20 Attributes that Influence a Link's Value

When it comes to the link building process, there’s a lot to keep track of — and that process today is pretty different from what it was a few short years ago. In this week's Whiteboard Friday, Rand scores 20 different attributes that can influence a link's value based on whether or not they still matter in 2015/2016.

Click on the whiteboard image above to open a high resolution version in a new tab!

Video Transcription

Howdy, Moz fans, and welcome to another edition of Whiteboard Friday. This week we're chatting about the attributes that influence a link's value and which ones we might still need to care about.

So in old-school SEO this was critically important. We had this very manual link building process where we'd go, "Well, here's a page. Do I want to get a link from that page? I know I'm going to have to put a bunch of outreach effort in, and that could be everything from submit to this page to email or pitch them or do business development and get a partnership going, maybe some press and PR. Maybe I can submit some guest content, get an embed going, try to get on get on a different page on the same site or some new angle. Maybe I can go buy the domain of the person they're already linking to and redirect that."

Anyway, very manual link building process, which meant every link had to be considered from a bunch of different attribute-based angles. I think that that was a little bit too manual. Probably a lot of those things that we were concerned about back in those days didn't actually make that big of a difference. Or if they did, it was only for a little bit of time until Google caught up with their algorithm and got better at it. Certainly, look, because the power of links used to be kind of the whole algorithm and today it's a much more complex and nuanced picture, this stuff really mattered.

But as we moved into this more "link-by-exposure" world, where you kind of fire the cannon, usually through social media and all sorts of other broadcast mediums, to your page and then hope that everybody goes, "Ooh. Aah, I'm going to link to it. Yeah, that's great," that's changed a little bit.

We've gotten less concerned with: "Who are these people, and do I have to worry about all the various attributes that they have or not? Do I have to check all the boxes?" No, not really. That world has gone away in a lot of senses.

However, it is true that today Google and Bing still look at a bunch of different attributes about a link to determine the relative value that it might have. I think there's a bucket of about 20. There are probably plenty more in the minutia and probably a couple of big ones that I've missed here. But these 20 are a good representative sample of kind of the things that most people in the SEO world think about or did think about.

So what I wanted to do is run through these attributes, talk about how and whether they were still important in 2015/2016 and moving forward, and how we might measure those and figure out whether those things are worthy of going out and getting links for, especially as some of us, I think some successful SEOs are actually returning to or never quit the model of having a manual outreach process for link acquisition.

We've seen that some old-school SEO folks have been really successful with this, and certainly in our studies with BuzzSumo, we saw that a lot of links that people tried to acquire through social or through broadcast means don't actually work all that well. So as we think about this again, which ones of these do we need to worry about? Let's go through here.

1. Anchor text

These are absolutely still important. I gave it a three out of five, five being the most important, zero being not important at all. The reason I only gave it a three is because in a lot of senses, while an anchor text does move the needle further than an exact match anchor text does, more than partial, and partial does more than non-match, a link is still valuable regardless of whether you get the anchor text. So if the choice is between, well, I could get no link at all, or I could get a link but it wouldn't have the anchor text that I want, get the link. Get the link.

With anchor text, obviously you can eyeball it, but you can use any of the link-based tools — Open Site Explorer. You can see this in the MozBar as well as Ahrefs or Majestic for seeing distribution of anchor text to a given page if that's what you're looking at.

2. PageRank

I gave that actually a very low score, because look Google's PageRank in the toolbar — obviously dead and been dead for years now. You can use MozRank as a substitute, and it's calculated in, well, I would say nearly the same way as the original PageRank patent paper. Who knows whether that is how Google is calculating PageRank, and I'm sure they have many other link graph-based metrics that they're doing these days.

But PageRank of the linking source, not that important for a bunch of reasons, including that if you're getting a link from a page that is not yet published or will be published soon or has been published in the recent past, it may not have a high MozRank or high PageRank score, and that doesn't matter very much. I'd much rather that we look at domain authority and importance of the site overall and importance of other pages on the site, relevance, etc.

3. Relevance

I gave relevance itself a three. Look, relevance is important, and you have to eyeball it. There's no tool that tells you exactly how relevant a page or site is going to be to your precise concept, at least not right now.

But to this degree I don't think it is critically important that you say, "Hey, I am writing about gopher food, and I need to go find websites that are only about food for rodents and marsupials, because any other site is just not going to be very relevant to my site." That's not true.

