What Does Machine Learning Mean For SEO?

  • By Spotzer
  • 22 Jun, 2016

Staying on top of top-secret algorithm updates from Google is tough enough. So how should SEO experts react as the algorithm starts to update itself?

Google has been in love with the idea of machine learning since forever, but it’s taken some time to inject their beloved search engine with the self-instructing intellect that has seen its popularity explode at forward-thinking tech conferences and Folgers-fueled startups across the globe. The latest addition to Google’s algorithm was revealed recently, and it’s even got a catchy name that’s sure to confuse the uninitiated.

Say hello to RankBrain (and thanks for helping to build it)!

The shift toward machine learning in Google’s algorithm can be summarized in a simple new term: RankBrain . As with everything Google, the layers go much deeper, but it’s a great starting point for anyone with an idle interest in SEO who wants to catch up.

In short, RankBrain is the machine learning component of Google’s complex Hummingbird algorithm (the “whole puzzle” of which RankBrain is one small but growing piece). On a whole, the algorithm pays attention to queries and the result of those queries, then runs the data in real-time through its magic transmogrifying box to deliver SERPs that users find useful. Delivering quality search results has always been an iterative process, but RankBrain is the first time Google has solicited maintenance assistance from the algorithm itself .

So Google is watching my every move?

No, not really – but yes, definitely it is. From what we know, RankBrain is paying attention to your behavior (and guessing at your satisfaction) using as much superhuman logic as necessary to create a personal experience.

Despite the human mimicry, it’s still fundamentally based on math, and it’s still just a piece of the greater Hummingbird whole. They aren’t revealing many specifics, and we know its involvement is currently limited to uncommon informational queries, but you can bet your bottom dollar that machine learning is going to become more important as cookies, meta tracking, localization and other personalization techniques further integrate into our lives.

We envision the use case for RankBrain and machine learning as looking something like this:

  1. User 1 enters a unique query
  2. Google delivers some results they might want using some RankBrainpower and some classic algorithm magic
  3. User 1 chooses a result and takes an action (bounce, click-through or revise search terms)
  4. Google listens and thinks really hard
  5. More users enter the query
  6. Time passes
  7. ??? (Transmogrification?)
  8. User 1,000,000,000 enters the same query
  9. Google delivers the exact result they are looking for using pure insight from RankBrain – even though the user intent and the delivered result are different from User 1!
  10. Google is renamed RankBrain upon being devoured by the now-sentient supercomputer lava monster
  11. Larry Page dies happy

Right now RankBrain is watching and, well, learning. It’s probably altered at least a few of the searches you’ve performed today, and while it’s not all the way there yet, we won’t be surprised if Google’s search algorithm is completely replaced by a machine learning system that has essentially written itself in a few short years.

Before you ask...

RankBrain is not PageRank
RankBrain! PageRank! BrainPage! PainBrank! BakePang!

No, RankBrain is not PageRank.

That’s different. PageRank looks at links and the way your website is “supported” by the web. It’s an increasingly outdated metric that Google has been backing away from for a while now; it’s too easy to game and fits into the “checklist” approach to SEO (the mentality that completing a series of tangible tasks – regardless of the product, scope or audience – is all you need to do to compete on SERPs).

Why did they name RankBrain so similarly? Beats me. It’s either a way of telling us that it’s a proprietary concept worth respecting, or it’s simply telling of their creativity.

It’s better than PageRank!

According to Google, at least.

A few months ago, RankBrain shocked the SEO world when it was unveiled as Google’s third most important ranking factor – an uncharacteristic admission from a company who is known for getting a bit over-excited when they think they’ve done something cool.

Only time will tell if our world-devouring lava monster prediction turns out to be true, but there’s no doubt that machine learning – and the ramifications of its implementation on a theoretical level – is huge for Google. They’re carrying the concept across all their products ( not just search ) and truly believe in its power to connect their users to the results (or, you know, the advertisements) they’re looking for.

Ok, but what does this all mean for SEO?

Machine learning is a nascent player in the search game, which means the iron is about as hot as it will get before it starts to burn anyone unprepared. If your company isn’t already planning a shift in SEO strategy to accommodate the “new Google” as it unveils itself in little leaps over the years, you’re already behind.

A few things will change (or continue as they are, but with greater gusto):

  • The SEO job will get harder (but more useful)
  • The user will become (even more) powerful
  • Creating useful content will become (even more) important

Twenty years ago – in the dark ages of Google – search engines were fertile ground for anyone looking to leverage some basic tech knowledge into quick cash. Black hat tactics and spam dominated the internet , funneling queries into their own private webs of messy “nontent” (auto-generated content that doesn’t actually deliver anything) in hopes for a referral click from a confused user. Casting a wide net and checking a few boxes usually meant success on the SERPs.

