March 19, 2018
Back in 2010, Google was getting beaten up in the media for the increasing amount of “content farm” clutter in the search results. That negative press was so overwhelming that Google felt it had no choice but to respond:
[We] hear the feedback from the web loud and clear: people are asking for even stronger action on content farms and sites that consist primarily of spammy or low-quality content.
Soon after that, in February 2011, the Google Panda update was released, which specifically targeted spammy and low-quality content.
Why do I bring this up today? Because the media has been hammering Google for promoting fake news for the past year and a half — a problem so extensive that search industry expert Danny Sullivan has referred to it as “Google’s biggest-ever search quality crisis.”
Needless to say, these accusations are hurting Google’s image in ways that cut far deeper than content farms. While the problem of rooting out false information is a difficult one, it is one that Google has a great deal of motivation to solve.
Google has already taken action to combat the issue in response to the negative press, including banning publishers who were promoting fake news ads, testing new ways for users to report offensive autocomplete suggestions, adjusting their algorithm to devalue “non-authoritative information” (such as Holocaust denial sites), and adding “fact check” tags to search results.
Of course, the issue of trustworthy search results has been on Google’s radar for years. In 2015, researchers from Google released a paper on Knowledge-Based Trust (KBT), a way of evaluating the quality of web pages based on their factual accuracy rather than the number of inbound links. If implemented, the Knowledge-Based Trust system would ultimately demote sites that repeatedly publish fake news (although there is a potential for it to go wrong if the incorrect facts become widely circulated).
Whether the Knowledge-Based Trust method is enough to combat fake news — or if some version of it has already been implemented without success — is difficult to say. But, it’s clear that Google is interested in making truthfulness a ranking factor, and they’ve never had a stronger motivation to do so than now.
One in five mobile search queries currently comes from voice search — a number that is likely to rise as Google Assistant-enabled devices such as Google Home continue to grow in popularity. And as voice search grows, we can expect to see an increase in featured snippets, from which Google often sources its voice search results.
Indeed, there is already evidence that this growth is taking place. A study released by Stone Temple Consulting last year confirmed that featured snippets are on the rise, appearing for roughly 30 percent of the 1.4 million queries they tested.
If this trend continues, featured snippets may even begin to rival the top organic listing as the place to be if you want to get noticed. (For more on featured snippets and how to target them, check out Stephan Spencer’s excellent primer on the subject.)
It’s now been over two years since we were first introduced to RankBrain, Google’s machine-learning AI system which helps to process its search results. Since its introduction, it’s gone from handling 15 percent of search queries to all of them.
Google’s interest in AI extends much further than RankBrain, however. They have developed the Cloud Vision API, which is capable of recognizing an enormous number of objects. Indeed, Google has so much machine-learning capacity that they are now selling it as its own product.
But perhaps most interestingly, Google has now built an AI that is better at building AI than humans are. This was a project by Google Brain, a team that specializes specifically in building AI for Google.
Unfortunately, AI is not without its issues. AIs tend to get stuck in local minima, where they arrive at a “good enough” solution and are unable to climb out of it in order to discover a better solution. They also have a tendency to confuse correlation with causation; one might even call them “superstitious” in that they draw connections between unrelated things. And since the developers only program the machine-learning algorithm, they themselves don’t understand how the final algorithm works, and as a result, have even more difficulty predicting how it will behave than in the case of traditional programs.
As Google continues to embrace AI and incorporate more of it into their search algorithms, we can expect search results to start behaving in less predictable ways. This will not always be a good thing, but it is something we should be prepared for.
AI doesn’t change much in the way of long-term SEO strategies. Optimizing for AI is essentially optimizing for humans, since the goal of a machine-learning algorithm is to make predictions similar to those of humans.
In May, Google warned webmasters that using article marketing as a large-scale link-building tactic is against its guidelines and could result in a penalty. Since this is already well known in the SEO community, Google’s announcement likely signals that an algorithm update targeting manipulative guest posting is on the horizon.
What counts as manipulative guest posting? To me, the most vital piece of information from Google’s guidelines has always been the recommendation to ask yourself, “Does this help my users? Would I do this if search engines didn’t exist?”
Guest posts that don’t expand brand awareness or send referral traffic aren’t worth doing, except for the possibility that they will positively impact your search engine rankings. The irony of taking that approach is that it isn’t likely to work well for your search engine rankings either — at least not in the long term.
I’m not saying anything that isn’t common knowledge in the SEO community, but I have a feeling that a lot of people in this industry are fooling themselves. All too often, I see marketers pursuing unsustainable guest posting practices and telling themselves that what they are doing is legitimate. That is what a lot of people were telling themselves about article marketing on sites like EzineArticles back in the day, too.
Bing has confirmed that they track unlinked brand mentions and use them as a ranking signal — and a patent by Google (along with observations from many SEO experts) indicates that Google may be doing this as well.
As AI begins to play a bigger part in rankings, it’s not unreasonable to expect “linkless” mentions of this type to start playing a bigger role in search rankings.
The tactics used to earn brand mentions are, of course, not much different from the tactics used to earn links, but since the number of people who mention brands is much higher than the number of people who link to them, this could provide a good boost for smaller brands that fall below the threshold of earning press.
This highlights the importance of being involved in conversations on the web, and the importance of inciting those conversations yourself.
The early 2017 mobile interstitial penalty update was a sign of Google’s continued battle against intrusive mobile ads. The hardest hit sites had aggressive advertising that blocked users from taking action, deceptive advertising placement and/or other issues that hindered use of the interface.
However, columnist and SEO expert Glenn Gabe noted that the impact of this penalty seemed… underwhelming. Big brands still seem to be getting away with interstitial ads, but Google may decide to crack down on these in the near future. The crucial factor seems to be the amount of trust big brands have accumulated in other ways. How all of this shakes out ultimately depends on how Google will reward branding vs. intrusive advertising.
It’s been nearly three years since Google announced that mobile searches had finally surpassed desktop searches on its search engine — and just last year, BrightEdge found that 57 percent of traffic among its clients came from mobile devices.
Google is responding to this shift in user behavior with mobile-first indexing, which means “Google will create and rank its search listings based on the mobile version of content, even for listings that are shown to desktop users.” Representatives from Google have stated that we can expect the mobile-first index to launch this year.
In other words, 2018 very well may be the year where signals that used to only impact searches from mobile devices become signals that impact all searches. Sites that fail to work on a mobile device may soon become obsolete.
Google has come a long way since it first hit the scene in the late 1990s. The prevalence of AI, the political climate and efforts and warnings against manipulative guest posts and distracting advertisements, all signal that change is coming. Focus on long-term SEO strategies that will keep you competitive in the year ahead.