Conversation about Semantic SEO

Semantic SEO

Everything started with my thoughts on a article written by Damian Thompson and called

What The Heck Is Semantic SEO & Should You Care At All? 

At the end of the day it doesn’t matter whether you call it Semantic SEO or just SEO. Adapting to the semantic search topography will be part of every SEO’s job description going forward. Like it or not.

But, a caveat: semantic web is NOT the same as Semantic SEO –http://goo.gl/B1N9Ww

Top tools to help you with your semantic SEO efforts:

Übersuggest – www.ubersuggest.org
CISE – Conceptual Intelligence Search Engine CISE – http://goo.gl/BdaUo8 – by Peter Hatherley

PS Did a bit of googling and found a good definition of the term ‘Semantic SEO”  by Mashable and Erin Everhart:

“Semantic search uses artificial intelligence in order to understand the searcher’s intent and the meaning of the query rather than parsing through keywords like a dictionary. When you search now, Google gives you results based solely on the text and the keywords that you put in that search. Essentially, Google gives you its best guess.” – http://goo.gl/MB8Mnv

In simple words Semantic SEO is dealing with search results that do not necessarily contain the keywords in your copy. But trust me, it’s not just a guessing game.

Then a conversation started.

Aaron Bradley: Omi SidoPS Did a bit of googling and found a good definition of the term ‘Semantic SEO”…

Actually what you provided was a definition of semantic search – a pretty good one, but one that doesn’t actually address semantic SEO which would (a la the “O” in “SEO”) give a nod to optimizing for semantic search.

So what is it?

I’ll have a go (culled from http://bit.ly/1aLpw09), and invite others to weigh in!

Semantic SEO is optimizing for the entities and entity relationships that search engines identify in queries and return in results.

Omi Sido: Aaron Bradley, you are absolutely right and your definition (denotation) is spot on. I can’t even believe you wrote this article in 2013 when not many people were talking about semantic search and #seo .

On the other side I still find the definition given by Erin Everhart (I was so sure somebody will challenge me on this one (lol) as neither the article, nor this exact quote are actually talking about Semantic SEO) a good base for explaining what Semantic SEO is, especially to people just getting into the semantic world.

And last but not least this is one of the articles Google and Bing (mobile and desktop) are showing on page one 🙂 when asking the question: ‘what is semantic seo’ – https://goo.gl/WGVxFi, https://goo.gl/MGFmNq.
Bad google. Still a lot of room for improvements 🙂

Aaron Bradley: Thanks a lot Omi Sido – and I agree that Erin’s thumbnail description is an excellent one.

For what it’s worth, my go-to definition of semantic search (which I quote at the end of http://bit.ly/semsemtechbiz2013) is from Tamas Doszkocs:

“Semantic search is a search or a question or an action that produces meaningful results, even when the retrieved items contain none of the query terms, or the search involves no query text at all.”

Teodora Petkova: Omi Sido thanks for igniting this discussion. Aaron Bradley thanks for the definition (again these relationships… :-))

Raja Rethinam: Omi Sido, this is a wonderful introduction to semantic SEO. Never thought about it, but now I’m gonna pay more attention!
Love the picture 🙂
Aaron Bradley your definition is an excellent one.

Omi Sido: PS I’ve just noticed that many people are using the same definition when trying to explain Semantic SEO (hmm I thought I was unique 🙂 lol )
One probability would be:  Google  is displaying this article on page one for this query.

Is Google trying to tell us something (bemuse us) or are we confusing the search engine (David Amerland, Aaron Bradley)?

Is it possible that in the future we may have wrong answers given by Google just because people like one thing more over another 🙂


Semantic SEO

David Amerland: Omi Sido, I have started defining some terms (still a work in progress): http://goo.gl/zb8Wl6 and the Veracity question of search is key to the trustworthiness of the results and the inherent difficulty in manipulating them. Within semantic search there is a weighing algorithm that ascribes veracity values through citations (as opposed to mentions) and authority. So, to break this down a bit, if a lot of us say the same thing (i.e. because it is popular) and only some of us say the right thing (and we have authority) Google will display the authoritative version as opposed to the popular one.

This has the same fluid caveats as any real life equivalent. So if a lot of us say the same thing and keep on saying the same thing then whatever we are saying will start to splinter from the authoritative source’s definition and will start to become a thing which will have the potential to provide an alternative, acceptable definition by widespread acceptance. In many ways we experience that in RL with the fluidity of language and definitions as well as some grammatical forms where the incorrect version is used “coz it’s used” as opposed to the correct form (the word “data” springs to mind as the perfect example here – it’s plural but many engineers use it as if it’s singular instead of writing “datum”).

