Do Social Signals Impact Ranking - The Definitive Guide
If you've kept an eye on search over the past 5-7 years, you've probably run across the topic of social signals and their impact on organic rankings. This has been a battlefield for marketers and SEO's with views changing rapidly based on the latest Google statement. I imagine it's been a nightmare for social media managers as well, who have suddenly found themselves under attack from SEOs deciding they now own some social turf.
We can't fully blame the SEOs for their actions (or maybe we can, but that's another discussion), as many of them are akin to cowboys; they're just trying to survive. They live in a wild world where new "threats" emerge practically everyday and their targets are always on the move. They have hodge-podge skill sets, that, between the lot, represents about every ability typically found in the marketing arena (creatives, analysts, technical junkies, hybrids, detail-oriented process fanatics, etc.). They're continually forced to adapt and evolve to stay relevant (aren't we all). This forced growth through adoption combined with diverse skills often leads them into dangerous waters with other disciplines.
SEO's try to take on too much and open their mouths when they should really be quiet. Just like our orc friend here:
...if only he'd kept his mouth shut :).
As search engines advance in their ability to understand and catalog the world around us, SEO's tend to get pushier. Quick side note: understanding team dynamics is more important than ever.
Over the last couple of weeks after having a go-round with a few other SEOs on the topic, I decided to do a deep-dive into the history of social signals and their impact on organic rankings and results. My intention was to round out my thinking and come to an actionable conclusion for myself.
Bonus, my work provides a great resource for anyone who is struggling to understand whether or not social signals play a role in determining content ranking in search engines (specifically Google and Bing). After all, it can be VERY challenging to sort through all the information available and decide what's relevant and what's not. With a little luck, maybe we can.
Let's get started.
History of Social Signals in Search
What follows is a high-level history of social signals and their use in search. I look at what's been said, what's been seen, and point out other important developments in the social & search timeline.
Enough with the background. Are they doing it? Yes, at least for Twitter. Both Google and Bing tell me that who you are as a person on Twitter can impact how well a page does in regular web search. Authoritative people on Twitter lend their authority to pages they tweet. - Danny Sullivan, SEL
The opening quote from Danny Sullivan clearly showcases how search engines were using social signals in 2010. I also recommend reading the Q&A section of the linked article which discusses a number of uses cases for the social data.
It's clear at the time, Twitter especially, was being used to understand people, gather sharing and link data, etc. Google and Bing BOTH had access to a Twitter "firehose" of data, which interestingly included links without the "nofollow" attribute.
A post from Barry Schwartz on SEL on Dec 20 2010 follows up with Danny's post after Matt Cutt's released a video discussing the usage of the social data. Matt makes the following points:
- We do use Twitter and Facebook links in ranking as we always have in our web-search rankings, but in addition, we are also trying to figure out a little bit about the reputation of an author or creator on Twitter or Facebook.
- Blog search, real time search, etc. people have all been using this information. We are studying now how to use this more widely in rankings.
- This is something used relatively lightly for now and we'll see how much it's used over time depending on how useful and robust it ends up being.
Matt also stated that if pages were uncrawlable and PageRank could not be assigned, that scenario would eliminate a post/profile/page from being used. Video Link
Despite Google losing the Twitter firehose of data and real-time search biting the dust in July, it appears (albeit loosely) that retweets, who shares your content, and other social metrics were still impacting rank as showcased by Cyrus Shepard of Moz and others. It's also important to note that followed links are not available to Google from tweets anymore, and social data takes longer to aggregate.
Two related Google projects kicked off/gained traction in 2011:
- Google+ Project Launch
- Near the time when Google lost it's social firehose from Twitter, it officially launched its Google+ project. Are they trying to get access to important data?
- Authorship Launch
- Also in June 2011, Google started promoting its usage of authorship data. In this blog post, they state:
We know that great content comes from great authors, and we’re looking closely at ways this markup could help us highlight authors and rank search results.
- Also in June 2011, Google started promoting its usage of authorship data. In this blog post, they state:
This seems to indicate that something important was lost when Twitter turned off its feed and Google is now attempting to continue on with their desire for social/author/people data in other ways.
- Freshness Update
- Google released a "freshness" update that impacted ~35% of queries in November of 2011.
- Agent Rank, Popularity, Authority, Freshness, and More
In the waning months of 2011, Bill Slawski and others including Justin Briggs and John Doherty begin posting about interconnected topics related to how search engines may be using activities on micro blogs and other social networks to understand an individuals topical authority, the freshness and immediacy of topics, and other subjects. I linked their names to respective posts, but please keep in mind they often wrote on these topics more than once. See the resources section down below for more information.
