Is The Keyword Irrelevant In Today’s SEO Game?
Google loves to change the game on those of us that practice Search Engine Optimization.
There is actually an SEO term for this: the Google Dance. It all started way way back when Google realized that people might use the meta keyword tag inappropriately. I mean, did you ever ‘stuff’ that tag with keywords that weren’t 100% relevant to your business? This is a safe space, you can tell us… we won’t share that info with Google.
Fast forward a few years from the first published Google update - the Boston update - to today. It has been a long and bumpy ride between all of the fuzzy little animals (Penguin, Panda, and Fred - yes there was an update named Fred) and the core algorithm that Google uses currently.
Today, we have to consider things like RankBrain and Neural Matching. What that means is that we are starting to witness the slow demise of the keyword.
Now, before you put me in ‘take jail,’ I am not saying keywords are dead or irrelevant anymore. They are EXTREMELY important. As long as we are writing things down, the quality and context of that content will be key to that content's success. What is less relevant in today’s SEO game is the exact structure and the specific nature of how we go about creating web content.
What is becoming more important today is for computers to better understand how we speak and what we say. Computers are trying to better understand and process our language - also known as Natural Language.
A Paradigm Shift So Powerful, Yet So Subtle, You Probably Missed It
Recently, we've seen a shift from "having to put in very specific information in order to get to a result" to computers simply understanding the intent of the search and providing a result. Think about it: those of us who have been around since the early days remember that if you wanted to search for something, you had to be pretty specific. I.e. “italian restaurant in town X” if you wanted to be sure to get a good local result. Now, you search for “best italian restaurant” and you get a much better result.
How we speak has changed - and that means the queries have changed, and the results those queries produce have also changed. The system is now able to understand what we mean by a qualitative topic, like ‘best’ or 'near me.' Google knows not just what you're saying, but how you're saying it - and it's responding to match your intent.
So What’s The 411 on NLP (Natural Language Processing)
Chances are, you've heard of NLP before - likely, in reference to Watson, IBM's Jeopardy!-playing computer. While it might sound silly to program a computer to play a game (this is actually the 2nd time IBM has done this, remember Deep Blue?), the game of Jeopardy! is probably the most difficult language-based game for a computer to play. That is exactly why the IBM engineers took the challenge.
How can a computer programmed in 1s and 0s dominate a language-based game show? The best example of this is the final question from day one of the tournament. The question needed to be a US city. Both human players got it correct knowing that the largest airport was named for a WWII hero and the 2nd largest was named for a WWII battle.
To answer. you would have to know your WWII history, as well as the names of at least some major airports in the United States - all things that can be programmed into a computer pretty easily. But what isn’t easy is connecting those two dots. Watson got it wrong. (It chose Toronto - I mean come on, Watson, Toronto isn’t even a US city!)
Here’s the scary part… Watson wagered less than $1,000 on the question. It made the decision based on the assumption that it would not do well. Is your mind blown?
Watson beat the two Jeopardy! champions… badly. To the point where Ken Jennings wrote "I for one welcome our new computer overlords" on his screen at the end. Is Ken onto something?
So What’s New In NLP?
NLP has become the (latest) final frontier of AI. Companies are racing to better understand the English language so that they can better understand us - and make more potent business deicisons as a result.
The big data houses (Google, Facebook, Microsoft etc.) want to produce products/services that regular people will consume without feeling like they're talking to a robot. This goes beyond search engines, and extends even into using their everyday devices and associated technology like Google Home, Alexa, Portal, and more.
Tech companies are investing HUGE dollars into researching and creating these NLP systems.
So Why Is General Better Than Specific for SEO?
Recently, Barry Schwartz of Search Engine Roundtable published an article based on a conversation he had with Google about RankBrain and Neural Matching that really sums things up.
A bit of history/context on Neural Matching (if this is new to you):
- Google first announced the tech back in September 2018 at their 20th birthday party
- It had already been in play for a few months at that time, and was impacting 30% of queries
- It is described by Google as “super synonyms”
Google describes the difference between Neural Matching and RankBrain as:
- RankBrain helps Google better relate pages to concepts.
- Neural matching helps Google better relate words to searches.
Google also gave a few examples of how Neural Matching could be impacting search results. For instance, searches for “why does my tv look strange” might spit out a rich snippet about the "soap opera effect."
Time out. You didn’t include a single word about soap opera in that query, yet here we are. Why? Yep - neural matching.
Google has learned that when someone searches for that term, there is often something wonky going on with their TV's frame rate.
Simply saying "You need to be general when you search" is a little inflated. But the sentiment is correct. If you are looking to rank for a particular query, you don’t need that exact query in your copy. That means you can actually write something that will rank for queries that aren’t directly in your content!
TL;DR: well-written content (which Google has been evangelizing for years) is better than keyword-stuffed, spider specific mumbo jumbo.
So What’s In A RankBrain
Google uses RankBrain to “better relate pages to concepts.” This is described as "contextual signals." I have also seen RankBrain referred to as Google’s "happiness algorithm." If Google thinks that the link you clicked on was a positive result for the user, it will rank that site higher for that particular query. The result made you happy, so the site gets a little ‘boost’ from RankBrain.
Here's a question: can you specifically optimize your content for RankBrain? Or Neural Matching?
This is Google’s end-game. They are trying to finger snap SEOs out of existence… for good.
(Actually that isn’t entirely true. SEOs are still very valuable as part of a cross-functional marketing team, bringing a specific skill set for finding relevant keywords and topics to create content around. We can evaluate content based on specific marketing objectives to be sure it is maximized for search... and we can still build links.
Love ‘em or hate ‘em, Google has a stranglehold on search. Avoiding deep philosophical discussions on the flow and control of information - and if it is appropriate for one company to have that power - the point is that RankBrain is out there doing what it was programmed to do to match pages to concepts. It's all based on how we (people) present those concepts in the form of search queries.
Where Do We Go From Here?
OK, so it's semi-inflammatory to say that keywords are starting to be minimized in search - but they are still very much in play, as is having a really solid SEO strategy. That includes one that focuses on user intent and voice search.
I think we are in for a very exciting few years. As computers start to really understand our language (mostly English... for now), it will only become easier and easier to interact with them. I don’t think that is a bad thing… but I reserve the right to amend my take and follow Ken Jennings lead should the need arise!