Google's use of AI in search changes everything.

Search is a Critical Component of Artificial Intelligence

When everything in the world is data, navigating the data successfully requires technology that starts with and is driven by search. As search has become more and more sophisticated its indexing has become more complicated, its categorization of information ever more deductive. Its understanding of the importance of what it has indexed bounded, as well as defined, by knowledge. 

In many ways, search has become ubiquitous and invisible. When Google Now lets me know that my “…drive home is 39 minutes” from where I am at a meeting it has taken into account where my car is (by looking at the GPS signal of my phone), the traffic conditions on the way (by using Waze to cross-check information in traffic conditions and possible delays), my own travel patterns (by checking the history of my travelling to and from specific meetings with clients and what I do afterwards). It has gleaned information which has created abstract layers of habits and intent (much like a human would) and has made a recommendation I am likely to find of value. 

Knowledge, once collected, analyzed and understood, frequently gives rise to metadata that is abstract and which connects real-life, tangible events with the more ethereal needs and intent of a person. This is why a murder (a tangible, real-life event) cannot reasonably have a suspect who does not have means, motive and opportunity (or in search terms: access, context and intent). These are the basic requirements for connecting what’s inside a person’s head with what he does in the world we see and touch (and search now plays a key role even in murder convictions). 

Search, Knowledge, Disambiguation and Natural Language Processing

The one definitive signal that search is evolving lies in its processing of natural language queries. In plain parlance search has changed from a tool which had its own language that we needed to learn in order to use it better to a tool that works hard to understand us so it serves us better. 

The onus of quality results has changed from us to search itself. Paradoxically this has also impacted hugely upon us by forcing us to focus on quality content, semantic density  in the way we serve the information we create and real relationship-building engagement practices. 

Google has always been ahead of the pack in disambiguation (the ‘little’ thing that tripped up Facebook when it came to search) and this has relied in understanding context which requires not just data, but knowledge. 

Despite the company having morphed to Alphabet and cast its web ambitiously wide to include robotics and stratospheric orbiting vehicles, search is still being evolved. Natural Language Processing requires not just understanding of the language that is being looked at (and in a semantic search world Google has to build that capability anew for each language) but also the context within which it is spoken. That last bit is tricky. 

Google’s announcement of SyntaxNet self-labelled “the world’s most accurate parser” is important for two very specific reasons: 

  • First, it showcases the approach Google has always used in search. It uses human subjects to train and test its algorithms and when those algorithms achieve a high degree of confidence in the results, it scales them.
  • Second, it shows that in order to get the meaningful mountains of data Google needs it has established an open-source, sharing approach whereby AI products are released to the wider web community to use. 

Some of the implications of this approach are documented in an MIT Technology Review piece where it suggests that a generation of intelligent (and therefore useful) chatbots may be just around the corner. The bottom line is that while it appears that search is making more and more use of artificial intelligence components it is actually breakthroughs in search that are making new developments in artificial intelligence possible.

So, … yes, Marketers

Geeky search enthusiasm over, the million dollar question, quite literally for some, has to lie in “so what?” what does this change and what should website owners and marketers do differently? For a start the ability of Google to understand and parse natural speech shifts the needle of perceived value for content to the content creator. Either you create content that truly resonates at a human level not because it is somehow ‘brilliant’ but because it really is valuable in terms of answering the “so what?” question on its existence, or you are simply creating content in order to be noticed which will work less and less and less as we move forward. 

Then there is the question of data density and identity. Is the content you are creating comprehensive in the way it references other content, expands its value through linking out to authoritative sources (and if I hear the argument about leaking page authority because of linking out, once more, I will scream), have you included rich media links and resources. Have you made it, in other words, compelling? Does it reflect the values you stand for? Is it part of the way you work? Is there a USP (unique selling proposition) reflected in its creation, setting or context? 

I mentioned data density and identity in SEO Help: 20 Semantic Search Steps that Will Help Your Business Grow so there should be no surprises there. Nevertheless the needle on actioning these now shifts from “sensible to have because of semantic search” to “vital because of the way Google search parses information”.  

In addition to the structural changes search is undergoing we now have: 

  • Mobile search in all its different forms (tablets, phones and wearables) is of greater importance and prevalence than desktop search.
  • Voice search is changing the traditional relationship between searcher and search to one where search is a built-in function of a device and Voice becomes the user/device interface.
  • Query structure (and nature) is changing radically because of expanded use which leads to a much broader field of contexts. As a result search is turning to automated programming (a.k.a. machine learning) and applications of AI (like Rankbrain)
  • The linear relationship between search and website visibility now forks out to include search and social media, email, direct sharing, peer-to-peer sharing, communities and groups.
  • Traditional SEO has expanded to include branding, marketing, aftersales care, customer management, customer acquisition and social enterprise initiatives.
  • The rise of predictive search based upon an ever more refined awareness of  context and intent that’s expressed as a particular service through a device (Think Google’s Inbox, Waze, Google Maps, YouTube, Google Now – the list goes on).
  • The coming of augmented reality (which is already here in many early forms) and the rise of virtual reality. 

The list marketers and website owners must now work through goes something like this: 

  • Analyze your audience, make sure you really understand what it is that makes the people you reach out to, tick. Know what’s important to them and what isn’t. Remember they are people who matter not numbers for your marketing to reach so it can look pretty on a spreadsheet.
  • Develop a true value proposition in your engagement across your website, marketing, sales message, customer contact and social media presence. Do not just create a thinly veiled “Here we are, buy stuff from us” presence.
  • Be interested in who you do business with. Develop relationships that allow you to understand the impact of your business and services rather than the pricepoint dynamic where you carefully try to balance the psychological triggers you use in your marketing with the amount of money you ask.
  • Develop your online identity as fully as you develop your offline one. Work to earn the trust of those you do business with and build your reputation.
  • Unify your SEO, marketing, branding, sales and promotional efforts to achieve efficiencies that also allow you to drive a consistent message across all your business activities.
  • Use real knowledge in your content and employ context to clearly disambiguate its meaning and relevance. 

It’s far from exhaustive as a list but it looks at some of the key areas your efforts should be focused on. Many of these have been covered by me many times before so it should not be surprising that they are now becoming critical to your online success.   

   

© 2017 David Amerland. All rights reserved