• Driverless cars, paired with companies like Uber, will soon transport us about the city, offering us the extra time to revise our notes before an important business meeting.
• Leap-motion interfaces are available.
• Marketing has become personalized and targeted.
• We have experienced a loss of privacy in the name of a supposed greater safety.

In many respects, we can say that the future is (almost) now.

Of all the things that were presented in Minority Report, the one that most concerns us as SEOs and inbound marketers is the personalization of experiences that our potential customers have when looking for a product and/or information, when they share things online, and when they interact with our brands on our websites.

Search marketing and personalization

Personalization in search marketing is not something new—it was (re)launched on Google in 2005. Still, it was only with the launch of "Search, plus Your World" (January 2012), the rollout of the Venice Update (February 2012), and the introduction of Google Now (July 2012), that the personalization factor has become predominant.

If we ask everyday Google users about personalized search, though, this is what they answer:

This data from the excellent infographic on seotraininglondon.org reveals something that we might have guessed in talking about rankings with our clients: the average user does not know that their Google SERPs are personalized.

To tell the truth, we SEOs also tend to forget that search is almost always personalized, and we examine concepts such as, for example, neutral search.

For example, we tend to act this way when we try to understand the rankings of our sites or when we do competitive analyses. It is certainly not incorrect—it is a necessary starting point—but in reality, it is not enough anymore.

Take the case where our site is national or global: In that case, the personalization of the search experience is such that we should not only check how our site ranks in the U.S. or the UK, but we should also in smaller geographic areas of our targeted country.

At the same time, we should see who our competitors are with a "micro-geographic" focus. In fact, while we might be on the first page in a totally neutral search with its geographical center being the political capital of the country we are analyzing, maybe we don't rank so highly in the searches done in a city that we consider a target as important as the "nation" (i.e. Seattle or Manchester).

Why? Because the often shamefully forgotten Venice Update enhances the localization of the user performing a search in terms of how their SERPs are shaped. Hence, local businesses, which might not be relevant on a national/global scale, are indeed relevant locally. In those cases, they can be shown at the expense of "national" or "global" sites, which often do not possess sufficient relevance at a local level.

And that's personalization (note: in the concept of personalization I personally include context, because without it, personalization would provide a poor search experience).

But that's not the only way localization influences the personalization of search.

In fact, as both Tom Anthony and Will Critchlow explained well, localization (and other contextual information) is a key component of what they defined as "new queries," which include both a explicit and implicit aspect.

An even stronger implementation of personalization is possible: implicit-only queries, as they are defined by Baris Gultekin in this video interview shot at Google I/O 2013.

These queries are those that users don't even actually perform, but that Google predicts they are implicitly performing. The results are shown in Google Now cards:

In the first case (personalization due to geolocalization), we can try to acquire more relevance on a local level by creating events (online and/or offline), connections with local web sites, and partnerships with local influencers. Those influencers can be found with tools that geographically map social media followers/fans, such as Followerwonk (all the better if they are already connected with us):

Or, we can take advantage of the geographical segmentation of the people we have circled on Google Plus (and of the local communities' pages, if they exist):

In the second case ("new queries" with implicit and explicit aspects), we can try to "enter" in the personalized SERPs of our users, creating content that is contextually relevant to a topic + location + device. For now, though, it is quite hard to determine how, from where, and for what a user is already searching on our own sites via Google search. This information can't be easily understood with tools like Google Analytics, and Google Webmaster Tools does not offer us the opportunity to dig deeper than the country level. Hence, the best way to get this information is by actively obtaining feedback directly from our targeted audience.

In the third case (totally implicit queries), we can go with the classic SEO's first reaction of fright and ask to have our site integrated in the Google Now ecosystem, as Zillow, Booking, Urbanspoon and many others have already done.

Personalization and Knowledge Base

Last May, at Google I/O 2013, Amit Singhal said, "The search of future will need to answer, converse, and anticipate."

With "answer," he refers to the Knowledge Graph, with "converse" to Voice Search, and finally with "anticipate" to Google Now. Knowledge Graph and Google Now are based mostly on the so-called Google Knowledge Base, and in both cases—as well as in Voice Search—semantics and entity recognition play an essential role.

Semantics, entity recognition and the Knowledge Base, then, are the foundation on which Google can really achieve the goal of creating its dreamed-of Star Trek computer, capable of providing information to the user by predicting its needs for information.

As I wrote in a previous post here on Moz, the Knowledge Base helps Google by answering how and why the documents are connected and searched, as well as an understanding of what named entities those same documents cite and are related to.

The most evident examples of this are the Knowledge Graph boxes:

This snapshot, though, shows another example of personalization.

