Recently, Google Webmaster Trends analyst Gary Illyes surprised many of us with a remark he made during his keynote Q&A with Danny Sullivan at SMX East in New York City. Illyes said that he recommended webmasters not remove the rel=author tag from their site content.
Google had used rel=author as part of its Google Authorship feature that (potentially) displayed a special author rich snippet in search results for content using the tag. Google ended support of this feature in August 2014.
The phrase that made everyone sit up and say, “Did he just say that?” was this: “…because it is possible Google might make use of [rel=author] again in the future.”
Even though Google’s John Mueller made the same recommendation after he announced that Google was no longer making use of Google Authorship in search (to be precise, Mueller said leaving the tag in place “did no harm”), Illyes’s statement seemed to shock many because Google has said nothing about Google Authorship or the rel=author tag since they said they stopped supporting it.
In a subsequent Twitter exchange I had with Gary Illyes, he explained that if enough users are implementing something, Google might consider using it. I asked him if that meant specifically if more people started using rel=author again, that Google might make use of it again. Illyes replied, “That would be safe to say.”
Before I provide my commentary on what all this means, and whether we should expect to see a resumption of Google Authorship in Google Search, let me provide a brief overview of Authorship for anyone who may not be familiar with it. If you already understand Google Authorship, feel free to skip down to the Will Google Bring Back Authorship? section.
A brief history of Google Authorship
Google Authorship was a feature that showed in Google Search results for about three years (from July 2011 until August 2014). It allowed authors and publishers to tag their content, linking it to an author’s Google+ profile, in order to provide a more-certain identification of the content author for Google.
In return, Google said they might display an authorship rich snippet for content so tagged in search results. The authorship rich snippet varied in form over the three years Authorship was in use, but generally it consisted of the author’s profile photo next to the result and his or her byline name under the title. For part of the run of Authorship, one could click on an author byline in search to see results showing related content from that author.
Google Authorship began with an official blog post in June of 2011 where Othar Hansson announced that Google would begin supporting the rel=author tag, but with no specifics on how they might use it.
Then in a July 2011 video, Hansson and Matt Cutts explained that Google+ would be the hub for author identification, and that Google might start showing a special Authorship rich snippet result for properly tagged content.
Those rich snippets slowly began appearing for more and more authors using rel=author over the next several months. During the three years of the program, Google experimented with many different configurations of the rich snippet, and also which authors and content would get it in response to various search queries.
Interest in Google Authorship from the SEO and online marketing communities was spurred even more by its possible connection to Google’s Agent Rank patent, first revealed by Bill Slawski. In this patent, Google described a system by which particular “agents” or “entities” could be identified, scored by their level of authority, and that score then be used as a search ranking factor.
Since one of the types of agents identified in the patent was a content author, the patent rapidly became known as “author rank” in the SEO community. The connection with Authorship in particular, though, came from Cutts and Hansson stating in the above-mentioned Authorship video that Google might someday use Authorship as a search ranking factor.
Speculation about so-called Author Rank, and whether or not it was “on” as a ranking factor, continued throughout the life of the Authorship program. Throughout that period, however, Cutts continued to refer to it as something Google might do in the future. (You can find my own take on why I believed Authorship was never used as a direct ranking factor here.)
The first hint that Google might be drawing back from Authorship came at Pubcon Las Vegas in October 2013 when Matt Cutts, in his keynote “State of Search” address, revealed that at some point in the near future Google would be cutting back on the amount of Authorship rich snippets shown by “around 15%.” Cutts said that in experiments, Google found that reducing Authorship rich snippets by that much “improved the quality of those results.”
Sure enough, in early December of that year, Moz’s Peter Meyers detected a rapid decline over several days in the number of Authorship rich snippets in search results, as measured by his Mozcast Features tool.
Around that same time Google implemented what I called “two-class Authorship,” a first class of authors who continued to get the full rich snippet, and a second class who now got only a byline (no author photo).
Finally, in August 2014, this author was contacted directly by John Mueller, offering to share some information under an NDA embargo until the information was made public. In my call with Mueller, he told me that he was letting me know 24 hours in advance that Google Authorship was going to be discontinued. He added that he was making this call as a courtesy to me since I had become the primary non-Google source of information about Authorship.
With that information, Eric Enge and I were able to compose an in-depth article on Authorship and its demise for Search Engine Land that went live within two minutes of John Mueller’s own public announcement on Google+. In our article linked above, Eric and I give our takes on the reasons behind the death of Authorship and the possible future of author authority on Google.
Will Google bring back Authorship?
From the day Authorship was “killed” in August 2013, we heard no more about it from Google—until Gary Illyes’s remarks at SMX East. So do Gary’s remarks mean we should expect to see a return of Google Authorship to search results?
I don’t think so, at least not in any form similar to what we saw before.
Let me explain why.