Aaron Bradley: David Amerland, what a great summary of how, in semantic search, authority can trump (or at least inform) popularity – thanks!  (Literally bookmarked under “Authority”.  Oh, did somebody say “Authoritah?” http://bit.ly/respect-my-authoritah)

The subject too brings to mind the Knowledge-Based Trust (KBT) method proposed by Google researchers (http://bit.ly/1zRQtdG), which would have Google rank a resource based on “the correctness of factual information provided by the source.”

KBT is a fascinating rabbit hole I urge you to dart down Omi Sido, but I won’t journey there myself now except to note the importance in KBT of the model that “jointly estimates the correctness of extractions and source data, and the trustworthiness of sources.”  That is, judging what’s the “best” answer to a query involves make an assessment of the quality of the dataalong with the trustworthiness and veracity of the data provider (of which David speaks).

It strikes me more and more as discussions like those stimulated by KBT arise that the challenges faced by the search engines now are increasingly – and increasingly baldly – epistemological.  Hell, even phenomenological. A bigger and bigger part of search engine research is being directed at figuring out how machines can answer philosophical questions that we humans not only grapple with, but often do so with great difficulty.

How do we tell a fact from a non-fact?  How do we distinguish a pretty good fact from a really good fact?  What are we to make of two seemingly authoritative but contradictory facts?

Those are now the sort of questions that machines are endeavoring to answer.  Or required to answer in the narrow sense that a search engine needs at the very least to always provide a preferential answer. A search engine responding to a query, unlike a human talking to another human, can’t say, “well, we disagree on that, but since it’s not important you hold on to your views, I’ll hold on to mine, and we’ll go have a beer together.” A search engine can’t waffle it’s way out of producing a result.

And these discussions are arising more often with our increased reliance on machines (and especially mobile devices) to provide us with answers, and however trivial it may seem for most queries, there’s always going to be an answer that’s selected from alternatives on your behalf. Fairly innocuous in consequence if your question is “where is the nearest Starbucks?” but perhaps less so for “best local physical therapy.”

Oh, and thanks Raja Rethinam and Teodora Petkova!

David Amerland: Aaron Bradley yep, love what you added here. This is where all the cognitive computing magic is taking us. OK for closed data sets as Watson demonstrated with medical research but a bit of problem in the open web. I have a suspicion that context, in this sense, creates a closed data set of sorts that would define a ‘best’ answer but that would mean that at some point my context might trump yours and I get a better answer than you, maybe? It would be great to see our devices deal with that. 😉

Teodora Petkova: On behalf of John Kellden (without him knowing that:)) I am upgrading this thread to epic [marginalia: “the challenges faced by the search engines now are increasingly – and increasingly baldly – epistemological.  Hell, even phenomenological. A bigger and bigger part of search engine research is being directed at figuring out how machines can answer philosophical questions that we humans not only grapple with, but often do so with great difficulty.” cit. Aaron Bradley]

John Kellden: Epic, Teodora Petkova is epic.

I believe we’re only in the beginning of the Avatar Metalogues – where we little by little rediscover that our digital selves are already engaged in metalogues.

With our ten bits per second processing speed, we can’t grok these processes – with one of the results a cornucopia of articles, bringing parts of this down to a level where we can at least begin to understand some of it.

However, if we switch to pattern recognition mode, we can already begin to see a convergence, between Intelligence Augmentation, Semantic Search, Semiotic Entities, Artificial Intelligence, Network-centric heuristics, and tons of other related fields.

Granted, that our human analog pattern recognition is somewhat buggy, belief-biased (shortcuts in) inference ladders, cognitive fallacies and much more – and here is where the algorithms brings a much welcome bonus gift – a slightly more objective additional feedback loop.

Aaron Bradley: Ha, thanks Teodora Petkova The “ha” because I’ve been writing a book about the phenomenological and epistemological implications of globally-used search engines for years.

Unfortunately (or perhaps fortunately for the reading public) I’ve mostly been writing it in my head, so Theory of the Search will almost certainly never see the light of day.

(Those of you of a post-structuralist bent may recognize that as a nod to Roland Barthes’ “Théorie du texte” (Encyclopædia Universalis, 1974), to which many of my ideas in this realm are indeed indebted, and is a piece which everyone should read.)

David Amerland: Aaron Bradley how can we get this book out of your head? Do we really need to hold you down and operate? 😀

Aaron Bradley: David Amerland You’d be disappointed. The dejected look on the face of that zombie that had been moaning “braaaaains” when he opened my skull cap was simply heartbreaking.

