‘Last time we talked about SEO myths and one of the biggest ones was links! I bet you’re surprised about that!?’
Nick Garner is founder of Oshi.io casino, a successful Curacao-licensed Crypto/Fiat money operator. He previously worked as search manager for Betfair, then head of search marketing for Unibet.
In 2012, Nick founded the igaming digital marketing agency 90 digital, which he then sold. Semi-retired and looking for new adventures, Nick founded Oshi casino. Nick advises several of his Oshi.io affiliate partners on ranking strategies and tactics.
This time I want to go through some of the research, findings and my own personal experience with the relationship between links and rankings.
Do links work?
Yes. But, it’s more complicated than that. And that’s why I’m writing this article.
In my previous article I talked about correlation not necessarily mean causation. In other words just because two metrics correlate with each other, doesn’t mean they are interdependent with one metric driving the other.
It is another interesting correlation for you:
You know how frustrating Internet Explorer had been in the past? Apparently it leads to an increased murder rate. Who would have known!
My main point here: Just because you you see a big ranking domain with a ton of links, doesn’t mean those links are driving rankings.
To frame this article a bit better, I have used this graphic from Moz ranking factors article which describes the apportionment of various factors that apparently drive rankings.
Overall I more or less agree with this apportionment.
Let’s go through the factors and I’ll give you my opinions on them…
Search visits: 47% correlation
In other words, sites that get more visits from search traffic are more likely to rank.
My take: Google wants content pages that its users want. If a Page gets a higher-than-expected click through and engagement rate, it’s obvious users want that content in thein the context of the search query they used.
Domain authority 39% correlation
Sites which have more ‘authority’ i.e. trust and expected satisfaction from users generally do better in rankings.
My take: Google wants websites that users want. If Google is confident about a website satisfying its users, it will rank its content with less upside friction than other websites.
Number of linking root domains: 31%
My take: If a link is a vote, then 100,000 footer links from a single domain is not 100,000 votes. The number of unique linking root domains makes more sense as an editorial vote in favour of a site.
Number of Facebook shares: 20%
My take: people share content they think other people will like. Google probably don’t use social media metrics, because they don’t own that data. Also just because a page has a lot of shares, doesn’t mean it’s satisfying in the context of any search query. This is just a loose indicator that content is possibly satisfying.
Keyword in the title: 3%
My take: title tags are the equivalent of a headline in AdWords placement. If the title tag describes the content and is ‘clickable’ , that increases click through to the page and therefore acts as a signal saying the content is satisfying. Although, 3% is a low correlation
Keyword meta description: -2% correlation
My take, there’s a negative correlation between meta descriptions and rankings. That’s probably because Google takes its meta description information from whatever part of the page is most relevant to the search query being used.
At this stage, you’re properly thinking ‘where’s the links in all of this?’
I want to give you context on where links sit in the big scheme of things. If we assume you’ve done your SEO housekeeping, sorted out your indexation and other basic hygiene, then in my view user satisfaction is probably 85% of what drives rankings, with links being around 15%.
Here are some very good research on links, correlations and ranking factors:
The stone temple article summary put it nicely:
“we don’t find that links can rescue poor quality content”
In a ruthlessly competitive segment like iGaming, that 15% you get from links can be the difference between making or breaking your project. However, in my experience people get fixated on links and forget about user satisfaction. In later articles we’ll talk a lot more about user satisfaction and how to analyse websites to work out what will satisfy your users and make you loads of profit.
Mental framework for link acquisition
There is a fresh website i.e. mylatestwinningcasinobonuses.com and its got zero reputation, history, links.
Affiliate publishes mylatestwinningcasinobonuses.com and there is no signal for Google to say whether this is an important site or not.
Google needs signals to make decisions about rankings, so you go and dump in a bunch of links that work.
Google sees this signal, ranks your website and you’re now being auditioned.
If your content is satisfying in the context of the key phrases you rank for, Google will want you to be more prominent in its search results, because users want you. Therefore you rank higher and get more traffic.
When engagement kicks in, you don’t need more links, although you’ll properly get them organically.
Google accumulates trust in your domain, you have more and more authority and rankings finally stabilise at the point where your rankings equate with how satisfying your content is in the context of the keywords it ranks on.
To back this hypothesis up you can read lots of big reports on user engagement driving rankings, but anecdotally if you:
- pick out a key phrase you care about
- look at the pages ranking above you
- put yourself in the mind of the Google user and their search query i.e. ‘casino bonuses’
- Do those top ranking sites satisfy you? Can you find casino bonuses you want? 10 years ago, I think you might not be so satisfied, but today I’m pretty confident on high traffic big-money phrases, the pages which rank deserve to rank.
Site gets links
Google gives you bump in ranking
User satisfaction then mostly drives rankings.
The flipside is when you’ve got a domain that has been auditioned, tanked in the rankings and then a site owner attempts to resuscitate the domain by pumping in a load of good links.
If the content hasn’t changed and is as unsatisfying as before, that site owner is going to burn their money.