1. Illyes made no promise. Far too often people take statements about what Google “could” or “might” do from spokespersons like Gary Illyes, Matt Cutts, and John Mueller and translate “could/might” to “will.” That is unfair to those spokespeople, and an abuse of what they are saying. Just because something is spoken of as a possibility, it does not follow that a promise is being made.
2. It ain’t broke so…. So if there are no actual plans by Google to restore Google Authorship, why would Illyes make a point of stating publicly that authors and publishers should continue to use the rel=author tag? I think a primary reason may be that once Google gets any set of people to begin using any kind of schema, they’d rather have it remain in place. Anything that helps better organize the information on web pages is good for a search engine, whether or not that particular information is “in play” at present.
In the case of rel=author, I think it still may be useful to Google to be able to have confidence about content connected with certain authors. When Authorship ended, many people asked me if I were going to remove the tags from my content. I responded why would I? Having them there doesn’t hurt anything. But more important, as an author trying to build my personal brand reputation online, why wouldn’t I want to give Google every possible hint about the content with which I should be identified?
3. The reasons why Authorship was killed still remain. As with any change in Google search, we’ll probably never know all the reasons behind it, but the public reasons stated by John Mueller centered around Google’s commitment to a “mobile first” user experience strategy. Mobile first is a recognition that search is more and more a mobile experience. Recently, Google announced that more of all searches are now done on mobile than desktop. That trend will likely never reverse.
In response, we’ve seen Google continually moving toward simpler, cleaner, less-cluttered design in all its products, including search. Even their recent logo redesign was motivated by the requirements of the small screen. According to Mueller, Authorship snippets were too much clutter for a mobile world, with not enough user benefit to justify their continuation.
In our Search Engine Land article, Eric Enge and I speculated that another reason Google may have ended the Authorship experiment was relatively poor adoption of the tagging, low participation in Google+ (which was being used as the “anchor” on Google’s side for author identification), and incorrect implementation of the tags by many who did try to use them.
On the latter point, Enge conducted a study of major publishers, which showed that even among those who bothered to implement the authorship tagging, the majority was doing it wrong. That was true even among high-tech and SEO publications!
Alt that points to a messy and lopsided signal, not the kind of signal a search engine wants. At the end of the day, Google couldn’t guarantee that a result showing an Authorship rich snippet was really any better than the surrounding results, so why give it such a prominent highlight?
Despite Gary Illyes saying that if more sites used rel=author Google might begin using it again, I don’t see that doing so would change any of the conditions stated above. Therefore, I believe that any future use of rel=author by Google, if it ever occurs, will look nothing like the Authorship program we knew and loved.
So is there any future for author authority in search?
To this question, I answer a resounding “Yes!”
Every indication I’ve had from Googlers, both publicly and privately, is that author authority continues to be of interest to them, even if they have no sound way to implement it yet.
So how would Google go about assessing author identity and authority in a world where authors and publishers will never mass-tag everything accurately?
The answer: the Knowledge Graph, entity search, and machine learning.
The very first attempts at search engines were mostly human-curated. For example, the original Yahoo search was fed by a group of editors who attempted to classify every web page they came across. But as the World Wide Web took off and started growing exponentially, it was quickly obvious that such attempts couldn’t scale. Hyperlinks between web pages as a means of assessing both the subject matter and relative authority of web pages proved to be a better solution. Search at the scale of the web was born.
Remember that Google’s actual mission statement is to “organize the world’s information.” Over time, Google realized that just knowing about web pages was not enough. The real world is organized by relationships between entities—persons, places, things, concepts—and Google needed a way to learn the relationships between those things, also at scale.
The Knowledge Graph is the repository of what Google is learning, and machine learning is the engine that helps them do that learning at scale. At a simple level, search engine machine learning is the development of an algorithm that learns on its own as a result of feedback mechanisms. Google is applying this technology to the acquisition of and linking together of entities and their relationships at scale.
It’s my contention that this process will be the next evolutionary step that will eventually enable Google to identify authors who matter on a given topic with their actual content, evaluate the relative authority of that content in the perceptions of readers, and use that as a search ranking factor.
In fact, Matt Cutts seemed to hint at a Knowledge Graph-based approach in a June 2013 video about the future of authorship where he talked about how Google was moving away from dependence on keywords, from “strings to things,” figuring out how to discover the “real-world people” behind web content and “their relationships” to improve search results.
Notice that nothing in a machine learning process is dependent upon humans doing anything other than what they already do on the web.
The project is already underway. Take a moment right now and ask Google, “Who is Mark Traphagen?” If you are in the US or most English-speaking countries, you’ll probably see this at the top of the results:
That’s a Knowledge Panel result from Google’s Knowledge Graph. It reveals a couple of things:
1. Google has a high confidence that I’m likely the droids, er, the “Mark Traphagen” you’re looking for. There are a few other Mark Traphagens in the world who potentially show up in Google Search, but Google sees that the vast majority of searchers who search for “Mark Traphagen” are looking for a result about me. Thanks, everybody!