David Amerland: Aaron Bradley you’re cracking me up on a busy Friday 🙂

John Kellden: According to latest Semantic Search theory, the cracks in our heads is where the entity can peer out into the world.   🙂

David Amerland: John Kellden LOL, in that case Aaron Bradley has got a headstart on us thanks to his zombie encounter.

Teodora Petkova: John Kellden, feels like our “human analog pattern recognition” is biased by our bodies too, plus our lost connection to the group. Funny that you said cracks 🙂 Yesterday I while I was walking in a park I was thinking about David Amerland‘s comment: “I have a suspicion that context, in this sense, creates a closed data set of sorts that would define a ‘best’ answer but that would mean that at some point my context might trump yours ” The train of thought went this way: by default context is circum-sribing text, in a way it encircles it. The only way we can link texts from one circle (i.e. context) to another is through cracks in the context. Thankfully we have these cracks in the form of ambiguity, of biases or other “human” stuff. On the open web, we can have them as bridges between ontologies.

Teodora Petkova: Aaron Bradley, this is such a wonderful thing. Thanks for deciding to write that book soon. :)) [excuse my attempt for a tiny rhetorical approach :D] That would be a magnificent read, plus I will feel less weird (officially) when I tell people that web and text have so many functional and formal similarities.

Aaron Bradley: Teodora Petkova, The web is a text, or at least a vast corpus from which narratives are endlessly constructed, like the journey that weaves in and out of queries and websites, queries and websites.

Which means you can approach it with the same textual analysis tools that the likes of Foucault and Barthes and Derrida wielded to produce so many devastating insights into language.

The post-structuralist notion that there is no way a text can be said to have a canonical meaning – that a text is in essence rewritten every time it is read, because the of the text’s consumption changes its meaning – could not be truer of search queries.

The hunt for meaning in search now does not get underway when Google starts to hunt for resources to satisfy a query, but before with Google hunting for the meaning in the query itself.

Teodora Petkova: Aaron Bradley 🙂 I will have to work on the last paragraph to grasp it fully 🙂 Thanks for these thoughts.

The thing that is exciting is that before these texts weren’t as accessible. Plus accessible is only (mainly) accessible via search.

The new version of Google Translate application

Google Translate gets smarter with language detection

Today in their article called ‘Google Translate gets smarter with language detection, Word Lens‘, www.mashable.com announced that Gooogle is updating their Google Translate by integrating  Word Lens, which instantaneously translates written text.

Google surprises us every single day. What was probably a science fiction just a few years ago, today is not just reality but also right there in our hands (yes this feature is coming to both the iOS and Android versions of Google Translate, so iPhone lovers are not left behind the digital revolution again – lol).

Google Translate gets smarter with language detection

 

It was only recently that David Amerland said: “Images Can Be Translated” (http://goo.gl/Ag8G7P)  and he really meant “into words”, but Google went one step ahead in no time and not just created an algorithm to “see” what’s in an image but even how to translate it into a different language. Right there before your eyes.

How did it all started? Well nobody really knows, but I will give a big credit to the research done by Bettina Harriehausen-Muhlbauer and Timm Heuss in 2012 called ‘Semantic Web based Machine Translation’ (http://mt-archive.info/EACL-2012-Harriehausen.pdf).

Two things to point out in this research.

“Especially with translations, it is often crucial to understand the source text correctly, as otherwise ambiguities may result in incomprehensible target language translations”.

Context is the king when it comes to semantic search. Everybody knew that from the very beginning.

“One of the leading users of SMT is Google and Google Translate engineer Anton Andryeyev, who explains SMT’s essence as follows:

SMT generates translations based on patterns found in large amounts of text. […]

Instead of trying to teach the machine all the rules of a language, SMT effectively lets computers discover the rules for themselves. It works by analysing millions of documents that have already been translated by humans […].”

Correct me if I am wrong, but are we not talking about Machine Learning here? Way before we started discussing it in the Internet world.

Semantic search technologies are slowly but surely coming to our everyday life. As they say in the article “it may not solve all the world’s problems”, but semantic search is making our life a lot easier.

2015 SEO Predictions

Semantic Search

Semantic Search is changing SEO forever.

It’s all about optimising your content and everything you do online for your customers rather than just Google, and this is a radical shift in thinking.

SEO 2015 Workout: How To Gain Weight from SEMrush

Google wants us to be more human. Words like trust, authority, personality are fashionable again.

“Identity is key” said David Amerland once.

In a way whatever you do offline to impress the people around you, now you have to incorporate in your everyday online activities. Both personal and for your company/brand.
A transition from corporate to human branding.

Is semantic search good for website owners?