Things have changed since then. Guestbook spam has died off and most linkbuilding schemes have long since squeezed their last unscrupulous penny (though that hasn’t stopped others from continuing to try). Google’s algorithm changes 500-600 times per year to accommodate the way we search for and discover content, and the RankBrain rollout is a presage of greater dependency on listening to the wishes of its users.

It’s simple: SEOs have always relied on analyzing the signals that Google uses to rank websites, working backwards from there to create something the search engine agrees should rank favorably. In the past, this allowed for checklists, bullet-point value propositions with unstable guarantees and a systematic, methodological approach to SEO. The game has changed. Google is moving away from technical analysis – an uneven playing field – toward a system wherein the end customer (you) is also the one delivering feedback on the product as it develops.

So if Google is looking to the user for its signals, shouldn’t you? And with the rules to the SEO game being in a constant state of flux (or, more accurately, the referee changing on every play) while SERP conversion continues to rise , don’t you think successful SEO campaigns will only become more noticeable/valuable?

You have less power than ever – use it!

On first glance, it does seem that machine learning will take control from the hands of SEO experts. It’s true to some extent – it will certainly make a lot of people who were spamming the “easy button” miserable, and it’s definitely one more nail in the black hat coffin.

That’s the pessimist’s guide to RankBrain. Others see it as the harbinger of a new landscape in which the goals of SEOs and Google are closer aligned than ever. It’s true that business models built on deception and spam will be negatively affected, but any honest business (SMBs, personal bloggers, major brands) will see this as a huge boost. Most business owners know what their customers want; updates like RankBrain allow them to eschew the checklist/black hat mentality so they can focus on delivering timely, useful content that increases engagement and loyalty. If you do it right, the SEO benefits will tag along.

To put it simply (with a touch of exaggeration for illumination purposes):

  • If Users teach RankBrain
  • And RankBrain teaches Google SERPs
  • And you are a User
  • Then you teach Google SERPs

All you have to do is figure out what you want. For SEOs (like our crack team at Spotzer !), that means putting yourself in the shoes of your target audience and serving yourself up a plate of fresh content. How’s it taste? You’re the chef and the critic – but remember, the critic lauding Kraft mac’n’cheese as a masterpiece usually reserves a table for one. As soon as you start debasing your morals – ok, as soon as your audience discovers you’ve debased your morals – they’ll run away.

In the past, when SEOs controlled much of the web, this spam-based catch-all approach was effective; your poor user would just run into another of your clever traps. That’s not the case any longer. Machines are in charge now, and that means we, the humans, are finally getting the SERPs we deserve.

Thanks, machines!

Spotzer Blog

By Spotzer 07 Sep, 2016
It’s been a busy summer at Spotzer: the IT team successfully concluded a number of complex partner launches, Jon Stribling settled into the Amsterdam office as Chief Digital Officer, our Italian team quadrupled in size, and the weather has kept us outdoors regularly (rare for a Dutch summer).

September is set to be Spotzer’s biggest month ever, which we are very excited about. Our teams continue to grow across the world, and it’s getting harder and harder to choose between the excellent performances we’ve seen in every office... but we gave it a try anyway!

Here are Spotzer’s Employees of the Month for August, along with a bit of explanation from their manager:
By Spotzer 29 Aug, 2016

Spotzer is proud to announce a partnership with Infusionsoft to provide an innovative best-in-class solution for websites and paid search campaigns within Google. The new digital products will take advantage of automation and analytics tools to give Infusionsoft's SMB customers an edge over their local competitors. 

Spotzer will create websites for SMBs and optimize them for Pay-Per-Click (PPC) performance using Spotzer’s proprietary ad serving technology, Ad-Direction . The unique localization tool will integrate with Infusionsoft’s powerful automation software to fine-tune campaigns and generate effective leads for their customers.

By Spotzer 23 Aug, 2016

If you’re planning to open an innovative technology company in Europe, your options just became clear. The European Union Committee of the Regions recently named their top three European cities for innovation , and we were quite pleased (though not surprised) by the results:

#1 – Amsterdam, Netherlands

#2 – Turin, Italy

#3 – Paris, France

Amsterdam edged out a field of 36 competitors across the EU to claim the title of “iCapital” (Innovation Capital of Europe). Each city received a substantial cash reward to support further efforts in sustaining growth in innovative tech fields.

Spotzer’s global headquarters are based in Amsterdam. It makes sense that these cities are topping the charts for promise and potential, as we’ve seen massive interest in high-quality digital products in these markets; some reactionary, wherein a disruptive offering could easily transform the status quo, and some traditional, following a more linear and predictable path to innovation.

Spotzer sees these awards as confirmation of our direction – not only from a product perspective, but geographically as well, and we look forward to continuing our contribution to these fantastic cities.

More Posts
Share by: