Where next, Google?
Today’s guest post is from Ian McCartney of Glaswegian digital marketing agency Equator. Enjoy!
The Exponential Curve of Smartness
There is no doubt about it. Google – already a smart, smart beast – is getting even smarter by the minute. And its ambitions don’t stop at being smart. Google wants to get prescient on yo’ ass. In an interview with UK tech journalist Emma Barnett back in June, Google’s Head of Search Amit Singhal had this to say:
“You need to be able to have a conversation with your search engine. I want my search engine to be the expert who knows me the best. It needs to know you so well that sometimes you don’t need to ask it the next question.”
So, full marks for ambition, Mr Singhal. But in order for that to happen, there’s gonna need to be a gigaton of work to be done yet. Hell, I barely know myself what I’m going to type into Google next – so without some kind of neural sensor (Google Synapse?) being invented, we’re stuck for the time being with ye olde search box. Plus the the much admired, rather complex, and literally awesome algorithm.
And as that algorithm sharpens, Google’s aim is to help us cut a better path through the jungle of data, and find exactly what we’re looking for – no matter what it may be: the mass of the Higgs boson, the diameter of the Death Star, or where to find decent pizza in Edinburgh, Scotland.
But in order to do that increasingly well, Google made some changes. The most recent of which of course were the Penguin update and its slightly scarier cousin, the Panda. When these updates were dropped into the algo, a familiar cry was heard. And that cry was “SEO is dead”. But that’s no bad thing, and here’s why…
Why “SEO is Dead” is a good thing
There was a happy time when SEO could be achieved mainly just by gaming the system. Get a great big pile of links with the anchor text “cheap sheds” and point then to your shed site. Bingo, you see your shed site move up the SERPS. Get more links! Move further up that page!
Didn’t matter where those links were, of course. They could be hidden in pages of unintelligible cyber-dreck, and very often were.
Then there were those odd, odd pages you used to see that were written in some kind of language that bore a resemblance to English, but didn’t make any sense at all. Mmm, automated spun articles – such great fun to read … not. The spinning software had a centrifugal force so strong (linguistically speaking) that all meaning was spun out of the text. Along with any vestige of user value too, I might add.
We all know these old grey and black hat tricks. I don’t need to enumerate them here. Suffice to say that prestidigitation it wasn’t, for the main part at least.
Google’s updates killed a lot of these practices stone dead. Now, those who practiced them were themselves smart, and knew these techniques wouldn’t work forever. But for many businesses it meant temporary gain in exchange for months of pain, their URLs just scorched earth, SERPs-wise.
So, “SEO is Dead” is a good thing – because without Google’s constant evolution and sporadic SERPS cleansing, there would be no SEO by now, it would have been extinct by the mid-2000s. The bulk of of link building would be automated, as would content production. It would simply be a game of whose software could build the most links.
What we have instead is a world where link building still exists, of course, but it’s all about context. Google can still see the links, but it’s beginning to read between the lines more than ever before.
Has anchor text had its day?
This is an open question – we all have our theories about what may happen, so have your say in the comments below. My answer to that question is: obviously anchor text isn’t dead quite yet. But as has recently been pointed out by (a strangely beardless) Rand Fishkin, sites do rank for terms that a) don’t show up on the page or in its title tags/ met descriptions. Or, for that matter, b) in the anchor text of links pointing to the site.
Now, it’s one thing when Google looks at the contextual relevance of anchor text. An anchor text with the words “cheap sheds” in a page that’s otherwise entirely about cats may be bona fide in terms of its placing there by a benign webmaster helping out a friend. Or it could have been placed there for (ahem) commercial reasons. But one thing’s for sure – its relevance is low.
And it’s another thing entirely when a site ranks highly for words/ phrases that aren’t contained in it and haven’t been pointed towards it. In short, this means that Google is using so-called ‘word relationship technologies’ ever more effectively.
In my own experiment to see if all this was true and provable, I did indeed find a site ranking in second place for my chosen term. The term I used was ‘ratings’ for a certain vertical – and the second site in the SERPS was a reviews site that didn’t happen to use the word ratings on its home page, or on any of the individual reviews pages I looked at. Nor in any of the anchors of its backlinks.
What’s happening here? In this case it’s fairly obvious that in certain contexts, “reviews” and “ratings” are – if not synonymous – then at least closely related at a very granular level, and are therefore treated similarly within the algo. This is the basis of latent semantic indexing – the contextual placing of a word being taken as a signal that other words in the same context will have a similar function (if not actual dictionary definition).
Anchor text is still massively useful, of course – when used in good faith. My guess would be that as Google becomes better able to interrogate context, the less it will rely on links generally – especially those that are embedded in pages (if not entire sites) of content that only exists to provide those links. As we have already seen, albeit just the gentle beginning, with the devaluation of web directories and other piled-high, built cheap link depositories. Plus the downright outlawing of spun content and black hat techniques.
So, where next, Google?
Imagine a world with no more anchor text, no more links as value items for SEO. It’s easy if you try. Actually, it isn’t – but it could still happen soon enough…
In this brave new post-link world, what would we see? The birth of hypotext replacing hypertext is one possibility. Here is my definition of hypotext: un- hyperlinked keywords whose power lies not in their semantic value or function but in the context where they are deployed.
And what would that mean for SEO? It would mean that all good content is good SEO. That content, uncontestably and eternally – is king after all. Just like search marketers always said.
As Yoda once said: “always in motion is the future”. The Jedi sage also said that fear is the path to the Dark Side. So, let’s not fear the future.
And if Amit Singhal is right, then Google hasn’t even hit fifth gear yet. Which in turn means that SEO has only just begun…
Ian McCartney works in search content development for Equator, Glasgow. Ian’s favourite topics within SEO and marketing are the Knowledge Graph, web linguistics, and brand semiotics. Outside of work, Ian’s interests and obsessions include sharks, futurology, ambient music, UFOs, science fiction, and decent sushi.