Link Maintenance Matters

[goodroi]

Someone recently asked me why one website performed so much better than another website. It was super easy to answer - backlinks. One site had backlinks from 19,000 domains and the other site had over 400,000 linking domains. To make matters even more impressive the larger backlink portfolio also had a much better quality aka trustworthy links.

As much as we talk about Google using over 100 ranking factors, we need to remember that Google places vastly different value on each ranking factor. Google has admitted that BACKLINKS ARE ONE OF THE MOST HEAVY RANKING INFLUENCERS. Don't believe me? Go back and watch the Google webmaster videos, interviews with Google engineers and their search conference speeches. Still don't believe me? Do your own research and you will find a very strong correlation between and backlinks & rankings.

So what have you done to maintain your links?

Are you regularly checking your backlink profile to spot and regain backlinks that former link partners deleted? Are you crawling your site for broken internal backlinks? Are you regularly reviewing your competition backlinks to find new link partners? Are you monitoring SERPs to identify and contact new link partners before your competition? Maintaining a strong backlink profile is critical to maintaining your Google rankings. Even if you don't care about Google rankings, you should care about growing referral traffic to your website so your business isn't 100% reliant on Google sending organic traffic. Make sure you invest in your backlinks!

*Disclaimer - I am not saying you should ignore content and other important pieces. Great content makes it much easier to convince websites to link to your useful website. Internet marketing often has best results with a holistic approach - so stop being lazy with your backlinks :)

[aristotle]

the other site had over 400,000 linking domains

Apparently these are "naturally-attracted" backlinks, not artifically-built or generated by some link-spam program.

But before a site can attract such a large number of natural backlinks, it has to have a lot of traffic. But in order to get high rankings (and high traffic) from google, it needs to have a lot of backlinks. So which came first -- the backlinks or the high rankings in google?

[goodroi]

I'm not saying 400,000 link domains is normal but when you are one of the most popular sites in the US, it is possible.

High rankings come first.

I've tested this out on a few new sites over the last 24 months. If a new site has strong content and social/usage signals you can have high rankings with very minimal backlinks. If the backlinks do not quickly catch up with your other rankings signals you will lose those short lived rankings. I am not saying a new site is going to magically rank #1 for the most popular term. A new site with minimal backlinks can successfully launch and if its something really unique & useful it will gain backlinks and the rankings will keep improving and spreading out to more keywords.

[jmccormac]

Most webmasters seem to have no idea of the fragility of the link graph of their sites. Every month approximately 2.3 million .com domain names are deleted and with them a considerable number of links. Far from being an unchanging monolith of a TLD, the .COM is continually changing with domains being deleted and new ones being registered each day. Verisign commented in a recent press conference that the one year renewal rates (renewal rates on new domain names registered for a year) was about 50%. That means that of the .COM domain names registered in August 2015, approximately 50% will be deleted in 2016. (This had to do with questions about the Chinese speculative bubble in .COM/NET in Q3/Q4 2015.) The one year renewal rate has been gradually declining for some of the gTLDs.

In engineering, there is a concept known as the "bathtub curve". It deals with the failure rate of components. ( [en.wikipedia.org...] ). The failure rate spikes at the beginning of a component's lifecycle and then towards the end of the lifecycle. The lifecycle of a link may follow a similar pattern with the exclusion of site redesigns. One would have to know the creation date of the domain name and correlate that with the link data. If a domain name renews for the second year, then the link may be stable for a few years. The renewal rates vary by country with some countries having very low renewal rates and others having high (>70%) rates. This is a kind of second order linking effect where links from particular countries may be quite sticky and others may be, in annual terms, transient links.

If your site has links from some of the new gTLDs that have used freebie or discounting to build registration volume then you may want to evaluate the security of your website as links from sites in these gTLDs particulary occur with link injection compromises of websites. As for Google, it and its algorithms are of secondary concern in this case.

Regards...jmcc

[aristotle]

jmccormac -- The backlinks that are short-lived are nearly always low quality backlinks.

