Someone asked me what the latest “small” Google update on September 27 was about. The only thing I could tell him was that the black hat link spammers saw it first.
Did that mean the recent update was related to spammy links? Probably not.
Black hats tend to build sites that are essentially a collection of landing pages built on the foundation of shopping keywords. So it could have been an update related to content. It could have been a lot of things.
If you’re going to put money on it, the surest bet would be on the likelihood it was about relevance.
Focusing on what kind of spam Google is “targeting” rarely, in my opinion, captures the improvements to relevance Google might actually be making. So let’s set aside considerations of what Google might have been targeting.
A better focus is on trying to understand how Google ranks websites today.
Google’s AI Algorithm
Google’s latest algorithm does remarkable things.
In my opinion, the most recent changes to the algorithm demands a re-evaluation of common SEO practices like focusing on things like title tags and heading elements.
As you’ll see below, Google is beyond matching title tags and heading elements to search queries.
If you want something to consider, consider how the author intent of a page of content relates to the intent of the searcher making the query.
Shopping sites should have words like shop, buy, and compare. Informational sites generally should not. These are signals of author intent, something that maybe isn’t discussed enough.
What Was the August 2018 Update?
The August 2018 Google update has been referred to as Medic by some in the SEO industry because anecdotal evidence suggested that Google was “targeting” medical sites.
But that was not the case.
In a Webmaster Hangout, Google’s John Mueller explicitly shot down the medical update idea with the following comments:
“The update we launched… around the first of August, was more of a general ranking update. Like we always do. So it’s not specific to medical sites. It’s something that could affect… any website out there.”
I declined to call the August update “Medic” because I understood that Google updates do not tend to “target” specific industries.
As you can see from Mueller’s comments, the August update was a “general ranking update.” And when we talk about ranking, we’re really talking about about relevance.
What search engines do is “rank” relevant websites.
Mueller followed up with the following comment:
“And in general it’s not something where we’d have any specific guidance to say, well this is what you need to change. This is really just a part of the normal changes that we make on the web overall.”
Neural Matching Algorithm
I wrote an article about what Google’s Neural Matching Algorithm, an AI algorithm that is involved in approximately 30 percent of search queries.
That article references various research papers to give an idea of how that kind of algorithm functions. In brief, the goal of that kind of algorithm is to come to a deep understanding of what a user means when they make a search query and also to understand what a page of content means. Then to put those two together, without the use of links.
I’m not saying that Google is ranking sites without links. Links may still play a role, as is indicated in the algorithm research papers. According to the algorithm research, this kind of algorithm kicks in after the ranking algorithm has done its part.
A Comparison of Ranking Algorithms
Comparing Google to Bing can help reveal what one or the other is doing. It can give an insight into how ranking is being accomplished.
An interesting example is a fishing lure that’s called the DR Minnow. It was discontinued for several years but has recently returned to production.
Some eBay sellers mistakenly call it the Doctor Minnow. The DR is a naming convention that has nothing to do with doctors. Humans mistake the DR for doctor, so let’s find out if AI does the same.
A Google search for “DR Minnow” shows about 20 results about the fishing lure, with the Daiwa website at the top of the search results. That makes sense, because users typing in the short phrase might prefer information about the lure, so the manufacturer page belongs at the top.
Bing, however, shows fishing related results in the top four positions and then begins to keyword match webpages. The DR part really trips up Bing because Bing thinks DR has something to do with a kind of street (Drive). Thus it shows a Zillow result of a house on Minnow Drive.
The lure manufacturer is not the top result on Bing, either. Bing’s shows results for doctors, too.
Bing appears to essentially be “guessing” that the DR means Drive or Doctor. What’s remarkable about Google’s results is that they are confident that the phrase DR Minnow is about a fishing lure and nothing but a fishing lure.
Now here is a demonstration of what I believe is an example of Neural Matching: A search for: “where to get a dr minnow.”
If I typed “where to buy a dr minnow” the intent is clear. A normal human understands that “get” means to “buy” something. Will an AI algorithm understand it?
Google does in fact understand that “get” within this context is the same as “buy.” This may be what Danny Sullivan was referring to when he mentioned “super synonyms.”
The search results begin with a shopping site, then the manufacturer, some videos talking about the lure and then more shopping sites.
Bing fails because it does not understand that the word “GET” is synonymous with “BUY” in the context of a search for “DR Minnow”, which Google clearly understands as being a fishing lure.
None of the algorithms mention the author intent or use that phrase. But I think it’s a useful phrase for understanding what Google is doing.
The concept of Author Intent can help us understand why a “medical” site about natural healing might not rank for medical phrases anymore. It could be because Google better understands that “natural” solutions are not, strictly speaking, medical ones and that vitamins are not pharmaceutical grade medicine.
And that may be why so-called “medical” sites seemed to have been “targeted.”
It wasn’t that they were targeted. And it wasn’t because they were medical. It was more likely because Google’s AI could better understand the Author Intent that the page is not about a “medical” solution, regardless of the keywords in the title tag and heading elements.
Screenshots by Author, Modified by Author