2. Google has high confidence that the Mark Traphagen you’re looking for is the guy who writes for Search Engine Land, so that site’s bio for me is likely a good instant answer to your lifelong quest to find the Real Mark Traphagen (a quest some compare to the search for the Holy Grail).
If Google can continue to do that at scale, then they can lick a problem like assessing author authority for search rankings without any help from us, thank you very much.
How does all this fit with Gary Illyes’s recommendation? I think that while Google knows it ultimately has to depend on machine learning to carry off such projects at scale, any help we can give the machine along the way is appreciated. Back in the Google Authorship I days, some of us (myself included) believed that one of the real purposes for the Authorship project was to enlist our help in training the machine learning algorithm. It may be that rel=author is still useful for that.
What might Authorship look like in the future?
Allow me to speculate a bit.
I don’t expect we’ll ever again see the mass implementation of author rich snippets we saw before, where almost anyone could get highlighted just for having used the tagging on their content and having a Google+ profile. As I stated above, I think Google saw that doing that was a non-useful skewing of the results, as more people were probably clicking on those rich snippets without necessarily getting a better piece of content on the other end.
Instead, I would expect that Google would see the most value in identifying the top few authors for any given topic, and boosting them. This would be similar to their behavior with major brands in search. We often see major, well-known brands dominating the top results for commercial queries because user behavior data tells Google that’s what people want to see. In a similar way, people might be happy to be led directly to authors they already know and trust. They really don’t care about anyone else, no matter how dashing their profile image might be.
Furthermore, for reasons also stated above, I don’t expect that we’ll see a return to the full rich snippets of the glory days of Authorship I. Instead, the boost to top authors might simply be algorithmic; that is, other factors being equal, their content would get a little ranking boost for queries where they are relevant to the topic and the searcher.
It’s also possible that such author’s content could be featured in a highlighted box, similar to how we see local search results or Google News results now.
But notice what I said above: “…when [the authors] are relevant to the topic and the searcher.” That latter part is important, because I believe it is likely that personalization will come into play here as well. It makes sense that boosting or highlighting a particular author has the most value when my search behavior shows that author already has value to me.
We already see this at work with Google+ posts in personalized (logged in) search. When I search for something that AJ Kohn has posted on Google+ while I’m logged in to my Google account, Google will elevate that result to my first page of results and even give it a good old-fashioned Authorship rich snippet! Google has high confidence that’s a result I might want to see because AJ is in my circles, and my interactions with him and his content show that he is probably very relevant and useful to me. Good guess, Google, you’re right!
It is now obvious that Google knows they have to expand beyond Google+ in entity identification and assessment. If Google+ had taken off and become a real rival to Facebook, Google’s job might have been a lot easier. But in the end, building machine learning algorithms that sniff out our “who’s who” and “who matters to whom” may be an even better, if vastly more difficult, solution.
So to sum up, I do expect that at some point in the future, author authority will become a factor in how Google assesses and ranks search results. However, I think that boost will be a “rich get richer” benefit for only the top, most reputable, most trusted authors in each topic. Finally, I think the output will be more subtle and personalized than we saw during the first attempt at Authorship in search.
How to prepare for Authorship II
Since it is unlikely that Authorship II, the future implementation of author identity and authority in search, will be anything like Authorship I, is there anything you can be doing to increase the odds that Authorship II will benefit you and your content? I think there are several things.
1. Set a goal of being the 10X content creator in your niche. Part of the Gospel According to Rand Fishkin these days is that “good, unique” content is not good enough anymore. In order to stand out and get the real benefits of content, you’ve got to be producing and publishing content that is ten times better than anything currently on page one of Google for your topic. That means it’s time to sacrifice quantity (churning out posts like a blogging machine) for quality (publishing only that which kicks butt and makes readers stand up, take notice, and share, recommend and link).
2. Publishers need to become 10X publishers. If you run a publishing site that accepts user-generated content, you’ve got to raise your standards. Accepting any article from any writer just to fill space on your pages won’t cut it.
3. Build and encourage your tribe. If you are authoring truly great, useful stuff, sooner or later you will start to attract some fans. Work hard to identify those fans, to draw them into a community around your work, and to reward and encourage them any way you can. Become insanely accessible to those people. They are the ones who will begin to transmit the signals that will say to Google, “This person matters!”
4. Work as hard offline as you do online. Maybe harder. More and more as I talk with other authors who have been working hard at building their personal brands and tribes, I’m hearing that their offline activities seem to be driving tremendous benefit that flows over into online. I’m talking about speaking at conferences and events, being available for interviews, being prominent in your participation in the organizations and communities around your topic, and dozens of other such opportunities.
BONUS: Doing all four of those recommendations will reap rewards for you in the here and now, whether or not Google ever implements any kind of “author rank.”
The natural power of the fact that people trust other people long before they will trust faceless brands continues, in my opinion, to be one of the least understood and underutilized methodologies in online marketing. Those who work hard to build real author authority in their topic areas will reap the rewards as Google begins to seek them out in the days to come.
