November 11, 2018 Harold Tor

Google Search is 20 years old: What to expect next and what Google still does not get

Recently, Google celebrated 20 years of Google Search, and I attended its anniversary event in Brussels. The event was jointly organised with Edima, ESBA, the Kingdom of the Netherlands and Allied for Startups.

What was not advertised about the event but of huge interest to SEO experts around the world, was the presentation by Pandu Nayak, Google Fellow and Vice President of Search. As part of Google’s core Search leadership team, Pandu Nayak is in charge of ranking.

Mr Nayak gave a peak of what is to be expected of Google Search and search ranking that will ultimately change much of what we now do for search engine optimisation.

Before proceeding to introduce the much-awaited future changes, Pandu Nayak first qualified Google’s approach. They had conducted 228,022 quality tests, 31,584 side-by-side experiments, 11,365 live traffic experiments, and 2453 (test) launches. How much of this was done only in English or in the United States is not clear. The US-centred approach of Google, as I’ve explained in the past, is the reason why it cannot penetrate certain markets and in Europe, search quality is often poorer universally other than in English.

Harold Tor - dontthinktoomuch.com - Google Search is 20 years old #SearchIs20Harold Tor - dontthinktoomuch.com - Google Search is 20 years old #SearchIs20
1. Google AI, Topic and the Knowledge Graph

Google acknowledges that the same words often mean different things in different contexts, and that is often the reason for poor search quality based solely on keywords.

Pandu Nayak gave the example of the word “change“, which can mean: adjust, convert, exchange, install, modify, replace, or switch.

To circumvent the problem, Google came up with an artificial intelligence system that studies the context in which such keywords are included, called Google AI.

What Google AI does is to continuously group keywords of similar context together into one group, and that group is a layer of dimension added to the search algorithm called Topic.

For example, if the word “change” keeps appearing with the words “brightness” and “screen“, then pages containing all three words will be added to the topic layer that focuses on “changing the brightness of one’s screen“.

Grouping pages into topics will greatly assist Google Search in contextualising the search results.

 

What does this mean for the future of SEO?

Recognising that topics are as vital for your SEO as choice and placement of keywords within your content, you should now focus on the context for which your content is written.

What do I mean by that?

For example, if you have a page on “brown leather sofa“. Try to think of the specific situations in which one would search for this content: for purchase, maintenance, repair, what styles, how to tell fake and real?

If your content focuses only on the maintenance of brown leather sofas, then think of inserting into your content the words that provide the right context so that Google AI can group your page with all the others that write specifically on that:

  • Brown leather sofas are not that difficult to upkeep
  • You need a little patience when it comes to the maintenance of brown leather sofas
  • To keep your brown leather sofa looking as new as when it was first bought…
  • To maintain the look and shine of your brown leather sofa…

In short, contextualise your content.

Harold Tor - dontthinktoomuch.com - Google Search is 20 years old #SearchIs20Harold Tor - dontthinktoomuch.com - Google Search is 20 years old #SearchIs20Harold Tor - dontthinktoomuch.com - Google Search is 20 years old #SearchIs20

2. The Activity Cards, Evergreen vs New content

Google AI and the Topics have another function: to feed content to the upcoming Activity Cards.

Pandu Nayak presented this new feature that is supposed to enhance one’s search experience. Many users conducted searches but left them sometimes without finding satisfactory results. Perhaps new content that appears will answer your needs a few days later. Or, if you are interested in “bonsai cultivation in Sri Lanka“, it will save you time conducting this search all the time, because new content will be delivered to you once they have been ranked. As a personalised feature, your recent and frequent searches are now grouped by “Topics“.

If you go onto your Google Search or Google Home, these cards will appear showing you the latest content for each card.
If this works, it will enhance the user experience for both abandoned searches and frequent searches.

What does this mean for the future of SEO?

The focus of the Activity Cards is certainly the newness of the content.

Does that mean only completely new content gets featured in the activity cards? As we know, the crawling is done based on the frequency of updates to your content. The more updates you have, the more often it gets crawled. Yet even if your content gets crawled, will updated or renewed content appear in the Activity Cards?

The answer provided by Pandu Nayak was not clear nor satisfactory.
He said for Google, they have a special status for “Evergreen Content“. Evergreen content are pages that have been visited and therefore validated by millions of users, and they will retain their high ranking in searches. While these will not get featured in the Activity Cards, they do not get penalised by getting “ousted” by new content.

While I accept that this will promote the visibility of new and young content, most of the content out there are in the middle: neither owning that “Evergreen” status nor are freshly created.

To circumvent this issue, I suggest that we conduct periodic reviews of our pages and try to refresh and update the content based on the latest developments. If the change is drastic, why not change the title and URL so as to brand it as a new content?

