From AMP to Algo Updates - Rank Ranger's SEO Industry Research
December 19, 2017 |
There were a lot of important changes and updates that had an impact on the industry in 2017. We're both honored and humbled to be a part of reporting and analyzing the big stories that affect the industry. I can proudly say that our team has really done a great job digging into SERP feature data shifts, Google algorithm updates, and more. With another year behind us, let's have a look back at some of the weightier findings we brought to the forefront in 2017.
AMP - Pushing Slower Results Out of Mobile Carousels Since 2017
We covered a lot when it came to AMP this year. In April we reported that organic AMP results took some losses
, in July we showed that new markets had started seeing AMP in the news carousel
, etc. In truth, we've reported on dozens of SERP feature data shifts over the year, but these data discoveries are not our purpose here.
Well actually, there is one such discovery that I thought was worthy of being mentioned, even though it was found via careful monitoring more than diving into data or running an experiment, and that is the huge uptick of AMP in the news carousel.
AMP in the News Carousel
Just as the new year was upon us, our Mobile SERP Feature Tracker
saw what was just a huge spike in AMP's presence within the Top Stories carousel
. When AMP first hit the news carousel in April 2016, roughly 15% of all of its cards contained an article optimized for AMP.
Baseline data for AMP within the mobile news carousel
Jump forward to January 26 and that number starts to really jump. By the end of January 2017 AMP occupied anywhere from 40% to 85% of news carousel cards depending on the country. The new data pattern has, for the most part, remained consistent up until the time of the writing of this article nearly a year later.
Mobile news carousels show more AMP results globally
AMP in Organic Carousels
The mobile news carousel was not the only SERP real estate where AMP became the dominant force. About a month after reporting on AMP's news carousel proliferation, we reported on AMP preference within organic carousels.
The story behind this realization is actually a bit funny. I'm a semi-serious football fan (American football that is). One cold day I happened to have my browser displaying as a mobile device and did a search related to some sort of football news. While scrolling through the results I happened to notice that sports juggernaut ESPN did not have a mobile carousel within their result, but some no-named sports site did. How odd, right? The only difference I could spot was that the John Doe of sports sites was showing up as an AMP site, whereas ESPN was not.
An AMP site showing with a mobile carousel while a non-AMP ESPN does not
As a result of my fascination with men bashing their brains in on their way to an early onset of Alzheimer's, I decided to take a look at some of the top sites from various industries that had not moved over to AMP in order to see if they showed mobile organic carousels. The plan was pretty simple. Go to these sites, see the top stories they were running, and run a Google search on mobile for keywords that aligned to the titles of these stories. In other words, I was targeting the stories in my searches.
The result of query after query was that if Google had an AMP page available on the SERP that was "carousel ready," Google would show a carousel for that site and not for a site that had not gone AMP, no matter how prominent the non-AMP site was.
The non-AMP cookingchannel.com shows without a mobile carousel
Thus, heavy-hitters like Politico, Rolling Stone, ESPN, Vogue, or the Cooking Channel, were organic carousel no-shows. Interestingly enough, when Google had no AMP alternative, these very sites showed a carousel within their mobile results.
Rolling Stone's results display with a mobile carousel when no other AMP sites are available on the SERP
The moral of the story, if you want your organic results to display with a carousel on mobile, go AMP.
Read the full study on Google giving mobile carousel preference to AMP results
Breaking Down and Analyzing 2017's Google Updates
Google's says that they (or should I say RankBrain) are constantly updating the algorithm. That said, some updates are more significant than others of course. This past year we tried to analyze the impact of a few such updates. At times we were unsuccessful in showing any sort of pattern, but we did have our moments in 2017.
Google's Mid-May Update
On May 18th the industry was abuzz with talk of a pretty significant Google update. There were rumors flying around that a Featured Snippet shakeup was behind the increased rank fluctuations. Around the same time as our Rank Risk Index
caught the fluctuation, Google had removed the URL shown within a Featured Snippet from the organic results. We looked into it and found that only 4.5% of all Featured Snippet URLs did not also appear in the organic results. Not nearly enough to cause a rank shakeup.
In the course of our research we did find an interesting rank pattern related to some eCommerce sites. A few of the major retail sites saw an abnormal drop in their visibility. While we wouldn't define the update by its targeting of retail sites, the data shifts were quite significant. For example, Amazon saw a 7% visibility loss and had their average rank fall from .21 of a position to 7.34.
Amazon was not alone as other big name sites such as Best Buy and Overstock saw similar losses.
eBay on the other hand experienced a significant increase in their visibility (29%) and had their average rank move up an entire position.
Access the full write up on the May 18th Google update
Google's June 2017 Update
The end of June 2017 saw a massive Google update that everyone in the industry was talking about. In fact, this update was the longest running update we had ever tracked, lasting an astonishing 10 days. Digging deep and trying to ascertain what exactly transpired was of course quite tempting, like the way candy is to a kid in a candy store.
The roll out of June's algorithm update was the longest ever tracked by the Rank Risk Index
Not wanting to let this historic update pass us by, we relentlessly dug into our ranking data. I say relentless because we were continuously hitting a brick wall and grasping at straws. That is until we found that most of the ranking changes had occurred between the 6th and 10th position. Meaning, Google left the top results alone, relatively speaking. We looked at the data niche by niche and across multiple industries and found that rank fluctuations significantly picked up past the 5th position on the SERP.