High-quality backlinks from high-quality websites are usually stable and long-lived.

So most of the backlinks that a site loses are low-quality, and therefore the loss doesn't have much effect. It's mainly the high-quality backlinks that determine what happens over the long term.

[JS_Harris]

I'm not surprised. Backlinks will continue to matter a great deal to Google, at least until they let AI serve up all data and make websites obsolete.

I just wanted to add that before you embark on a backlink generating campaign it's probably a good idea to refresh your own internal link structure first. If your best pages can only be reached from page 52 of a category it's not benefiting from incoming links to other pages nearly as much as it could. When your best/money content sits in the most prominent locations within your internal link structure they benefit from incoming links to virtually any page, by virtue of being 1-2 clicks away, while the buried pages need direct links to them to achieve the same.

[aristotle]

It's best for a site to have a lot of different pages that can attract backlinks. If nearly all of the backlinks point to the same page, such as the home page, that's not as good.

On my sites each separate article can potentially attract its own backlinks. It happens that some of my articles attract a lot more backlinks than others do, and some don't attract very many at all. But if all of the articles relate to the same general subject, then they tend to support each other and the whole site is stronger overall.

[engine]

Unless I get a warning of a problem, I usually assess each site on a schedule throughout the year. That way i'll eventually get to them all.

Key aspects for me are to ensure the inbound points to the correct resource, and the resource is up-to-date and accurate. Oh, and, of course, not broken.

Ideally, i'm looking for some organic link growth without intervention, and that's all it needs. A quality resource usually generates its own links when released, but tails off fairly rapidly, then may only achieve very slow growth, and eventually achieves little or no growth. That timescale varies considerably, and in these days of topics going viral, might only last a couple of weeks, ending up being ephemeral.

An important factor for me is to check the inbound has not gone bad, and it occasionally happens as domains get sold, or even hijacked.
Trust really is important.

[aristotle]

An important factor for me is to check the inbound has not gone bad, and it occasionally happens as domains get sold, or even hijacked.

By "gone bad", don't you mean "disappeared"? If the other site is sold or the domain allowed to expire, there's not much you can do.

But like I said earlier, backlinks that disappear are usually low-quality backlinks anyway.

[engine]

Gone bad can mean a number of things, and if you've ever found a hijacked site linking to you, it's not good. We must have all seen sites that still link, but have now changed their topic. I've seen some nasty hijacks in the past, and sometimes it's not pleasant at all. Sometimes the site owner doesn't realise there's an abuse going on.

Yes, low quality links will eventually go.

[aristotle]

What about people who buy expired domains for sites in their niche and then re-direct these previously-expired domains to their own site? Thus any backlinks that these domains still have pointing to them will be re-directed to their own site. These re-directed backlinks are not "earned" in the same way as naturally-attracted backlinks. So does google count them or not?

[ergophobe]

One minor technical issue to point out to folks who are doing an annual look at their links.

One tool you're likely to use is looking at your referrer traffic, right? Which sites, aside from SEO factors, are sending you the most traffic. Have some sites just dropped off your referrer list? Big contributors who are completely gone?

Well, if they have gone https and you are still on http, you will lose the referrer data. It no longer gets passed and will not show up in your analytics.

[aristotle]

Well, if they have gone https and you are still on http, you will lose the referrer data. It no longer gets passed and will not show up in your analytics.

I don't understand why this would be the case. Is it a bug in Analytics or does the visitor's browser simply not provide the referer info to the server. And if so, why not? I don't understand it.

[Walt Hartwell]

What about people who buy expired domains for sites in their niche and then re-direct these previously-expired domains to their own site?

You mention earlier in the thread that having links to many pages on your site is better than links just to the main page. Redirecting a dropped domain to a new domain won't, by itself, carry any potential link benefit from individual sub-pages that used to exist on the dropped domain.

[aristotle]

Walt Hartwell -- Well individual pages on one site can be redirected to individual pages on another site if the content is related.