Harold Tor - dontthinktoomuch.com - Google Search is 20 years old #SearchIs20Harold Tor - dontthinktoomuch.com - Google Search is 20 years old #SearchIs20Harold Tor - dontthinktoomuch.com - Google Search is 20 years old #SearchIs20
3. Search by image

According to Pandu Nayak, Google Search will focus a lot more on search by images.
Apparently more and more users prefer to use image rather than text results. This may most likely due to the majority of searches was conducted on consumer goods.

Of course, this calls into question the quality and ownership of images, and the context in which the images are placed on the pages.

Because the event took place in Brussels and in the EU district, Pandu Nayak was quick to point out that Google supports creators and will insist on displaying and prioritising copyrighted images that are properly accredited. This, according to him, promotes creation and creators.

What does this mean for the future of SEO?

It is already a best practise amongst SEO experts that you HAVE to place appropriate images on your page and at the same time, provide keywords and description and other meta data to your image. Now, with Google AI, you will have to go one step further: ensure the images fit the context – the topic layer – in which your content is situated.

Back to the “brown leather sofa” example: if you have a page that focus on the maintenance of brown leather sofas, you should not place just any image and fill in the meta tags with “the maintenance of brown leather sofas“. It should in increasing priority be an image of:

  • leather sofa
  • brown leather sofa
  • a hand cleaning a brown leather sofa
  • a hand and a product cleaning a brown leather sofa

Most importantly, make sure that it is your owned media.
Images purchased from image banks tend to be too generic and will have been replicated elsewhere. Use your own images and ensure the copyright information is there on your page.

Harold Tor - dontthinktoomuch.com - Google Search is 20 years old #SearchIs20Harold Tor - dontthinktoomuch.com - Google Search is 20 years old #SearchIs20
Other problems I have with the new Google Search

All this sounds like Google is continuously improving their algorithm and product. That’s good news.
But there are always challenges when one tries to use technology to categorise, imitate or analyse human communications.

The new Google Search is plunging us into the hell of context

As Pandu Nayak explained, Google AI and the topic layer is an attempt to contextualise content so as to match the context of the search.

Pages with both their textual and image content are now linked to topics. As time goes on, Google AI will grow. The end result being much of the internet as we now see will be pigeonholed into millions and millions of contexts. The depth of the context will only go deeper.

It very much goes against the nature of the internet as it is now when users are free to decide the depth of their searches. They are free to lightly explore, or fervently research. By creating a huge machine to identify context, I’m afraid Google is refining search so much that it is taking away the right of the user to make that decision about how deep they want to search.

I highly suspect the “228,022 quality tests” Google has conducted were very much focused on consumer goods or “things” in general.

While it is a very basic notion, is it right that Google has decided for us that most people are placing it as their number one priority is the way we want the internet to grow?

Yet if we look at how we use the internet outside of the world of Google Search, it definitely goes beyond that.
Google helps us answer the “whats” and the “hows” of our lives. But can it help us answer the “whys“?

With the new Google AI, I get the feeling this is not their priority.

The new Google Search got Context all wrong

Do you still remember the time when we used to go to the library?
It was a wonderful experience. You have the different sections, archaeology, history, art, languages, engineering, literature etc. There is always a corner with huge sofas where one can pick a book and immerse in it for hours. There are also desks with very concentrated lightbulbs, for people who have a pile of reading to do and who are always scribbling away. There are times when I would pick up a magazine, just very quickly browsing through the pictures. Other times, I just wanted the latest news – politics, culture, cinema, art, and they always have the best newspapers.

Despite we now have the internet, humans are still humans.

Our need for knowledge, the situations in which we want to acquire the knowledge, the context of our desires, the ways we want to eschew that knowledge are all very ad hoc, emotionally tied, culturally and socially influenced decisions and events.

Google Search only sees it one way, that people want “content“.
Things” are not “content“, just as “content” is not knowledge.
One glaring fact that Google Search does not understand human aspect of search, is the lack of distinction between news and new content.

I posed my question to Pandu Nayak but he did not answer my question at all.
Google shows new content (newly-crawled) in the Activity Card. These, to Google, are “news” that cater to your interests that they got to know via the topic layers you searched on before.

By locking me in the contextual abyss of the topics I have searched before, Google is showing me an extremely narrow view of the world.

To take the example of the library, there are times when I want to find out more about “cars“.
While the context of my quest for knowledge is important, the history, the mechanics, the different models etc (which Google Search is trying to replicate), what is more important is there are times when I want to go to the newspaper section to read about all the latest news on all topics possible. Yet Google has me locked in the sections and refuse to let me out.

 

In my opinion, Google has lost its way.
The original neutral, no-nonsense search engine created by a group of normal guys has given way to a huge tech conglomerate that thinks it can impose its products on everyone without even understanding the human aspects of our behaviour, including our desire to know and understand.

, ,
%d bloggers like this: