David Amerland

How Semantic Search is Changing End-User Behavior

Google Semantic Search

Search and psychology have always walked side by side. This has not always been obvious. But it’s a truth. Because search is the only viable window we have for looking out at the web and its massive amount of data every tiny change that happens at the Google search box in time produces a change in human behavior.

When Google introduced Google Instant for example it also, quite intentionally, created an “instant feedback” mechanism that enabled the end user to “formulate a better search term”, this in itself had a remarkable and far-reaching impact on SEO and online marketing with website content now being guided, to a degree, by the search terms revealed from the drop down suggestions, right at the Google search box. 


As the Google video shows below this resonated with end-users who ended up “adapting their search on the fly”.





Before semantic search came along this is how search used to work:

1. We’d try and put in search a keyword we thought would most likely lead us to the results we wanted.

2. The person whose web page perhaps best answered the search query we wanted to look for would also think in terms of keywords, trying to guess how we would look for what they had in the page. In order to best approach this they would also do their keyword-research that would reveal to them two things: A. The number of people using that particular keyword B. Peripheral keywords associated with the subject and the original keyword.

Here we have a classic case of second-guessing on all fronts. When we are looking for something and want to find it fast the last thing we want to do is go through a series of searches where we try to understand what we don’t know and find the descriptive term that best fits what we are looking to find, yet this is what we used to say find ourselves doing. At the information or service provider end there was an equal amount of second guessing taking place. Those who sold something as simple as beer, for instance, would also have to try and optimize their websites for ale, lager, bitter, stout, brew and the less than obvious “amber brew” all because they were trying to guess the terms we would use when looking for them. And all that before I add the complexity of a fresh layer of synonyms activated by the keyword “pub”.

That was a losing game, designed to appeal to those who love scramble perhaps, but hardly fit for the information age.

I say “was” because it is not done like this any more. Google processes 100 billion queries a month and on any given day 15% of that, some 500 million have never been seen before. When you consider that the traditional way of optimizing a website to show up in search required keywords, you can instantly see that there is a gaping hole in the methodology that opened 500 million times a day and swallowed up a search without a trace.

The semantic web is different and semantic search works differently within it:

Step 1. On the end user side, semantic search works hard, pulling in personalized data from various points that include location, time of the day, past search history, search patterns and social activity to understand the intent behind the query.

Step 2. On the website side Google indexes the page, uses latent semantic indexing (LSI) to understand the connections between words, extracts document metadata so it can begin the process of entity extraction (a requirement for proper semantic indexing) and then cross-references the web page with all the connections, mentions, social activity and citations it can find on the web.

Step 3. At the Google search box, Google now matches its understanding of the query (step 1) with its understanding of the web page (step 2) to present the best possible match regardless of keywords used.

This process seems better suited to an age where time is precious and context and relevancy are key. The effect it is having on end users is remarkable, particularly where Voice Search (a totally semantic search vertical) and Google Now (a semantic predictive search vertical) are concerned. By freeing us of the double-guessing that takes place within the search process, it allows webmasters and those who use search to bypass the double-guessing game and focus on what they each need to do.

Webmasters have to make sure their websites present things in the best way possible so that Google search understands who they are, what they do and how they do it. Search users simply need to ask for that they are looking for.

Is the process perfect right this minute? No, it isn’t. We are only six months in the semantic web world and Google already is making gain in it with each passing day. The fact remains that the web is changing, search has changed and the way we operate as individuals, as well as marketers, has changed with it.

The takeaway for marketers and webmasters here is that within the semantic web the focus is on creating

  • Content that answers specific search queries as opposed to content that surfaces thanks to keywords.
  • A valuable end-user experience on their websites.
  • A strong social presence for their websites.
  • Brand behavior when it comes to how they use the web.

 

What You Missed

Web 3.0 and Semantic Search
Most Common Questions On Semantic Search Answered
The Concept of Symmetry in a Social Network Setting
Semantics, Context and Personalisation
Google Semantic Search (book)


External Links

Google Search Baffled 500 million times a day

© 2017 David Amerland. All rights reserved