I would shout loudly “Yes” and some of my main points are:

Semantic search is all about content, so writers with good knowledge of a particular subject will benefit even if they don’t know much about SEO.

Semantic search is less complex than normal indexing algorithms. People will see search results based on synonyms rather than exact keywords. Looking for “car shipping” will also give you results for “auto” and “vehicle” shipping. Think about good content rather than keyword optimization (goodbye black hat seo). Or in simple words – “write for the people, they are your clients, not the search engines”.

But what about “keywords”?

Well I am afraid Google is shifting towards more contextual search so focusing on keywords alone will be disastrous for any business. On the other hand contextual search is boosting small business web visibility immensely across the globe. Imagine a few years ago a small restaurant appearing on the first page of the search results for generic terms like “restaurant” or “pizza”.
No more injustice just because you are small business. Local businesses can rank highly for those keywords within a certain context.

Which brings me to my last but not least vital 2015 SEO point.

Local and National/International Results.

Location is the most obvious contextual feature Google uses when catering for a search query. Even if you don’t know or don’t care about semantic search you must have noticed that for certain keywords Google is displaying only local businesses. Try searching Google for the keyword “plumber”. Even if you wanted to find the meaning of this word or a job description, you are presented with local plumbers. Based on your location Google decides whether the keyword used should provide local or national results. And honestly do you want to see national results for “petrol station” or “Indian restaurant”.On the other hand do you want to see local results when looking for something like a bank account or mortgage?
Local search is boosting small/local businesses visibility so if you are small business this should be you main starting point.

Next year “expect SEO to be harder, better, faster, stronger”.

P.S. Article written as a comment to a post by David Amerland

When Obama & +Barry Schwartz Became King Of The United States For The Day

King Of The United States

Conversation with David Amerland and Peter Hatherley

“Why would Google show this answer when it is clearly wrong? Is it an Easter Egg? I strongly doubt it. But it does look to be a case where Google got it wrong when it comes to getting answers from third party sources.” – Barry Schwartz

King Of The United States

Peter Hatherley: Still semantically relevant of course Omi Sido. This is where context comes into it’s own, I believe, and that remains a major part of the deep learning process that Google is currently engaged in …

Omi Sido: Peter Hatherley I was about to say something similar, but I got stuck in the thinking of how this has happened.

Not a long time ago Google was talking about entity salience. As far as I know names are always salience entities. And the Knowledge Graph has a lot of background information about Barak Obama and the difference between a king and a president.

As you said yourself : The deep learning continues.

Peter Hatherley: Omi Sido the how of the incident is as much a part of the picture too. Appreciate you bringing it to our attention 🙂

David Amerland: Omi Sido this shows the weakness that exists, still, in the veracity component of semantic search.

Peter Hatherley: This is why contextual categorization is so important David Amerland. This is why a simple contextual system that isolates positive, neutral and negative elements could answer correctly. The analysis of the textual content obviously being related more clearly to entities and authority etc.

The truth is that king’s and president’s are intrinsically linked semantically yet there is no King of America.

This would be easily solved by matching these titles against countries as a boolean query.

David Amerland: Peter Hatherley agreed, it shows the complexity of the problem: i.e. what is semantically correct is incorrect at a human understanding level and in this particular example the issue is compounded further by the sourcing of the Barry Shwartz photograph to illustrate it. It shows that Google’s social-mining component is not yet linked quite the same way, in the verification process.

Peter Hatherley: Yes quite true David Amerland it does expose that flaw. The point is that now that this has been activated will it remain part of his semantic profile for the future?

David Amerland: Peter Hatherley no, I suspect it will drop off quite quickly. The checks and controls are in place, though the speed we expect them to work at, is not. One unique element of the semantic web is that results are cached so that new calculations are made on top of new ones instead of starting from scratch each time like we had to with Boolean search. Thus, with time, veracity and the speed at which it is calculated will improve. What we are seeing here are the issues of a nascent technology that has not yet scaled adequately.

Peter Hatherley: Yes that makes perfect sense David Amerland eventually it will be far far more organic and serendipitous in nature

Topic Modeling & Semantic Connectivity

Topic Modeling & Semantic Connectivity

Topic Modeling & Semantic Connectivity By Rand Fishkin .

Notice when he says “use your own mental intelligence to say, Are these terms and phrases relevant? Should they be included? Are these things that people would be looking for? Are they topically relevant?”.

So true – “use your own mental intelligence”.

For a long time I have been advising people to use a simple trick for long tail keywords, entities and their relationships to one another. Many people know this trick, yet not that many people are doing it.

