"I give you the light of Eärendil ... May it be a light for you in dark places,
when all other lights go out."

Even if I'm interested in large-scale correlation tests like the Moz Search Engine Ranking Factors, in reality I am convinced that the science in which we best excel is that of hindsight.

For example, when Caffeine was introduced, almost no one imagined that that magnification of the SERPs would have meant its deterioration too.

Probably not even Google had calculated the side effects of that epochal infrastructural change, and only the obvious decline in the quality of the SERPs (who remembers this post by Rand) led to Panda, Penguin, and EMD.

But we understood just after they rolled out that Panda and Co. were needed consequences of Caffeine (and of spammers' greed).

And despite my thinking that every technical marketer (as SEOs and social media marketers are) should devote part of their time to conducting experiments that test their theories, actually the best science we tend to apply is the science of inference.

AuthorRank is a good example of that. Give us a Patent, give us some new mark-up and new social-based user profiling, and we will create a new theory from scratch that may include some fundamentals but is not proven by the facts.

Hindsight and deduction, however, are not to blame. On the contrary; if done wisely, reading into the news (albeit avoiding paranoid theories) can help us perceive with some degree of accuracy what the future of our industry may be, and can prepare us for the changes that will come.

While we were distracted...

While we were distracted—first by the increasingly spammy nature of Google and, secondly, by the updates Google rolled out to fight those same spammy SERPs—Big G was silently working on its evolution.

Our (justified) obsession with the Google zoo made us underestimate what were actually the most relevant Google "updates:" the Knowledge Graph, Google Now, and MyAnswers.

The first—which has become a sort of new obsession for us SEOs—was telling us that Google didn't need an explicit query for showing us relevant information, and even more importantly, that people could stay inside Google to find that information.

The second was a clear declaration of which field Google is focusing its complete interest on: mobile.

The third, MyAnswers, tells us that Personalization—or, better, über-personalization—is the present and future of Google.

MyAnswers, recently rolled out in the regional Googles, is a good example of just how much we were distracted. Tell me: How many of you still talk about SPYW? And how many of you know that its page now redirects to the MyAnswers one? Try it: www.google.com/insidesearch/features/plus/‎.

What about Hummingbird?

Yes, Hummingbird, the update no SEO noticed was rolled out.

Hummingbird, as I described in my latest post here on Moz, is an infrastructural update that essentially governs how Google understands a query, applying to all the existing "ranking factors" (sigh) that draw the SERPs.

From the very few things we know, it is based over the synonym dictionaries Google was already using, but applies a concept based analysis over them where entities (both named and search) and "word coupling" play a very important role.

But, still, Google is attending primary school and must learn a lot, for instance not confusing Spain with France when analyzing the word "tapas" (or Italy with the USA for "pizza"):

But we also know that Google has bought DNNresearch Inc. and its deep neural networks, which had gained great experience in machine learning with Panda, and that people like Andrew Ng moved from the Google X team to the Knowledge Team (the same of Amit Singhal and Matt Cutts), so it is quite probable that Google will be a very disciplined student and will learn very fast.

The missing pieces of the "future" puzzle

As with any other infrastructural change, Hummingbird will lead to visible changes. Some might already be here (the turmoil in the Local Search as described by David Mihm), but the most interesting ones are still to come.

Do you want to know what they are? Then watch and listen to what Oren Etzioni of Wavii (bought by Google last April 2013) says in this video:

As well described by Bill Slawski here:

The [open information] extraction approach identifies nouns and how they might be related to each other by the verbs that create a relationship between them, and rates the quality of those relationships. A “classifier” determines how trustworthy each relationship might be, and retains only the trustworthy relationships.

These terms within these relationships (each considered a “tuple”) are stored in an inverted index that can be used to respond to queries.

So, it can improve the usage of the immense Knowledge Base of Google, along with the predictive answers to queries based on context. Doesn't all this remind you what we already see in the SERPs?

Moreover, do you see the connection with Hummingbird and how it can link together the Knowledge Graph, Google Now, and MyAnswers; and ultimately also determine how classic organic results (and ads) will be shown to the users over a pure entity-based and semantic analysis, where links will still play a role, but not be so overly determinant?

So, if I have to preview the news that will shake our industry in 2014, I would look to the path Wavii has shown us, but also especially to the solutions that Google finds for answering the questions Etzioni himself was presenting in the video above as the challenges Wavii still needed to solve.

