You have nothing to worry about.

Also, in my experience, if you have 200 totally irrelevant expressions you have nothing to worry about either.

[Walt Hartwell]

If you are attempting to rank for SEARCH ENGINE OPTIMIZATION and you have external link anchors with exact matches of SEARCH ENGINE OPTIMIZATION I would indeed disavow that..

I think it's a good thing you've perhaps started to understand what people are experiencing. A little discouraging that you throw negative SEO issues into the thread, none of it has anything to do with negative SEO.