This of course is not to say that there weren't other dynamics at work during the update or that other rank positions did not fluctuate more than normal. In fact, there was a significant increase in rank fluctuations when we analyzed positions 3 - 5 as well. The fluctuations seen in these positions only lost their consequence when compared to the increase in fluctuations seen towards the bottom of the SERP.
See the full study for data on other niches that saw substantial rank fluctuations during Google's June algorithm update
Google ccTLD Update Impacts SERP Features Globally
As of October 27th you can no longer access international search results via a Google specific ccTLD. In less technical terms, you can only see search results for your location. Thus, if you are in the US, you can no longer go to google.ca to access the Canadian SERP or google.com.au to see the Australian SERP, etc. (As an aside, there is a nice little workaround that does allow you access to country-specific SERPs
Providing hyper-local results with no country overlap (as was common in some European countries) required changes on the algorithmic level. These changes resulted in a series of peculiar SERP feature behaviors. We found that Ads and Local Packs in markets across the globe underwent significant data alterations as a result of the ccTLD update.
Ads in Europe underwent an unusually sharp increase that paralleled Google's change to its ccTLD policy
Delve into the full analysis of all of the Google ccTLD update SERP feature data alterations
Case Study on How to Analyze Google Updates in the RankBrain Era
After seeing some downright odd algorithmic activity in September, we went on the hunt for an explanation. To be honest, pinpointing any sort of pattern was unusually difficult. However, along the way we came across some sites that highlighted just how tricky tracking an algorithm's impact on your site can be in the age of RankBrain.
A series of one-day Google updates dominated the second half of September
Long story short, the way we traditionally think about Google updates doesn't really apply to RankBrain. One time large rank losses (or gains), while easy to identify, are not exactly the norm in the RankBrain era. When it comes to tracking an update's impact vis a vis RankBrain, it pays to look at the whole picture. Often enough, a sharp loss is the reversal of a previous spike and in many instances is quickly undone as a site's rank returns to previous norms.
Visibility for the site emedicinehealth.com underwent a reversal of previous gains in September 2017 and was followed by strong gains in October and November
Also, as opposed to being hit by say Panda, a serious rank loss that results from RankBrain may play itself out over time. Whereas the more "traditional" algorithms are quick and messy, RankBrain tends to push a site down gradually. Which of course means you need to be vigilant and on top of your rankings.
angielist.com sees a slow and gradual decline in visibility over the course of 7 months
Check out our case study in full to see which mistakes to avoid when tracking Google algorithms in the RankBrain era
Google's Uphill Battle Fighting Fake News
Of all the research pieces and case studies we did this year, the piece we did on Google's ability to fight fake news is by far my favorite, which is why I saved it for last. It all started with a "Hey Mordy, check this out, Google is showing a News Box for a query on reverse mortgages." I thought that was odd. I mean how could Google show a News Box for a topic that had no news. It was not as if a new law passed that impacted reverse mortgages. In fact, all of the results within the SERP feature were opinion pieces. It made me wonder if Google really could identify genuine news content. If my hypothesis that it could not was true, then how could Google fight fake news?
As it turns out, my hunch that Google has a hard time identifying news content was an understatement. During the course of my investigation I found News Boxes containing not only how to articles, but actual spam and even promotional content.
A News Box that shows for the query "bad credit" presents a series of spammy results
As hard as it is to believe, things only went downhill from here. I ran a series of generic and news-centric queries. What I got back was... less to be desired. When interpreting general, yet highly relevant news queries, Google has tremendous difficulty showing accurate results. That is, keywords such as most important news
, big news
, and even breaking news
all pose a problem for Google.
Google is unclear as to the intent of the query "big news" as it displays a Disambiguation Box and shows only those news results that contain the word "big" in the article's title
What became clear was that Google relies far too heavily on a site's status as a news site and excessively on an article's title. It being the case that Google cannot decipher news content qualitatively, the search giant is intrinsically incapable from preventing fake news content from appearing on the SERP.
The proof is in the pudding. Google is continuously subject to fake news scandals. Google has responded to these fake news incidents with a series of external fixes that are all but patchwork. Meaning, instead of dealing with the underpinnings of preventing fake news from appearing in the News Box, Google resorts to mending its news quality problems with things like Publisher Knowledge Panels or quality raters. None of which impact news content placement algorithmically. Which is odd, or is it?
Definitely sift through our case study on why Google can't really fight fake news
, there are truly a lot of implications to the findings shown within it.
A Game of Hide and Seek
In an age where Google won't officially confirm its algorithm updates it becomes more and more important for the industry to plant our feet and dig deep. We feel it's extremely important to provide industry insights so that we can all navigate the playing field. It's actually an honor to be a part of the conversation and we hope to bring you even more research and analysis in 2018!
Do you have an SEO topic you would like us to look into and research? Reach out and let me know
, we'd love to hear from you.
Before I bid you adieu, I want to thank the entire Rank Ranger team. Sure, my name is on the research, and sure I worked hard to dig through a lot of mind numbing data. But this is not a one man show by any stretch. I want to thank the development team for accommodating my requests and making the research process so much more the easier. I also need to thank she who has probably lost her mind annotating my content, Teresa Hessler. Of course, I need to thank our fearless leader (and Rank Ranger CEO), Shay Harel, for working through the thick of it all with me. And last, and 1,000% not least, my partner in crime and our SEO manager, Tzvi Barish, who has not only inspired a lot of my posts with "Hey Mordy, you have to see this," but has been a constant source of brainstorming (and even editing!).