David Amerland

How Search Engines Discover Who You Are in the Semantic Web (and why this matters to your business)

Identity and Reputation in a Semantic Web
Verifying your identity is key in the semantic web. When the indexing of information leads to inferential judgments that produce answers right on the search results pages (SERPs) Veracity (the 4th of the 4Vs of semantic search) is predicated on identity.

Simply put how can you trust whether a piece of information is true or not when you are uncertain of its origins (which then means you are uncertain of the intent behind its dissemination)?

Semantic search’s entire success depended on search engines being able to identify the originator of a piece of information and then assess their profile with an algorithmically determined trust score. This is not new, of course. In the pre-semantic search days the link graph signature of a website could be used as a trust-assessment system (a fact consolidated in Yahoo’s patented TrustRank the methodologyof which is detailed in a research paper) to help weed out spam from the search results pages.

Trust however continues to be a huge requirement on the web. It featured centrally in Google’s leaked Human Raters Guidelines and both Eric Enge and Bill Slawski have talked about in in detail in relation to ranking in search. 

Identity, Real Names and Trust

Google+ (which Eric Schmidt described as a digital identity service) kicked-off Google’s efforts to map real profiles and understand the interactions between them and became core to some of semantic search’s initial scaling.

Now, just because someone is willing to put a name and a picture up on the web does not mean that immediately all barriers come down and trust is instantly offered to them. But it does go some way towards creating a sense of investment in one’s presence. Trust is always a function of risk assessment. A profile that’s been built up over time through activity and engagement has a reputational value that is hard to get and easy to lose, so its owner is unlikely to engage in activity that destroys it.

This is one of the reasons that the advent of semantic search became such a game-changer for SEO and marketing. When the shortcuts make no sense in terms of providing any time and effort gains, then you may as well be real and work hard to deliver value as the only viable means to gain.

The question frequently asked is why would a real name make a difference in the first place? The reason is mostly psychological. You only have one real name whereas pseudonyms can be create ad infinitum so in the very first instance it creates a perception that makes for fewer deceptive practices. The whole idea behind the Real Names Policy on Google+ and a little later on Google+ Authorship were predicated on this.

Naturally, as many on Google+ know already, lack of a real name does not immediately reduce trust. The creation trustworthiness is based upon the four-step process of:

• Contact

• Perception

• Assessment

• Connection

It seems reasonable to suggest then that even pseudonymic profiles can gain the trust of others over time provided an investment of time and effort is made in their reputation.

When Google relaxed the Real Name Policy and finally withdrew support of Google+ authorship from search the fact that it could now follow and map personal profiles across the web as well as its own network was one of the reasons behind the decision.

Reconciling Identities Across Different Social Networks

For Google this is important for several reasons: First, verification of data. This is a big deal. Google indexes the entire world (or wants to according to its mission statement). There is little point in having the “world’s information” in your index if that index cannot then be trusted to give trustworthy replies to search queries. Second, scaling. The indexing of information goes on all the time. As information multiplies the more trusted sources Google can find the faster it can validate it which means the faster it can respond accurately to search queries. Trusted sources are important to scaling. Third, spam (and perhaps trolling as a side issue). Both of these activities pollute the web and sour users’ experience of it. Being able to filter out these two from search allows Google to create a better end user experience.

The principle behind the reconciling of user identities across the web is a simple one. When we establish a social network we usually do so as a specialized subset of our real world ones. We, for instance, create a subset of professional connections at LinkedIn, a subset of familial ones on Facebook and a subset of special interests or hobbies on Google+. In addition there are other fragments of ourselves found in communities, interactions with posts and other content across the web.

How Google Discovers Who You Are on the Web

Google can use a variety of techniques that allow it to index identities across social networks on the web:

Influencers – Those who follow an influencer in one social network also follow them in another, even if they use a different name to do so.

Domain expertise – Your interests, knowledge, likes and dislikes inevitably seep into whatever online activity you get into.

Login details – Many of us use one or two emails addresses when we create profiles in social networks.

Your contacts – Few people are truly capable of keeping digital profiles that have no overlap in their contacts.

Your network connections – Inevitably the people we choose to follow and interact with provide an overlap that can begin to identify us.

Google+ - Google’s own social network is a great springboard to start off from.

Personal data – The personal data we choose to make public in order to experience a better service is difficult to erase as we switch between social media profiles and (even) identities.

Your location – Location and IP addresses provide additional data.