But I didn't go into the details because this isn't something that i've ever done or intend to do, or recommend that anyone else do. I brought it up because some people have done it and I was wondering if google gives credit for for that type of backlink since it isn't actually "earned'.

[ergophobe]

I don't understand why this would be the case. Is it a bug in Analytics or does the visitor's browser simply not provide the referer info to the server. And if so, why not? I don't understand it.

Passing referrer data from a secure page to an insecure page is prohibited by the HTTP spec. Only old or non-compliant user agents will do so.

A user agent MUST NOT send a Referer header field in an
unsecured HTTP request if the referring page was received with a
secure protocol.

[tools.ietf.org...]

[Walt Hartwell]

I was attempting to communicate that there are many nuances with a redirect that eliminate or severely reduce the perceived benefit of links pointing to an expired/dropped domain.

Can redirecting a dropped domain provide benefit? Yes. Will most people get no benefit from it? Also yes.

As a fairly lame example, I got a blog post on 1st page SERPs for a minor term with nothing but Chinese links. There was one English word in the domain which was part of my target. Why? Because I wanted to know the effect of links having nothing in common with the target page content. It wasn't a churn and burn project, but pretty close in concept.

[ergophobe]

And section 9.4 on security

Since the Referer header field tells a target site about the context
that resulted in a request, it has the potential to reveal
information about the user's immediate browsing history and any
personal information that might be found in the referring resource's
URI.

[aristotle]

Walt Hartwell -- I was trying to answer your point about re-directing everything to the home page of your site. I agree that this probably doesn't have much SEO benefit.

ergophobe -- My sites are http, but the server logs for all of them show numerous referrals from https://www.google.com/, https://www.bing.com/, and other https sites. So I'm still confused.

[smallcompany]

So which came first -- the backlinks or the high rankings in google?

Good/useful/original/real life content > shared among people and sites > more real backlinks > higher rankings (I believe)

[iamlost]

Link maintenance is serious business.

In my view there are three major categories of links:
* those within a site;
* those pointing to a site aka back links;
* those pointing away from a site aka external links.
If one has link rot and navigation problems within their site... RIP...
So I'll limit comment to the others.

Many/most webdevs that concern themselves with back links seem to view them through Google coloured glasses; they peruse various back link services and peer over the results for various perceived SEO reasons. Barring the one time I was attacked by negative link building I could care less about any links that don't show in my log files; if a link doesn't refer actual human traffic I don't have the time to waste on it.

In links:
On the other hand, if a new to me link sends traffic I (via bot) check it out; loading URL, domain name, title, meta description, all headers, anchor text, surrounding text into a database. Depending on the traffic level I subsequently check back regularly, at least annually, to confirm status/change.

Why? Because I view those referring visitors to my sites as affiliates: a mindset not a paid relationship. :) I don't much care about what my competitors are doing, I most certainly do care about my referrers.
* what are they saying about my niche and about my sites that brings their visitors to me?
* is the link pointing to the best resource page or page location?
* what is their visitor market segment breakout?
* what is their perceived personal/niche qualifications, authority, trust et al.
* how are their referrals converting?
* etc.
Most webdevs parse Google and Google traffic - I parse just about everyone but Google. :) Google comes along free with everything pretty much regardless, no need to get fixated.
Note: if one has to constantly chase Google there are serious fundamental site/business problems that need attention.

Out links:
Are my recommendations. I am saying 'if you like my stuff you'll enjoy this as well'. My reputation is riding on each and every one. So link rot is a serious matter. If gone the link must be removed asap and a replacement discovered if appropriate (and surrounding, anchor text adjusted as necessary). If resource has changed (1) for the worse, the link must also be immediately stayed and re-pointed or replaced , (2) for the same or better the anchor and surrounding text may need adjustment.

The easy if irritating side is removal/replacement. However, adjustment may be almost as important so that leaving my site and arriving on the other is a smooth transition. A bumpy landing isn't professional. I want return visitors. And that means being the best in niche from arrival to after leaving.