First everybody knows that most of the information in Google Knowledge Graph comes from Wikipedia. Secondly it is by far the largest online keyword and entity library.

So the trick is to use the string: inurl:Wikipedia. Say you are in the “Social media marketing tips” business like Ana Hoffman. Just type Social media marketing tips inurl:Wikipedia in Google and hit the search button. An ocean of Wikipedia pages coming your way (optimized headlines I love). Have a look at the highlighted words in black, the words surrounding them and write them down. Scroll down to the related searches and have a look at the suggestions there. Write them down as well. Now click on the first Wikipedia result. Just have a look at the “Content” section main headlines. Keyword/entities heaven. No other keywords on this planet are more related to your business niche than those navigational links in a Wikipedia page and Google is ranking them high.

Scroll down to the reference section. It looks a bit messy, but don’t give up. Look properly and  you will see things like “Generation of Business Engagement” or “Social Network in marketing Opportunities”.So many of them I don’t even have enough time to mention them here. Don’t you have enough highly targeted keywords and entities relationships now?!

Before going to the last step just type “Generation of Business Engagement” in Google and hit the search button. What do you see? Four articles talking about Social Media Marketing, one of them even in fourth organic position. But wait….click on the ones that do not have Social Media Marketing keywords in the title or description. Go for the third one “The Next Generation of Business Engagement – Wiley”. I know you are smiling as the page you see is actually talking about a book called “Social Media Marketing: The Next Generation of Business Engagement“. Suddenly all the entities and their relationships in semantic search David Amerland is talking about are becoming a bit more clear to you.

Now for the final step of the trick just go to Google Keyword Planner and use all the keywords from Wikipedia to get ideas. That’s it….you are ahead of the competition. You just had to  “use your own mental intelligence”.

Data Density and Semantic Search – Hangout On Air with David Amerland and Mark Traphagen

Data Density and Semantic Search - Hangout On Air with David Amerland and Mark Traphagen

Semantic search is going to change everything about the way we do our SEO (Search Engine Optimisation).

In this video David Amerland and Mark Traphagen talk about what’s new in the Semantic World, what we are seeing and what needs to be done by web marketers and website owners in order to utilise the power of the emerging semantic technologies.

But first thing first.

What is Semantic Search?

According to Wikipedia:

Semantic search seeks to improve search accuracy by understanding searcher intent and the contextual meaning of terms as they appear in the searchable dataspace, whether on the Web or within a closed system, to generate more relevant results.

(On the other side to understand what Semantic Search is not you have to read this article by Teodora Petkova)

One thing comes to mind straight away when looking at the statement above: Searcher intent and relevant results in one sentence mean only one thing. We have to know our audience. And we have to know it good.

Semantic Search brings the human element back to the search engines. So in order to satisfy the human being and the search engine at the same time we have to figure out what is that thing that is going to attract visitors to our web pages and at the same give reason to Google to show our pages in the SERP’s.

There is only one answer. Content.

I’m not going to talk about content marketing here, but I am going to talk about a common misconception when it comes to content.

Many people think that writing ‘good’ content alone is enough to succeed in the war of the SERP’s. Good content is important, but it does not make your website unique in any way. If you don’t know what your audience’s biggest problems are and if you are not giving them a solution for these problems then you just have ‘good’ content that nobody is interested in. 

Listing your qualities/products without serving your potential customer’s need gives them…just a list. Zara Altar 

So in order to satisfy the search engines at the same time as the human visitors David Amerland is giving us 3 points:

  • Context
  • Relevance
  • Clarity

Which gives us a new definition of Semantic Search from a business point of view:

Semantic Search is taking into account the contextual relevance of your content in order to deliver highly personalised answer to the searcher’s query.

Notice I’ve used the word “query” which according to Google have the following definition: a question, especially one expressing doubt. Semantic search is seeking to bring clarity to the web. I can predict that even if all current Google algorithms change in the future the semantic notion will stay the same. So if you want to bulletproof your future marketing efforts you should start seeking clarity to your content and brand voice.

Now let’s go back to the hangout and the wisdom coming from this online chat:

Semantic Search/Web is here to stay.

“Everything you do from now on “have to have meaning“.  There are no shortcuts and there are no longer any easy wins in term of rankings. Google now requires proof of concept before anything works for you.
So if you say that you are absolutely the best in something and your website should be there, Google now needs additional proof beyond you saying this on your website.

And the way to do that is  trough “relational connections” across the web.  So if your website is actually really good at something and your business is great and you are winning in terms of quality and customer experience than is highly unlikely that your website is in the dark side and it’s never mentioned anywhere on the web apart of your own website.” – David Amerland