Recently Yandex, Russia’s main search engine, revealed that its algorithm has evolved to be able to serve up regional search ranking results without looking at backlinks. For a practice like SEO that has long been built on the “”links-as-currency” model, this move will likely be watched closely by other search engines like Google in 2014.
The original PageRank algorithm that ran Google was innovative. Its ability to assign authority to pages based on who linked to those pages and who linked to the pages that linked to those pages was novel and proved very effective in the beginning. However, the ability for this model to be exploited by black hat SEOs, spammers, domainers, and anyone with the ability to hack and scale links made the original algorithm an imperfect one.
Google has been making its own transition to more “natural” signals of influence and site popularity. Google has put much more weight on social signals—link sharing activity and brand/website mentions across social networks such as Twitter, Facebook, LinkedIn, Google Plus, etc. While social signals are still link-based (a link from Twitter to a blog post, for example), they’re sometimes a better, more immediate indicator of quality than traditional backlink metrics are.
So how important have social signals been to SEO success?
For one, most of the supporting data we have about social signals is correlational. High-ranking content happens to have more social signals than low-ranking content, but this doesn’t control for other factors. A large news website may have a solid distribution infrastructure and higher site authority due to traditional links, so its content ranks better, more people see it, and more social sharing naturally occurs regardless of the content quality.
What is “Normal” for You?
Google’s machine learning has really perfected the art of identifying what is “natural” in the world of links and social shares. Google has been straightforward in naming social signals as a ranking factor, since normal businesses – good businesses – have products and services or content worth talking about online. But these online conversations are more or less important to rankings depending on the industry. More social conversations are expected around consumer electronics than financial investment services, so most data suggests that Google ranks and rewards websites relative to their respective industry norms (which is good if you’re in a “boring” industry).
Local SEO is a good domain for comparison when thinking about social signals as ranking factors. While the number and quality of reviews on a business’s Google Local page is a local search ranking factor, Google bases this on the industry. Restaurants naturally accumulate reviews faster than cleaning services. So you only need to be concerned with the reviews of your competitors rather than having some specific number of reviews to get.
- To-do: determine what is a typical amount of social sharing for your industry by comparing product pages or sampling blog posts of your competitors and measuring the amount of Facebook shares, Tweets, and Google Plus shares each receives. This can serve as a baseline for comparison, and can be done either manually or with the help of SEO tools like Ahrefs.
Social Signals Will Never Be Enough
SEOs have been bantering more feverishly lately about a future “link-less” Google, where all rankings are based purely on content entities, “co-occurrence,” and social signals (let’s not forget that social signals are also link-based, but that’s a discussion for another day).
What I’ve observed over the last several months on multiple websites is that social signals are sometimes more “shiny object” than “silver bullet.” What I mean is that the core keyword optimisation factors are still the primary determinants of what rank. Overall domain authority is also weighted heavily and a website would be hard-pressed to build a substantial amount of domain authority on social signal links alone. It just doesn’t work that way.
As a result, I’ve seen instances where a large number of social signals does very little for SEO and rankings. I’ve created content that garners hundreds of natural shares on Facebook or dozens of shares on Google Plus that still doesn’t rank well, even for moderate/low competition keywords. What many SEOs think is a major ranking boost tied to social signals is, in fact, a ranking boost tied to an increase in traditional backlinks due to social discovery by content creators. My thoughts are corroborated by some very detailed experiments by Eric Enge and Stone Temple, who found very little correlation between Google Plus shares alone and organic rankings. It’s doubtful that in 2014, Google will decide to switch to a completely linkless algorithm.
Social Signals Will Still Be Critical
Let’s be clear about what social signals do provide, in additional to exposure and reach.
- Legitimacy – Social signals provide meaningful information to search engines about the nature of sites and companies. As search engines seek to eliminate web spam, social signals add an additional layer of proof that a website deserves to rank as a whole because it acts as an extra barrier to entry for illegitimate sites to gain favor.
- Differentiation – Among legitimate companies in an industry, social signals provide a way for search engines to compare how companies are received by their customers and the marketplace. In the perfect algorithm, Google would rank the best product or service at the top. But in the real world, Google uses proxies for product/service/company quality, including word-of-mouth, which it can measure through social signals and traditional links. Two sites with seemingly equal site authority and on-page optimisation are likely to be compared on secondary measures of quality like this.
- Freshness power – One area in which social signals do impact SEO greatly is in Google’s ranking of new content or fresh queries. New news and content tends to generate a lot more natural conversation online, which is likely to put greater onus on social signals to impact SEO.
2014 Prediction: Improved Social Spam Detection
A best guess based on Google’s webspam and algorithm history is a focus on improving social spam detection in 2014. They’ve been in the business of improving webspam detection for over a decade. There’s no good reason to abandon this in favor of a linkless model. But Google’s ability to detect social spam still needs work. The rabbit hole of automated, fabricated social profile networks is deeper than ever (a week ago, I ran into a huge network of fake people with fake Facebook profiles, Pinterest accounts, and Twitter accounts – all with almost natural-looking activity). As Google’s social spam detection improves, we’ll likely see changes to the algorithm that will make it appear that social signals are given more weight. The result will be better spam detection, but not necessarily greater social signal weighting for SEO.