Why does this make sense? Google wants satisfying content on its search results. If content is satisfying AND it gets more links, Google will audition the page across more search queries and with higher ranking positions.
What is a good link?
At a very high level, it’s a link that acts as an editorial vote in favour of another piece of content. In other words, I like that content, I’m linking to it and I’m helping my users journey around the Internet.
As you Bobby know, Google’s page rank algorithm, the founding algorithm for Google was based on academic papers and the number of citations to those papers.
Prominent and important academic papers are cited more often by other academics than unimportant academic papers. Prominent and important pages are linked to more often than unimportant pages.
In the beginning page rank algorithm drove far superior search results than any other search engine. And then the spammers came. For a while in early 2000’s, anyone with enough links could rank.
Unsatisfying content started to percolate all over Google search results and Google were unhappy.
Eventually machine learning kicked in and the Penguin algorithm made a good job of zeroing with no editorial vote and links became a more useful signal to Google.
Why I don’t believe in trust flow/domain authority and other similar metrics
Somebody decides to launch an SEO tool, index the Internet and emulate Google’s algorithms. They come up with some metrics that end up as a standard currency on the Internet. Anyone tied in with SEO ends up with a deluge of link sellers peddling high domain authority links…
For me, the right link is:
From pages and domains Google trusts. Google will trust pages and sites it ranks.
So when we analyse domains, we look at:
Organic search traffic
Number of referring unique domains
Rankings for those referring unique domains
I.e. if a domain ranks and has lots of other ranking domains linking into it, it’s probably a very authoritative, trusted site. And I want a link from that domain.
I use a tool called digimetr.com for bulk analysis of domains. It’s a pay-as-you-go tool and costs around two dollars to query around a thousand domains. You can get proportion of search traffic, volumes of traffic, sources of the traffic and so on. These metrics are very helpful for outline analysis.
If I want to go deeper, I will probably use either SEMrush or ahrefs to see the ranking trends and keywords the site ranks for.
If it’s a strong site, then we will do a deal with the publisher ‘share a little money’ and get that link placed.
Does link context matter?
In my opinion not so much so. It’s a huge topic, but the rule of thumb is stay within norms. I.e. if you’re a casino affiliate site and you’ve got a vast proportion of links from gamer sites, that might hold together.
I don’t think geolocation matters for link placement. Again, if you look at link profiles of domains geolocation can be very divergent.
Rule of thumb: look at competitors, see where most of the links are coming from, filter out anything that doesn’t rank and what’s left over is a good indicator for the kind of content range that is acceptable for your link placement efforts.
This is a very nuanced subject, because one has to discriminate between links that are just junk, which Google ignores and ones which Google pays attention to that are toxic.
In Brighton SEO 2017, Gary Illyes (Google Webmaster trends analyst) responded to the following question:
Interviewer: The link selling market has exploded obviously since Penguin came out because there’s, it’s being discounted. There’s a lot of people saying yeah we’ll get you all these links from forums and all these various publications, some better than others. And because, we don’t have to worry about Google because Google’s just going to ignore the bad ones and we keep all the good ones. What do you say about that?
Gary Illyes: If you are thinking about link building by buying links then yeah go ahead and buy links from forums for example. That would be great for your business because you lose like $3,000 or something like that and you will not benefit from that. Another thing is that you do have to keep in mind, that we still have manual actions for links and I’m actually very happy when I see that some bigger site has been hit with a manual action for buying links. Mainly because we are really clear on a lot of this in our guidelines and if you go against that then, if you actively trying to manipulate our search results, you deserve that manual action. And, unfortunately, it’s very easy to get rid of that manual action because we have the disavow tool for example, which I don’t necessarily like. But then I’m not in a position to make or for example remove this tool. I would do it, but maybe because I’m an ass.
Essentially he’s saying ‘burn your money on junk links if you want, but it’s not going to help you’
So in theory no link is toxic, but I have had experience with negative SEO attacks and if you get hit by junk auto generated spam links, ironically, amongst those millions of ‘zeroed’ links might be a few that help you. (This has been the experience of a couple of SEO’s I know)
The links to watch out for are from legitimate sites which aren’t built for link farming and have ‘illegal/antisocial/horrible/toxic’ content. Those sites link to you, Google sees they are ‘valid editorial votes’ of the worst kind.
Fortunately if these sites aren’t in the dark web, they are indexed and you can disavow them.
And yes, disavowing was a huge social engineering project to gather data on sites that are seen as ‘toxic’. This data was then fed into Google’s machine learning algorithms and ultimately became part of Penguin. You helped Google fight link spam.
Link acquisition can be very simple. Get links from sites Google trusts. Google trusts sites it ranks. Don’t get too divergent with the content match, but you got lots of latitude i.e. casino bonuses on a gamer site. And remember that links are probably only 15% of the ranking story.
If you’ve got a link acquisition budget and you haven’t invested in user satisfaction… Sort out user satisfaction first, then get your links. If you do the other way round, you get ranked temporarily, (probably crash gracefully) and then it’s far harder to get back once you sort your content out.
Next time were going to talk about project planning so you can organise your ranking campaign systematically and profitably.