I really believe in data being the source of answers to big questions. When data is well organised and structured, it makes complexity relatively simple and allows you to compare like for like. That’s why value the Searchmetrics ranking factors 2015 report so much. You can download it here: Ranking factors 2015
There is also another report which comes out each year from Moz: https://moz.com/search-ranking-factors/survey , which polls a group of SEO experts and aggregates their opinions.
To me, the best way to understand what’s going on in the world of SEO is to read the Searchmetrics support and then read the Moz report to get the ‘human’ take on how the engines are moving.
But for now, I want to dig into the Searchmetrics support in some more detail because it’s a good way of outlining my general philosophy and views on which ranking signals count.
The most comprehensive SEO report of the year: released.
If you’re not into SEO, I’ll just give you a brief explanation on what I’m talking about. Searchmetrics is a German company that have a suite of enterprise level SEO tools and analytics. Each year they release a White Paper analysing their datasets for correlations that relate to SEO. Essentially, the correlations suggest certain things for instance click through rates from search results affect rankings…
If you believe like I do, that data gives clarity, then the Searchmetrics report is the clearest, most useful baseline indicator of where the search engines are going. So we will go into this in a bit more detail later, focusing on iGaming
The reason these things are important for iGaming is because of the huge amounts of money spent on SEO. And if operators are spending money, they want to make sure it’s well employed. A good example in my view is the declining importance of links, despite the fact that enormous sums are spent buying links unnecessarily.
The Searchmetrics report uses Google.com as its data source. Because Google is essentially a unified infrastructure, irrespective of what’s English-speaking variant of Google you use, the metrics are going to be comparable.
The report looks at correlations and then draws some insight from those. It’s up to the reader to decide whether those correlations are meaningful are not. A good example of a correlation I believe in is around click through rate and rankings. Put simply, if you get more clicks than expected for the search position you’re in, you’re likely to go up in the rankings. Conversely less clicks and you go down.
There are other correlations like social signals, where more social signals seem to indicate higher rankings. Based on this idea, there is a whole world of fake Facebook/LinkedIn/Pinterest likes and interactions.
People believe what they believe… And I suppose the reason I’ve been asked to write this column is because I’m experienced and hopefully I can see which signals count and which don’t.
Let’s go through some signals.
Because the report covers 60 different signals and because you would probably get bored reading this article after an hour, let’s cover off the high-level headers and particular signals which are most interesting. Along the way I’ll apply these ideas to iGaming
- Technical – things like site speed, H1 tags, secure search
- User experience – internal linking, video integration, users signals
- Content – word count, keywords on page, keywords in links
- Dominating sites – Wikipedia, Facebook
- Social signals – their effect
- Mobile ranking factors – mobile friendliness, recent mobile update
- Backlinks – number of backlinks, anchor text, back link age
H1 tags etc
The report just talks about the existence of description, H1 tags, H2 tags in over 80% of top ranking sites. The question is: does this help you rank and in my view, not necessarily but it’s a very good housekeeping thing to do.
Keywords in a domain.
This is very interesting, because in my view there is a persistent misconception that you should have a relevant keyword in a domain for it to rank. This chart shows you the percentages for 2015 as compared to 2014. Essentially, there are fewer sites now with keywords in the domain than last year. Importantly, we are talking about an average of around only 6% of sites having keywords in their domain names, within the top 30.
Comment: In iGaming, there is a theme around buying junk dotcom domain names, loaded up with keywords like casinoroulettecasino.com . As you’ll see, user engagement is the biggest driver for rankings, and behind that is links. If that’s the case, then brand is really important… So why ever think of registering a domain just because it’s got keywords in it?! For clarity, think brand first.
Search volume of a domain name
If your domain name is searched for, you’re more likely to rank well on other key phrases.
Comment: This touches on one of the great problems affiliates have: no brand.
No brand equity means little engagement with your website and therefore less chance of building authority and trust from Google and users. By the way, Panda is all about helping the most trusted and authority websites get ranked.
Anyhow, the correlations around search volume for your brand name and subsequent rankings and other key phrases has jumped dramatically from 2014. Lesson: brand.
Ratio of homepage rankings to interior pages
You may have noticed, there are fewer and fewer homepages ranking on search results. This is because Google is getting far better at giving you pages offer trusted website that are most relevant to your search.
Approximately 87% of the results within rankings from 2nd position to 30th are interior pages. And this trend has increased steadily since 2014.
Comment: The main point? In iGaming there is a big thing around shovelling links into money pages and hoping they rank. Whilst it sort of works, what works far better now is to concentrate on link acquisition to the homepage, good internal linking and relevant content on a page level. In other words there is a huge disconnect between where the links go and where the rankings happen.
Proportion of .com domains in top 30 search results
The proportion is of.com domains has increased slightly against last year to around 84% in the top 10 search results and 81% for the next 20 search results.
Comment: There is a big idea that because you own a.com domain, it’s going to help you rank. Not true. With the emergence of hundreds of new domain extensions coming through, I think you will see the fall of the .com dominance.
I’ve seen so many sectors, for instance bit coin gambling where the most fashionable domain extension seems to be .io and sites with these extensions dominate rankings.
The average loading time for desktop pages within the top 30 search results is 1.2 seconds. For the top 10 search results, it is 1.16 seconds.
There is an argument that site speed is a ranking factor, but I think it’s an indirect one. In other words, slow sites put users off and they bounce straight out, giving a horrible signal to Google saying users do not like the site. Of course Google would pick up on this and do you rank the site.
