Publishers, your shares data is wrong (part one of two)

Social is the new SEO.

SEO is still a major traffic driver for publishers, but search is search is slowly losing its grip — constituting less than 50 percent of traffic sent to publishers — to an exploding number of social referrers. But, as publishers turn to social insights to optimize content and fuel site revenue, they run into a big problem. The data that they’re looking at to make these high and low-level decisions about social is inaccurate.

As the CEO of Parse.ly, a data and analytics platform provider for large-scale content sites, I’ve worked with teams from some of the biggest online publishers to resolve their data challenges. Social shares comprise a new data source that presents technical complexity specific to this market, and I have watched my company’s customers struggle with poor data leading to the wrong decisions.

Historically, publishers lamented over “fly-bys,” readers who discovered a site through a search engine, read the article, and then left without recognizing the brand or ever coming back. Though the explosion of social won’t directly convert these users, it does give publishers a whole new way to build engaged audiences that can translate to a decrease in fly-bys in the long run. With social, the strategy changes from passive (SEO) to active engagement, and that can have a dramatic affect on a user’s ability to recognize a publisher’s brand.

How does this happen? Well, it is certainly not just about tweeting all the new articles that you post on your site. Instead, it’s about understanding how your social audience is engaging with your content on and off your property. Publishers should be able to answer such questions as: What type of content resonates best with my audience from Facebook? Who are the top influencers of traffic to my site from Twitter? And if I focus on Pinterest as a social strategy, what type of traffic can I expect out of it?

The best publishers are getting even more savvy and looking at the virality of individual stories to judge success. And then extracting the viral characteristics of successful stories to make editorial decisions that hopefully continue to tickle reader curiosity and keep them coming back. For example, Buzzfeed wisely engineered a way to backfill Facebook timelines with “nostalgia-drenched” content through posts on their site, catering to readers’ desires to digitally re-live memories from an era before Facebook.

More and more successful publishers are dedicating resources to crack the nut on social the same way they did with SEO. The potential upside locked in social is huge.

Let me say it again: Unfortunately, your shares data is wrong

Simply put, the margin of error for shares data is unacceptable, and publishers are putting themselves at risk by making important decisions based on shot-in-the-dark conclusions.

On most publisher sites you see share buttons with the number of shares associated with each of the large social networks. Most publishers assume that these numbers correctly represent the number of shares associated with a particular post. However, there’s much more than what meets the eye.

The underlying problem with share buttons is that they only represent the number of shares associated with a single URL. For small sites that’s not a problem. But when you’re dealing with mid- to large-sized publishers, one URL does not equal one post. To explain why this is the case we have to go into a bit of history.

History of the publisher URL structure

Due to the way content and the Web has evolved, online publishers have been toying with the URL structure of their article pages for years. This has led to a slew of important “link alias” patterns for a variety of content types, including traditional news stories, articles, and slideshows. Here’s a list of the different subsets of URLs that publishers tend to deploy:

Mobile versions, which tend to use m.domain.com
Multi-page stories, which tend to have mulitple URLs all referring to the same story
“Single-page” and “print-friendly” versions of a long-form article
Comment threads
Article updates (such as SBNation’s “story stream”)
Alternative slugs for SEO testing
Shortened URL redirects (such as bitly) for the purpose of squeezing into Twitter character limits and tracking click-throughs

Even worse, social networks do not agree on how to normalize a URL. For example, consider a link to this article at Gizmodo: http://gizmodo.com/5909215. Should a link to http://www.gizmodo.com/5909215

be treated separately (notice the www.)? How about a link to http://gizmodo.com/5909215/ (notice the new trailing forward slash)? How about a link to http://gizmo.do/KpJhhN (Gizmodo’s bitly redirect to the same story)?

Many publishers have as many as 50 different URLs that all point to the same “piece of content” on their site. But the only number they look at is the point-in-time share count for a single URL provided by the official share buttons, which often drastically undercounts the real number. In some cases, a publisher could be missing a majority of their shares data because not all URLs are fired against the social networks. Any shares data, no matter how trivial it appears, is still important, because a single share can translate to multiple page views. If you don’t have the right data, you can’t make the right decisions.

The editorial importance of accurate shares data

Many publishers find themselves crossing their fingers and hoping for a story to “go viral.” If a publisher is waiting for a story to go viral, they haven’t pulled out a microscope to examine their social strategy.

With accurate shares data in-hand, informed publishers can start to answer really interesting questions (such as the ones referenced at the beginning of this piece) about their content and the behavior of their readership.

Further, if publishers start to keep historical data about shares, then they can start doing trend-based analysis across the data to gain even deeper insight. Real-time shares data is important, but it doesn’t tell you, for example, which stories are increasing in their velocity of shares (10 shares in the past ten minutes versus 20 shares in the following 10 minutes). Even more, with a historical look you start to understand the cause-and-effect of why stories did well and what led to their success (in-house efforts, other referrals, organic, partnerships). Since most of the current social network APIs don’t support this type of data, publishers would have to either build the infrastructure to capture this information or leverage a third-party service.

Diving into social data is incredibly important. It can dramatically influence the voice and content strategies of a publication. This is why it’s absolutely critical for publishers to take social-as-a-strategy seriously. Tomorrow, I’ll discuss solutions to the shares problem as well as larger initiatives the industry can take to reconcile the issue.

Sachin Kamdar is an NYC entrepreneur and the co-founder and CEO of Parse.ly. Parse.ly provides insights to online publishers through its flagship product, Dash. Dash captures publisher specific data to present the best opportunities for engagement and to improve key metrics for publishers.

Image courtesy of Flickr user KungPaoCajun.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.