Your sentiment graph – What we say, to whom, when and how. What is said to us, by whom, when and how, now all become subject to analysis.

Your interest graph – What we engage with in terms of interests, content, websites, and subject matter (and the people we follow as a result).

Your devices – Many devices send specific profile information to Google, Apple, Facebook and Microsoft.

Your Comment Graph – The sentiment expressed in our comments, writing style and comment style as well as more granular signals like the time of day we engage or reply, who we choose to engage with or reply to, the subject matter that gets us going, the length of our replies and the further responses they generate, can become part of a very unique signature that expresses us.

Offline/Online Mapping – The gap between the online and offline world is getting smaller. As Entities are being created it becomes harder and harder to create an online profile that is, essentially and ‘orphan’ existing independently of anything else and find that it is effective in doing anything.

What This Means for Individuals

The semantic web is inescapable. It is scaling fast and it is joining the proverbial dots between online and offline. This will make it harder to take to the web with an intent to vent, bully, stalk or harass people as the repercussions from such activities will potentially be a lot more severe than just the closing of an account.

In a recent Google+ post that’s worth revisiting I noted how finance companies will be using the social web to work out whether to lend money to, or not, to people. As semantic technologies take hold and our online profile becomes part of an algorithm that’s determining something about us (our eligibility for a job, the next blind date we ought to go to, how desirable we are to an organization) we will find that it makes sense to invest in its built up the same way that we invest in the way we build our reputation and personal network offline.

This ties online activity to offline human economic behavior in a way that makes sense for everyone.

Is this the end of Trolling?

Anonymity across the web has frequently allowed unwarranted behavior that has placed emotional and psychological pressure on individuals that were on the receiving end of it. Will the ability to map profiles allow the pasting of a “Troll” label across them?

Even if that were possible it may not be desirable. By degrees the web is turning into a transparent, interlinked community. Tarring and feathering is not really the best way to move forward. If they find themselves being increasingly isolated, their trolling activity returning no intrinsic rewards trolls too will either desist or “get with the program”.

What It Means for Marketers

The web is still about connection which means discovery and search. But search is changing. It is now driven by context and answers. Entities are important because they become trusted points. Marketers who are serious about succeeding on the web should:

• Have interlinked social media profiles.

• Try to achieve as great a data density as possible (place detailed information about what you do and why where your audience is most likely to find it)

• Tie every detail into your unique way of doing things (in other words develop a distinct online identity the same way you develop a distinct offline one).

• Create valuable content regularly in a way that establishes a real connection with your audience.

• Do not market. I know this sounds ridiculous but marketing is all about actions that have to be done without really asking “why?”. Spend your time, instead, providing to your audience the things they really want to know.

• Use more than one social network. I know time is precious. Different social network platforms have different audiences and unless you’re doing something that’s incredibly focused and narrow (like brain surgery or rocket science) then the chances are that you will find your audience is scattered all over the place. Make it easy for them to find you and connect with you by having a diverse, widely-cast online presence.

• Bring online and offline closer together. Make sure you have QR codes in your offline presence. Use Google+ for business for your business listing. Update your online presence in real time to match your offline activities.

• Create a brand. Have a mission statement, clearly communicated brand values and a methodology to achieve transparency, in place.

• Create relationships. When just one bad transaction can cost you thousands or even millions in reputational damage then it’s worth investing a lot of time and effort in creating a good experience every single time.

• Be personable. If you’re a business this is hard to do. Yet you really do need to find a ‘voice’ and develop a ‘style’ much as you would if you were a person.

• Be reasonable. We talk about creating exceptional content but exceptional content really is the kind of content that delivers value to its audience when they need it most.

• Be precise. On the web it’s easy to try to be all things to all people. That will only serve to diffuse your marketing efforts and dilute your sense of identity. Be what you really want to be and trust that it will help you connect with those who are your audience.

Semantic search is scaling fast and the semantic web is being created as we speak. Acting sooner rather than later will help because there really are no shortcuts to building a profile and a reputation, any more.

Posts That Expand This One

How to be Exceptional in your Marketing
Marketing is All About Personality
Google’s Knowledge Vault
Sentiment Analysis in Semantic Search

 

Sources

An efficient reconciliation algorithm for social networks (Google Research Paper)
Navigation in a small world. Macmillan Magazine.
Friendship and mobility: User movement in location-based social networks (Stanford Research Paper)
Bridging the gap between physical location and online social networks. (Carnegie Mellon University Research Paper)

 

© 2019 David Amerland. All rights reserved