With this mean for iGaming? There’s a great temptation to just overdo it on heavy images, video, interactive content when in fact people want speed so they can leaf through a site quickly and find what they want. Now that I am working on the operator side, with our new casino I’m going to make site speed and absolutely top priority. And I would rather have fewer graphics in favour of site speed all day long.
Once upon a time, search engines didn’t have the computational power to factor in user experience into their algorithms, that’s why links were so important. Links as a signal are relatively low cost computationally and in many ways are a good proxy for popularity of content.
But as the processing power of Google and other search engines has increased, they can go deeper into user’s signals to decide which content is actually popular. Hence, the slow death of links as a way of ranking.
In this section, there are a number of indicators like the amount of advertising on a page, usage of images, video integration and so on. The big one here is click through rate.
Click through rate
Comment: Put simply, higher search results get more click through rates. Of course! But what people in SEO forget, is that when a page gets a higher click through rate, it then ranks higher… It’s a very simple signal for Google to use because they have all the data they need to determine whether a page is getting a disproportionately high click through rate for its position.
I’ve written up a long post outline my arguments for engagement here:
http://90digital.com/seo-comment/engagement-as-a-ranking-factor-my-best-arguments-for-it-8289.html -Google search: “90digital engagement”
In summary, there is a 67% correlation between click through rate and search result position. That is the biggest correlation Searchmetrics have in their whole 2015 ranking factors White Paper.
The average word count for a top 10 search result per page is 1285 words, up from 975 words in 2014.
Comment: I think this statistic can lead to some really nasty misconceptions around what should rank. What do I mean? If you saw this statistic, you might think you should add 1285 words to your casino homepage to rank.
I have seen so many sites with no words on a given page, rank on extremely competitive phrases. The only reason they can rank is because of their title tags.
If you think of engagement as the primary ranking driver, if people click on a search result, then Google will do whatever it can to rank it irrespective of the amount of content on the page.
The reason the average word count is longer than it was in the previous year is because Google are ranking far more internal pages, which naturally have more words on them covering a specific topic.
It typically ranks in first place around 20% of the time, with second place at 29% of the time. One very interesting point is that Wikipedia has lost roughly 20% of its first-place positions since 2014. (It previously ranked number 1 26% of the time)
Comment: I think Google had given Wikipedia a massive free pass to rank across its index, however now we have engagement coming through as a really strong ranking factor, it means brand owned pages can outrank Wikipedia, simply because people are more interested in those.
As with Wikipedia, Facebook seem to be given a free pass on Google to rank wherever relevant. The most common ranking is around fifth position for a Facebook page.
This touches on a very interesting point about reputation for operators. Facebook is a good way of having a conversation with the user, but it doesn’t actually help sell a brand because the content is owned by the brand and therefore by definition is biased. This is why Wikipedia pages are often more regarded, because of the rigourous (and irritating) editorial process in place to screen content.
Social signals generally correlate with higher rankings.
- Facebook: 28% correlation
- Google+: 31% correlation
- Twitter: 23% correlation
- Pinterest: 23% correlation
Comment: In other words, sites that have a lot of user engagement typically rank higher. The question is whether social signals create rankings or rankings create social signals…
Because I believe engagement is the dominant ranking factor, sites that have a lot of social engagement are simply sites people engage with and therefore rank because of engagement not because of social.
Some of you may have heard about ‘Mobilegeddon’ , it was a big update Google made two separate desktop search results from mobile search results, lifting mobile friendly sites up the rankings.
In summary, Google is slowly improving mobile search results, but the changes that were expected earlier this year, have not been as dramatic as forecast.
We’re talking about a 3% change in the number of mobile friendly sites being ranked over a four-month period from the Google mobile update.
Big subject for iGaming…
There’s a 22% increase in correlation between number of backlinks and rankings from 2014. The chart below shows you referring domains by search position.
What’s most interesting is how the first search result has far fewer referring domains than the next 17 search results after that. I put this down to huge authority sites ranking across vast swathes of the Internet filling up the lower search results.
Comment: I have a theory about why people go on about link acquisition far more than engagement in SEO. I think it’s because engagement is tied in with brand and with the operators themselves, so it’s hard for an SEO agency to get right into the core make up of a website.
Whereas with links, it’s an externally driven activity and is easily accounted for within the business i.e. give me ‘X number of links for Y amount of money’. Therefore, it’s led to a massive spamming of the Internet, with search engine updates like Penguin coming in to clean up the noise.
Personally, the writing is on the wall from me – link acquisition is dying. What I’ve seen are high quality links from trusted domains having a huge impact, but not lower level links.
When link acquisition is married up with engagement optimisation, the rankings will follow.
No follow links
The ratio of no follow links has increased to 9% for the top 10 search results.
Comment: this just highlights two things. Firstly, most of the links on the Internet or follow links, secondly people are afraid to give follow links for fear of repercussions with Google, hence the small rise.
My overall conclusions
Search engine results are closer reflecting what people actually want, than ever before. Links are still important, but there’s an inflated marketplace for them, because it’s such an easy thing to sell relatively speaking.
Engagement is the biggest factor of all, but it’s also the toughest because it requires creativity, imagination and most importantly a willingness to align yourself with a users agenda.
Nick Garner is CEO of Oshi BitCoin Casino and founder of 90 Digital, the iGaming SEO agency.