But another acquisition may hide the key to those questions: the team from Behav.io.

I say team, because Google did not buy Behav.io as a society, but the entire team, which became part of the Google Now area.

What was the objective of Behav.io? It was looking at how peoples’ locations, networks of phone contacts, physical proximity, and movement throughout the day could help in predicting a range of behaviors.

More over, Behav.io was based over the smart analysis of the all the data the sensors in our smartphones could tell about us. Not only GPS data (have you ever looked at your Location History?), but also the speakers/microphones, the proximity detection between two or more sensors, which apps we use and which we download and discard, the lighting sensors, browser history (no matter which search engines we use), the accelerometer, SMS...

You can imagine how Google could use all this information: Again, for enhancing the predictive solution of any query that could matter to us. The repercussions of this technology will be obvious for Google Now, but also for MyAnswers, which substantially is very similar to Google Now in its purposes.

The ability to understand app usage could allow Google to create an interest graph for each one of us, which could enhance the "simple" personalization offered by our web history. For instance, I usually read the news directly from the official apps of the newspapers and magazines I like, not from Google News or a browser. I also read 70% of the posts I'm interested in from my Feedly app. All that information would normally not be accessible by Google, but now that it owns the Behav.io technology, it could access it.

But the Behav.io technology could also be very important for helping Google understand what the real social graph of every single person is. The social graph is not just the connection between profiles in Facebook or Twitter or Google Plus or any other social network, nor is it the sum of all the connections of every social network. The "real life social graph" (this definition is mine) is also composed of the relations between people that we don't have in our circles/followers/fans, people we contact only by phone, short text messages or WhatsApp.

Finally, we should remember that back in 2011 Google acquired two other interesting startups: PostRank and Social Grapple. It is quite sure that Google has already used their technology, especially for Google Plus Analytics, but I have the feeling that it (or its evolution) will be used to analyze the quality of the connections we have in our own "real life social graph," hence helping Google distinguish who our real influencers are, and therefore to personalize our searches in any facet (predictive or not predictive).

Image credit: Niemanlab.org

Another aspect that we probably will see introduced once and for all will be sentiment analysis as a pre-rendering phase of the SERPs (something that Google could easily do with the science behind its Prediction API). Sentiment Analysis is needed, not just because it could help distinguishing between documents that are appreciated by its users and those that are not. If we agree that semantic search is key in Hummingbird; if we agree that Semantic is not just about the triptych subject, verb, and object; and if we agree that natural language understanding is becoming essential for Google due to Voice Search, then sentiment analysis is needed in order to understand rhetoric figures (i.e. the use of metaphors and allegories) and emotional inflections of the voice (the ironic and sarcastic tones, for instance).

Maybe it is also for these reasons that Google is so interested in buying companies like Boston Dynamics? No, I am not thinking of Skynet; I am thinking of HAL 9000, which could be the ideal objective of Google in the years to come, even more so than the often-cited "Star Trek Computer."

What about us?

Sincerely, I don't think that our daily lives as SEOs and inbound marketers will radically change in 2014 from what they are now.

Websites will still need to be crawled, parsed, and indexed; hence technical SEO will still maintain a huge role in the future.

Maybe from a technical point of view, those ones who still have not embraced structured data will need to do so, even though structured data by itself is not enough to say that we are doing semantic SEO.

Updates like Panda and Penguin will still be rolled out, with Penguin possibly introduced as a layer in the Link Graph in order to automate it, as it happens now with Panda.

And Matt Cutts will still announce to us that some link network has been "retired."

What I can predict with some sort of clarity—and for the simple reason that people and not search engines definitely are our targets—is that real audience analysis and cohort analysis, not just keywords and competitor research, will become even more important as SEO tasks.

But if we already were putting people at the center of our jobs—if we already were considering SEO as Search Experience Optimization—then we won't change the they we work that much.

We will still create digital assets that will help our sites be considered useful by the users, and we will organize our jobs in synergy with social, content, and email marketing in order to earn the status of thought leaders in our niche, and in doing so will enter into the "real life social graph" of our audience members, hence being visible in their private SERPs.

The future I painted is telling us that is the route to follow. The only thing it is urging us to do better is integrate our internet marketing strategy with our "offline" marketing strategy, because that distinction makes no sense anymore for the users, nor does it make sense to our clients. Because marketing, not just analytics, is universal.