You can go get a broad new source. You could even go get tech industry news, which is completely unrelated. You could get links from a fashion site or from non-related sites that are in your geographical area. Those would still be providing a lot of value.

So relevance is something to think about. It's certainly a way to find folks who might have an affinity for linking to you, but it is not a requirement, and I would urge you not to cut out people who are irrelevant.

4. Domain authority

I'm not talking here about Moz's Domain Authority, although that's a reasonably good way to measure it, but rather I'm talking about the concept of domain authority. How important is this domain from a content and relevance independence perspective, just from the links and how Google considers that site?

This is pretty important. I think this is the number one way that I would go about, from a metrics-driven standpoint, trying to judge whether a link is more valuable than another and whether I should go pursue it first. I want to go after the highest domain authority targets that I can, as long as I think they're equally easy to get, and work my way down in terms of authority of the site. That's just going to be one of the best predictors of how a link will perform for you. It's not just from a link perspective, like link that will increase my ranking. It's also a link that will be seen by people, be clicked, be engaged upon, have these amplifying effects. So lots of good reasons to think about domain authority.

5. Location on page

Look, you can eyeball it, but that stuff barely matters anymore. The only reason I'm giving it a two instead of a zero is because if a link is in the footer of a page, in the sidebar or the advertising section of a page, it's likely to be seen and potentially classified by Google as being an advertising link. If it's not no-followed and it looks like it probably is advertising or sketchy somehow, that could get you in trouble. So something to consider, at least a tiny bit.

6. Internal versus external

That one's obvious. It is important, but we don't need to worry about it really. Basically, an external link, a link from a site that is not your own site will help you much more than one that's on your own site.

7. Quality of the page's other links

This is actually pretty high. Unfortunately, there's no great way to do this on a page-by-page basis. You could look at Moz's Spam Score on a domain or other metrics like that. You're going to have to eyeball this one and say, "Hey, are they linking to really shady places? Is this an open, free-for-all submission page that anyone can submit to? If so, maybe I shouldn't go after it. Maybe it's not all worthwhile." If, however, the links on the page and on the site generally are well curated, you're good.

8. Editorial integrity

This is a total judgment call, and we could quibble about which news venues, which websites we think have true editorial integrity. All I would say is, be aware that it's probably in the consideration set of things that Google will try and look at via machine learning or deep learning by analyzing content and seeing how people have regarded that content and judged it. So just know that it's coming.

9. User engagement

This one is really important. I almost want to put this at a five. It is tough to measure though. The reason user engagement matters so much is because if a page is visited a lot and people engage with that page and they click on those links, Google is seeing that through things like Chrome, through Android, and through all the Wi-Fi services, etc, and ISP data if they're buying that.

You can use something like SimilarWeb on the domain to see how relatively engaged with a particular domain is, or you can use your own web analytics, Google Analytics or Omniture or what have you, to say, "Is this particular link from this page sending me traffic?"

That is a good indication that it's getting engagement, and you probably want to go find links like that. But it's hard to do before you get the link. However, if you can use your best judgment and figure out whether you think a page is going to be engaged with or if it is engaged with — for example if it's ranking well or if it gets socially shared a lot — those are good signs, and user engagement we've seen to be highly predictive of rankings as well as rank influence.

10. Follow versus no-follow

Yes, it still matters. Every test we've run, a no-follow link is not treated like a followed link. I know they're well correlated, but remember that followed and no-followed links themselves are well correlated, so that could explain it. You can use the MozBar and turn on the highlighting for no-followed links, or you can eyeball it if you want.

11. Source depth

This was something that back in the day when Google's crawlers were not as robust as they are today — and actually Bing's crawlers are fairly robust today too — you could see that pages that were basically higher up in a site structure, fewer links away from the homepage, more likely to get crawled and crawled more frequently looked like they did well. Whereas ones further down not so much. These days it barely matters. I wouldn't worry about it.

Google is crawling everything. They judge all, see all, know all. ScreamingFrog, you could look at this, and I think they will calculate depth away from a homepage or a category page for you.

12. Text versus image

Don't care about it and it's obvious.

I also don't care about this one. With both of these, yes, there were tests back in the day that showed that, yeah look, an image link passes slightly less value through its anchor text, which is the alt tag than a raw text link does. But it's so infinitesimally minute.

If you can get a link, get the link. If you think the image is going to get more clicks than the text, go for it. If you think the text will get more links because people recognize it as a link, unlike the image, go for it.