Google presented me Saint Peter the Apostle because in a neutral search I performed before, Google agnostically presented me all the entities the Knowledge Graph could relate to the query "Saint Peter".

As you can see, neutral "objective" searches still play a huge role in Google... but is this really so? No, it isn't.

Even in a neutral search, personalization of search is present. Here are a couple of examples:

Knowledge Graph disambiguation boxes in Google.it neutral search for "San Pietro"

Knowledge Graph disambiguation boxes in Google.com neutral search for "San Pietro"

Knowledge Graph disambiguation boxes in Google.fr neutral search for "Saint Pierre"

Knowledge Graph disambiguation boxes in Google.com neutral search for "Saint Pierre"

Localization of the users—both geographically and linguistically—plays an evident role in the personalization of search.

But that's not all. In fact—as I said before—personalization is always acting, not just when users are logged in. When you're not signed in, Google uses a cookie to personalize your search experience based on past search information linked to your browser.

The more someone uses Google for search, even logged out, the more Google understands and refines the search experience for that user. Knowing that there are about 5,134,000,000 searches performed every day, we can understand how the Google Knowledge Base is endlessly updating itself. That is not Big Data, that's Gigantic Data, all used for one purpose: to offer more personalized search and ad results.

How does Google personalize search?

Search History is surely the most important factor, but as we saw, localization has assumed an increasing relevance, especially because of the rise of mobile search.

Google seriously knows a lot about us. Crazypants! as a friend of mine would say.

How does search history shape the personalized SERPs, and how can Google strengthen the personalization of SERPs in relation to a query when search history is not present or is not sufficient by itself?

Google does this thanks to search entities, a concept that is explained in depth by Bill Slawski in this post.

Search entities, as described by Bill, are:

• A query a searcher submits
• Documents responsive to the query
• The search session during which the searcher submits the query
• The time at which the query is submitted
• Advertisements presented in response to the query
• Anchor text in a link in a document
• The domain associated with a document

The relationships between these search entities can create a "Probability Score," which may determine if a web document is shown in a determined SERP or not.

I warmly suggest you read Bill's post to find out more about all the possible relationships that can exist between these search entities, but for this post, I'd like to focus on these ones:

1. The strength of relationships between these entities can be measured using a metric obtained from direct relationship strengths (derived from data indicating user behavior, such as user search history data) and indirect relationship strengths (derived from the direct relationship strengths).
2. A relationship between a first entity that has insufficient support (e.g., not enough search history data) to associate a given property with the first entity and a second entity that does have sufficient support to associate the given property with the second entity can be identified, and the given property can be associated with the first entity with higher confidence.

From an SEO point of view, these two cases are telling us that even though we aim for a neutral search environment, we should never forget that 99% percent of a user's search experience is personalized. We could define this attitude as "growth hacking SEO."

Moreover, we could take advantage of the personalization of search thanks not only to being included in the personal search history of the users, but also to connections created with entities that are already in those users' search history. This connection can be a link, a citation, or a co-occurrence in a document, which is considered more relevant than the query alone or the search history of the users.

Somehow this is not something new. In fact, when Richard Baxter talks about doing really targeted outreach, we know it is good from the point of view of being discovered by the audience. Creating content for other sites that are used by the people influencing our target market will often result in new users of our own site.

But now, this patent about search entities is evidence that typically inbound tactics can have a direct reflection on a purely search-related level.

Semantic web

When we talk about entities, we usually think about people, places, and things (i.e., a brand). But web documents are also entities.

And, in light of what is described in the patent cited above, the "probability score" of a web document, which can determine its presence in a SERP or its visibility in results for a determined query based on all the classic on-page "ranking factors," can be improved by the use of structured data.

Structured data, from schema.org, Microdata and Open Graph, are important not just because they can gift our site's search results with a rich snippet. That snippet is the facade of something more important: helping the search engines better understand what a document is all about.

For instance, the breadcrumb schema is surely important because it can help add mini-sitelinks to our snippets, but it is even more important because it clearly tells search engines how the documents in our site are hierarchically related between them.

Or, using an even better example, the article schema is the only way (or at least so it is described by Google) to obtain visibility in the In-Depth Articles search blend.

Therefore, the use of structured data has become essential, not only because rich snippets offer us a greater visibility in the SERPs, but also because not many people are using it (36.9% of URLs use Open Graph, and 9.9% use Schema.org, as reported by Matthew Brown at MozCon). In addition, structured data can help increase the relevance of a document for a determined query simply because it "helps our systems to better understand your website’s content, and improves the chances of it appearing in this new set of search results."

The social layer