These posts and related patents/cited sources showcase the past & present desire search engines have to understand and catalog instintanious changes in the real world. Essentially, they need and want to be relevant with what's happening "now".
These individuals do a wonderful job of providing insights into various sides of the problem engineers face in accomplishing this goal and the approaches likely used to solve them. In each of these posts, social signals and the social web/social graph have prominent places in acheiving these goals.
2012 & 2013
I reviewed multiple discussions and articles, but ultimately just kept one article. In the article, Matt Cutts is asked about social signals vs. links. He points out that social signals are likely to become more and more integrated in the future, but that the big impact/change everyone is wanting to know about is likely 10 years down the road.
Matt Cutts is asked a very similar question to the one he received in 2010 and 2012/2013. A few key issues are brought up to explain why utilizing signals from social networks for ranking may be challenging:
- Crawling (consistent data access and signal access) - Matt specifically mentions a period of time when data was unaccessible because they were blocked by a particular social site. If this can happen and crucial data is withheld, it makes it difficult for engineers to rely on that data.
- Cost -As part of the crawling item above, Matt also discusses the cost issue. This is a reminder that Google is limited by finite resources in different areas and they have to constantly balance goals and impact with constraints.
- Change (data accuracy) - Matt also gives an example of how important changes to information could take place almost instantaneously that would dramatically change how that information should be interpreted. This temporal aspect makes the data difficult to trust.
Matt also states that Facebook and Twitter pages are treated just like any other page within the index and are part of the standard index.
Toward the end, Matt does provide a nod toward the future by referencing topics that have been hot over the past few years related to identity, connections between individuals and their content, an individuals authority, etc. The phrasing makes me believe the challenges they face in integrating this type of information are still present despite recent (arguably massive) leaps. Interestingly, he also uses the "in 10 years" timeframe he did in 2012.
This year, AJ Kohn also wrote a nice piece on social in which he makes the case for links being the result of social activity and that being what's driving rank increase - not social signals. This concept is often discussed at Moz when ranking factor survey's are released and is a popular opinion of many in the industry.
It's a tough thing to research, discuss, test, and analyze information and never be able to come to a conclusive answer. The fact is, search engine algorithims are well protected secrets and it's not possible to understand everything at play. Fortunately, we don't have to be 100% positive about how something is working to draw useful insights. Keep that in mind.
I think it's important to provide a definition for what constitutes a "social signal" and even a "social media website". I wish it was black and white, but unfortunatley I believe many of the individuals' who's content I referenced above may define those elements slightly differently. This poses it's own problems when comparing and contrasting, but in general, let's assume:
- Social websites are those with a primary purpose for existence (besides money) to allow for and foster interaction between different people-based entities including businesses, individuals, groups, etc.
- Social signals are any data points that can be collected from social websites including but not limited to content shared and created (including links shared, entities mentioned, etc.) an individuals identity, business identity, and group identities. Social signals also encompass data points that can tie an individual or people-group to any URL through sharing/other social interactions with that URL within a people-group, or a people-group to another people-group, or an individual to a people-group or other individual where a people-group is defined as any brand, or set of individuals grouped under a common point of view, organization, structure, etc.
Now that we've got that covered, let's get to it.
The temporal impact
I believe that social signals, as defined above, can have a direct impact on the order in which results are presented for a specific user's query in Google.
I believe this because social is part of dealing with the issue of freshness. I believe that social signals, as defined above, can be combined with other signals used to determine "freshness" and other metrics to help a piece of content rank well for specific queries. I believe there is enough data to do this and Google has enough consistent access to data to make this happen, albeit imperfectly.
I believe social signals and usability/user satisfaction metrics are connected.
However, I also believe that Google is constantly struggling with the resource battle and making determinations on what information to utilize based on many of the same constraints we all face. I also believe they are working toward a representation of our world that seems completely impossible now. This future version of Google must include social components, as the social part of life is practically impossible to remove. But, for now, access to information, resources constraints, and the sheer problem of scale make this a daunting task.
So, what "social signals" matter most? That's not really the point. In fact, if you're worried about getting keywords into social posts, getting that link shared by so-and-so primarily for the potential of it increasing ranking, I think you're headed in the wrong direction.
Get to the point
MAKE IT EASY TO SHARE
It's VERY important that you consider the content you are trying to rank and the SERP you believe it could show in. I honestly don't think social signals can really help a piece of content move up the ranks for non-temporal queries, or even temporal queries that are aging (their time is passing).
