Understanding Google Algorithms - Read Between the Lines

Googles search algorithms is a huge and very complex monster that we, as SEOs, have always tried to reverse engineer and understand to the best of our abilities. The better we understand how the hundreds of parameters involved work the better we can help our clients rank their websites in Google.

So naturally, when Google publicly speak about how elements of their algorithm work we listen. And they actually quite often do that – at their webmaster blog, at industry conferences and in interviews.

But here is the catch – you can blindly trust what Google say.

I am not saying Google is (always) lying. Technically what they say may be true (although it is definitely not always the case!) but the effect may be different than what they try to imply.

In this post I will explore a few examples of Google algorithmic comments and how you can learn to read between the lines.

Google don’t count Facebook likes but value authority

In 2011 and again in 2017 Google stated that they do not count Facebook likes. At some point I even remember them saying that they did not really know anything about what goes on at Facebook. A funny statement, given the fact that they have billions of pages indexed from Facebook, so a little bit about what goes on over there they do in fact know.

So why do Google repeatedly make this statement about not contributing any value to Facebook Likes and is it the hole truth?

To read between the lines of this statement you first have to understand the far from friendly relationship there is between Google and Facebook. Google obviously don’t want to imply that Facebook hold any kind of important key to how Google rank websites. There may even be legal problems if they publicly acknowledged that they “scrape” Facebook details and use it as an important ranking factor.

Also,in general Google don’t want to provide us with too much insight to what actually make a website rank. Over the years they have come out with similar statements about the value of links and how “impossible” they are to manipulate (because Google is too smart). We all know thats not the hole truth.

If you are an experienced SEO you will know that Googles statements about Facebook is also not entirely true. There is a correlation between how much your website is mentioned on Facebook and the authority of the most that mention you. And Authority in Facebook is very much (although not entirely) about likes.

On a side note, you should  off course always be very carefull reading correlation analysis. As you are probably well aware of correlation does not necessarily imply causation. So my assumption that Facebook likes do have a value does not just rely on the many correlation analysis that “prove” this. It is rather based on real life experience.

So if we can’t trust correlation analysis and Googles statement about not counting Facebook likes may be true what is really happening here?

It is really very simple …

First of all I am very sure that Google do actually contribute some value to Facebook authority. But more than that. If many people talk about you on Facebook and like posts about you it is very likely that this interest also travel outside of Facebook. Google read these signals too.

So all in all maybe Google is not directly counting likes in Facebook and may technically be right when they say this but the direct and indirect effects of making you content popular on Facebook do translate into better rankings.

Google say GEO-targeting is not a problem

Another good example of the type of statements Google come out with, that also requires lot of interpretation, is the question of GEO-targeting and redirecting users based on IP.

A couple of years ago former head of the Google quality team Matt Cutts explained it in this video …

The main problem with this video is that Matt is really only teaching you how to not spam Google and how to make it easy for Google to build a good index. Not how to SEO your website the best way. There is a difference.

But its understandable and fair enough. Googles job is to create a great search experience for its users. Not to promote your website for free.

So is what Matt Cutts is saying in the video not true? Well, again, technically it is true. If you follow his advice you most likely won’t get flagged for spamming Google. However, if you goal is to get the best rankings in every region for your website his advise in the video is just not very good.

He say you should treat Google just like you treat any other user. So if a user comes in from a French IP you can redirect that user to the French website. If a user from the US comes enter your French website you can redirect that user to the US site. And Matt say you should do the same for Google bots.

But then how is Google going to crawl the French site?

Googles crawlers most often comes from US IPs. So if you redirect all US IPs to the US site and don’t allow them to enter the French site then they simply can’t crawl it. And if they can’t crawl it then you won’t off course rank for anything on the French site, so French users can find you in Google searching for French content.

A much better solution is to suggest to users visiting from another region that there is another version of your website available that may be better for them. A suggestion in the form of a pop-up or note at the top of the site rather then a full forced redirect is a much better solution if you want great regional rankings.

CONONICAL tags are not always working

Despite Googles unconditional recommendation to use CANONICAL tags to deal with most kinds of duplicate content the fact is that its not a very good solution.

CANONICAL tags are what I call a “band aid solution”.

If you cut yourself and you bleed a a lot putting on a band aid is off course wise. However, it is much better to not cut yourself in the first place. And thats the approach I suggest you take with SEO too.

Many technical SEO problems, such as Duplicate Content issues, can basically be fixed in two ways: On the server or in the client layer.

CANONICAL tags is a client layer solution. The problem with this kind of solutions is that you have no control over how it is read and interpreted (by search engines or others). You can just send it out and hope for the best.

In the past 2 years I have had 3 major cases with large US websites that did perfect CANONICAL tag implementation but for whatever reason Google did not get it right and the sites ended up with serious indexing issues.

On the other hand, server based solutions do not require the visiting agent (browser or search engine spider) to interpret anything. It is as such a much more solid and bullet proof solution.

Off course you can’t always solve all SEO problems such as Duplicate Content on the server. When impossible please go ahead and use a band aid solution. After all, its better to try and stop the bleeding than bleeding to death. Even if you can’t be absolutely sure it works.

Don’t trust Google when they say you can rely on CANONICAL tags. They don’t give you any guarantee that it will. And when it don’t you are the one left with the bill for it – not Google.

In summary – read between the lines and trust real life experience!

Life would be so much easier if there was a SEO manual. If Google would just tell us the truth about how to rank our websites we would not have to read between the lines. But unfortunately thats not how it is.

Off course you should listen to Google when they talk about how Google works. But not only that. For every “fact” they present you have to ask yourself if that is the hole truth. Often its not.

In addition to what Google say you should also listen what we – as experienced SEOs have to say. Or sometimes even better – conduct your own experiments. What exactly works for your business in your market may not always be exactly the same as it is for others.

Whatever Google say reality counts more.

Write comment