13. Link age and page age

You can use the Wayback Machine if you want to see how long a page has been around or how long a link has been on that page. But the whole old, crusty links thing doesn't look like it has good correlation or good connection to rankings today. In fact, the opposite is true. We've seen where new pages pop up that have links on them and those will actually outperform old crusty links.

14. Topical authority of a source

This is basically saying not how authoritative is this website overall from a domain authority perspective, but how authoritative is it in this topical area. A great way to measure this is to look at either Google, you can run some searches around a topic and if you consistently see a site performing well in the top 10, that might be a signal. Then you can use SEMrush and see all the keywords that site ranks for and how it performs and how it's performing over time.

That's a great indication of potential topic authority in Google's eyes, and it is moderately important. If you do find that great gopher food website that ranks for all the terms and they link to your gopher food site, probably a good thing.

15. JavaScript vs HTML

The only reason I didn't give this a zero is because there are still some forms of JavaScript rendering and weird stuff that people can do on pages that Google doesn't crawl as perfectly as they say they crawl it. You can look at Google's cached snapshot for any given URL to see whether the links are getting picked up or if it's post loading JavaScript and they are crawling and indexing it. Try to go for stuff that isn't too fancy. I wouldn't urge you to get too complex with your AJAX and all that kind of stuff.

16. Spam signals

I gave this one a four. It is relatively important. Google will discount or even penalize sites based on lots of spammy links. I like to try and keep the MozBar open while I'm surfing and have Spam Score in there and just have a sense of like, "Hey, does this domain look sketchy or not?" Or you don't even need Spam Score. You can use your own judgment. Ninety-nine percent of the time when you look at a site, you get a good sense of like, "Hmm, do they look suspicious? Yeah, maybe they do."

17. Speed of link acquisition

If you are doing white hat, non-manipulative, non-spammy linking, you never need to worry about this. Speed up, slow down, hey, that's just how it goes. Fine, don't worry about it. The only time you ever worry about speed of link acquisition is if you are doing something sketchy and you think you can maybe do it smarter than Google can catch you doing it, which good luck. Majestic is a way to go look at people like that, and you can see link growth over time and see if someone has gotten a bunch of sketchy-looking links — or a bunch of good-looking links, which could be the case, too.

18. Author authority

Google has sort of wavered on are they're completely deprecating it, or is it still a thing at all? So maybe it's still a little bit of a thing. I think where it really becomes valuable for us is not as link builders from an SEO perspective, but as link builders from an authority and influence perspective. If you can get an authoritative author in a space to write about you and link to you, that's a very good sign. You can use the quotes that that person has said about you in your marketing. You can use their endorsement as you pitch people. It's helpful in all sorts of ways.

You can measure that still via Google+ or use your own judgment. Google+ meaning for whatever Google might be doing with author authority, you could take a look at that.

19. First link to target in HTML

I don't know if you remember these tests, but we used to look at a page and we'd say, "Oh, if these two links both point to the same page, it looks like Google will only consider this first one. The first link is the one that passes anchor text and ranking signals and all that, and they sort of ignore the second one."

Does this matter all that much? No, it doesn't matter. The page linked to you twice. Great. More of a chance to get more user engagement, more of a chance to get more clicks, more traffic, and more exposure. The fact that the first link doesn't have your ideal anchor text, that would be sketchy to convince me that that is the thing that you should do with your outreach time, but maybe it is. One.

20. Prior links from this domain

Again, this is an issue where people said, "Hey, I'm looking for a lot of domain diversity and link diversity specifically." We had an old Whiteboard Friday that this one is now replacing about link diversity. I think we lost the video in the archive somewhere. But link diversity used to matter a tremendous amount back in the old days of SEO, kind of pre-2010/2011.

Today, link quality matters so much more, and getting a lot of good links from a small set of high-authority domains and high-influence domains is probably just as good, if not better than getting a high degree of links from a lot of diverse domains, none of which have all that much authority.

Plus, when you get linked to by authoritative sites and you're building up credibility in your field, that tends to trickle down to a lot of other second [various] sites in your field. So not as much of a concern. Something to think about only because the domain diversity of links is still well correlated with rankings. If all the links you get are on the same C block or from the same few set of domains or subdomains, that's not going to help your rankings all that much.

All right everyone. Hope that this is helpful for you. I'm sure there's a few factors that you're welcomed to jump into and add to this list and tell me what you think about their varying importance in the comments. We'll see you again next week for another edition of Whiteboard Friday. Take care.

Video transcription by Speechpad.com