Obviously, I have gathered the same information on the sites to which I refer as those who refer to me. And check even more frequently.
Note: without that site/page information how would I know about change? Given the shear number of links they must be tested automatically, not enough time for manual check of any not flagged; barely enough for those.

Links are the web. They are absolutely business critical.
Too bad so few treat them as such.

[ergophobe]

server logs for all of them show numerous referrals from [google.com...]

As I mentioned, that's common. What are the User Agents? Generally, it should be non-compliant UAs, but people do report a fair number of referrals from https sites.

I can't post images here, but in another place I just posted an image that shows referral traffic from a site as it changed to https. It was sending 4,000 referals per month and it went to about 10 overnight and then dropped to zero.

[Walt Hartwell]

Note: if one has to constantly chase Google there are serious fundamental site/business problems that need attention.

I've heard that a few times before. Why would anyone chase an algorithm? In the first part of the day, today, I established a new site with no links and three pages of solid information. By the end of the week I'd guess it will have some very viable links to the site.

This isn't 1996 or 2004, it's a good idea to adapt to the current environment.

Second part of the day I hacked the hell out of some hedges. Not that it's relevant, it's just how I stay in some kind of physical shape.

[aristotle]

ergophobe -- So as more and more browsers become compliant, I can expect to see fewer and fewer explicit referals from google and other https sites in my logs, and more and more entries with "no referer".

I don't have time to do a detailed analysis, but my genertal impression from looking at my logs is that most current browsers still aren't compliant.

[iamlost]

Just a side note on the switch from HTTP to HTTPS and the associated loss of HTTPS referrer data to HTTP sites:

Currently, globally, over 70% of browsers fully support HTTP/2 and another 5% plus are partially supported. However, while HTTP/2 spec allows for both secure and non-secure connections, the browsers ONLY support it over TLS (aka HTTPS).

As hosts offer the service (most cloud providers already do) and webdevs are shown network time savings and rendering results especially for mobile I expect the next couple of years switch to HTTPS::HTTP/2 to be similar to the mobile takeup, starting slow but building to a crescendo surprisingly quickly.

Indeed, the only obvious drag on the switch that I see is Google and non-compliant (aka dropped from auction) HTTP ads. As many in the AdSense forum have noted HTTPS ad fill is a problem. Once Google requires all ads on their networks to be HTTPS compliant...

So, even sites that really don't have a need for HTTPS will be pushed in that direction. Ready or not.

[Nutterum]

Very interesting discussion about the nature and importance of links. What I wanted to steer the conversation to is the more and more popular "paid" do-follow links. Even websites like Huffington and Mashable offer quite a few juicy "packages" where they can write one or more articles and then provide do-follow links to your website. These articles then get scrapped and you get even more links and referrals. Even on ccTLD level there are many news outlets that offer similar "services". Despite Google warning about such types of link building, they have little to no way of finding out, nor they do anything about it. So what is your standing on a "paid" way of acquiring links from trust heavy domains?

[Walt Hartwell]

@Nutterum
"Despite Google warning about such types of link building, they have little to no way of finding out, nor they do anything about it"

What was your question?

You probably don't want to build webspam, but arbitrary guidelines that may, or may not be relevant to your projects aren't something you should worry about.

[aristotle]

Nutterum -- You described one method that people use to cheat their way to higher rankings and traffic than their site deserves. That's not the only method that people use. Google isn't that hard to trick, especially if you've got plenty of money to spend.

[martinibuster]

So what is your standing on a "paid" way of acquiring links from trust heavy domains?

This is 2016, not 2006.

If you understand how links are rated then you'll understand that those kinds of general news type of links have limited to zero ranking value. Many popular link schemes are hamster wheels.

Imo, getting your site into the news (in a non-paid way) can be a path to ranking better but it's not a path for link popularity in itself.

[aristotle]

Occasionally I get an email from someone who wants to buy a backlink from one of my sites. I've never done it, but if I did, and was prudent about where and how I placed the backlink, then I'm almost positive that google's algorithm wouldn't be able to detect that it was a paid link.