Instead, it seems more likely to me based on my observations and what I've read, that social signals (for now) act more as initial catalysts for getting content discovered and showing an initial public value in relationship to a specific topic. Again, most likely for queries with a temporal component. Remember, topics can become more and less temporal as Justin Briggs discussed.
Finally, I'd like to note that if you look deeper into some of the studies that seem to show very strong evidence for number of followers improving rank, etc., I don't think the conclusions or methodologies really hold up. Sometimes, there just isn't enough data around a study to really trust it.
Do your research, understand how Google is treating different types of queries in your space and the relevance aspects that make up those queries such as time, location, query type (informational, navigational, transactional), etc. The complexity of search, what search engines can do with that data, and what data they can access and how frequently is only growing. Pay attention!
Based on this information and possibly your own conclusions, what does it practically look like to start implementing this knowledge? To help you get started...
For your industry/vertical/specific product/service:
- What makes people discuss it?
- Where do they discuss it?
- Who is discussing it (groups and individuals)
- How do they discuss it?
- How is our business actively involved in the discussion?
- How are we not actively involved (others discussing you)?
- How is content shared in our space?
- What kind of content is shared?
- Where is it shared?
- Why is it shared?
- What aspects (of our industry, etc....) are temporal?
- How are users likely to interact socialy with our content, and specifically, the content on our website?
- What devices do they use to access our web content/apps?
These questions help you build an understanding of the social aspects of your business and content. They should also help you understand in what situations search engines are likely to be considering social signals & metrics and where the usable social data is likely to come from.
A social strategy driven by your over-arching content strategy has become a necessity. The best way to capitalize on the social and search connection is to think strategically and integrate social into the appropriate areas of your business. I know, this sounds like an awful lot of strategy, but the fact is, without this work your wasting time, resources, and opportunities.
If you already have those elements, it's likely you've thought through many of the questions above. Good for you! It's likely you're well positioned to capitalize on the impact of social on organic search.
If these things are knew, walk through the questions, build the strategies, and start getting the social aspect of your business activated.
Social signals have been a hot topic in search for years. While the picture is never totally clear, there's enough information out there to understand that:
- Search engines want social signals
- Search engines are continuously at work to incorporate these signals and metrics into their algorithims
- Social signals are likely connected to queries that have a temporal aspect
As a result, a holistic approach to social driven by a content strategy and strong understanding of your target market(s) is likely to be a necessity for future success. The specific nitty-gritty reasons are not that important. Instead, focus on the user, understand the social aspects of what you do, and let the engines worry about catching up with you. It's the same advice often stated, but rarely listened to. Play the long game.
https://www.youtube.com/watch?v=ofhwPC-5Ub4 -December 2010
http://moz.com/google-algorithm-change - View Dec 2010
http://searchengineland.com/what-social-signals-do-google-bing-really-count-55389 Dec 2010
http://searchengineland.com/google-webmaster-video-reconfirms-use-of-social-signals-59320 Dec 2010
http://googlewebmastercentral.blogspot.com/2011/06/authorship-markup-and-web-search.html - June 2011
http://googleblog.blogspot.com/2011/06/introducing-google-project-real-life.html - June 2011
http://moz.com/blog/do-tweets-still-effect-rankings - July 2011
http://searchengineland.com/google-realtime-search-the-aftermath-of-the-google-twitter-split-84794 - July 2011
http://www.seobythesea.com/2011/10/authority-vs-popularity-in-search-engine-rankings/ October 2011
http://www.seobythesea.com/2011/11/agent-rank-or-google-plus-as-an-identity-service-or-digital-signature/ - November 2011
http://www.seobythesea.com/2011/11/googles-freshness-update-social-media-expectations-of-searchers/ - November 2011
http://insidesearch.blogspot.com/2011/11/giving-you-fresher-more-recent-search.html - November 2011
http://moz.com/blog/googles-freshness-update-whiteboard-friday - November 2011
http://justinbriggs.org/methods-for-evaluating-freshness - around December 2011
https://www.seroundtable.com/google-links-social-15396.html - July 2012
http://www.forbes.com/sites/jaysondemers/2014/06/26/does-google-use-social-signals-from-facebook-and-twitter-in-their-rankings/ - June 2014
http://www.searchenginejournal.com/matt-cutts-explains-important-social-media-signals-really/87277/ - January 2014
http://www.slideshare.net/randfish/google-ranking-factors-2014-correlations-testing-hypotheses - March 2014
http://www.blindfiveyearold.com/social-signals-and-seo - April 2014
http://www.quicksprout.com/2014/01/31/how-social-signals-impact-search-engine-rankings/ - 2014