Study Gives Insight Into Content Discovery Trends Across the Web’s Leading Publishers

|Kelly Reeves
External Traffic sources

We’re excited to announce our inaugural study on how people discover and engage with content on the web!

In the first quarter of 2011, for the first time ever, we decided to analyze our expansive data and share it with the industry to help inform conversations around content discovery. And now we’re ready to share it with everyone.

To do this, we looked at traffic patterns  from 100 million sessions across more than 100 premium publishers that are currently using our platform to see how readers are accessing content, where they’re finding it and how they’re engaging with that content. We’ve compiled this data into our inaugural report, and our hope is to use this as a benchmark against future quarterly trend analysis.

Let’s set the stage: Methodology

In order to set a baseline for future reports, we looked at the referral data into publisher articles to see which sources are currently driving the most traffic, as well as the level of engagement and quality of the traffic coming from these sources. For this study, we pulled a sample data set of 100 million sessions across more than 100 of Outbrain’s top publishers. A session is defined as a series of page views within a publisher’s site with no more than 30 minutes between one page view and the next.

Since Outbrain tracks traffic to content pages (articles and videos), we saw that approximately 33% of the overall sessions into such pages start from an external site. The remaining 67% of content sessions begin internally, from type-in traffic, bookmarks, clicks from the publisher’s homepage, other in-site links or are simply from unknown sources. For the purposes of this study, we evaluate only on the third of sessions that begin offsite in order to focus on how people discover content from outside sources.

Key Findings

  • While search still reigns supreme in terms of directing traffic to content pages (41% of external referrers), social is gaining share at 11%.
  • Of the six content verticals examined, stories in the news, entertainment and lifestyle categories are the most likely to receive traffic from social sources.
  • Traffic coming from social media sources has the highest tendency to bounce.
  • Readers who go from one content site to another (i.e. USA Today to The Daily Beast) are most likely to be engaged in what they’re reading, presumably because they are already in content consumption mode.
  • Facebook delivers a more diverse audience than Twitter.

Top Traffic Sources

According to our data, the following sources were the top referrers of traffic during Q1, by number of sessions:*

Top 20 Referring Sources to Content Pages

Top 20 Sources

For the purposes of our study, we are focusing on known external referral traffic to content pages, which accounts for 33% of our sample set (direct, in-site and unknown traffic sources account for the other 67%).

Here is a breakdown of referring sources in terms of the type of referrer represented:

Breakdown of All Incoming Traffic Sources

Excluding direct and in-site traffic, an adjusted breakdown of traffic share is illustrated by the following pie chart:

Breakdown of Known External Traffic Sources

Currently, search methods (including Google, AOL Search, Bing, Yahoo and Ask) send the largest slice of referral traffic to content. Links from publisher sites make up 31% of referral traffic to content pages (this includes manual partner syndication, linkswaps, Outbrain recommendations of content to content, etc.), portal homepages (,, account for 17% of traffic, and finally, social media sites (Facebook, Twitter, StumbleUpon,, reddit, Digg) send 11% of traffic to content pages.

In subsequent quarters, we plan to pay special attention to trends in this category in order to evaluate how people’s online behavior may be altering content consumption patterns.

What type of content do people share?

It’s no secret that people are spending an ever-increasing amount of time on social sites like Facebook and Twitter. One of the by-products of this shift is that these same people are now relying on their networks of friends and peers to alert them to interesting news and content. We thought it was worth looking at how this breaks down among the different content verticals. News stories, at 42%, were the most likely to receive traffic from social sites, followed by entertainment stories at 30% and lifestyle stories at 13%.

Breakdown of Social Traffic Referrals

Reader Engagement By Category

In addition to looking at which sources deliver the most traffic, we also analyzed how those sources stack up against each other in terms of delivering the most engaged audiences, as measured by metrics such as average page views per session and bounce rates.**

Average Page Views Per Session

Content sites have the lowest bounce rates, presumably because they are targeting an audience that is already engaged and in content consumption mode. Traffic coming from social media sources, on the other hand, has the highest tendency to bounce.

Bounce Rate Per Session

Delving into the topic of engagement a bit further, we looked at which sources are most likely to refer a hyper-engaged reader, which we defined as a reader who views five or more pages per session.

Hyper-Engaged Reader Sessions by Traffic Source

Readers referred from publisher sites — from content to content — are likely to consume more pages than those referred from other sources. Readers referred from search queries consume slightly more pages than average, though Yahoo, whose sessions tend to be twice as likely to be hyper-engaged than the average referrer, skews this number.

Comparing Facebook and Twitter as Referral Sources

Facebook vs Twitter

Given the popularity of both Facebook and Twitter, we thought it was worth comparing their relative traffic quality to see what differences exist. Surprisingly, the two sites drive similarly engaged audiences in terms of page views per session, bounce rates and hyper-engaged reader sessions. The one key difference is in their relative reach, which we define as the number of unique visitors per 1,000 sessions. Specifically, we found about 72% of sessions originating from Facebook were from a unique visitor, versus only 52% in the case of Twitter, suggesting that Twitter’s audience is more likely to be made up of repeat visitors.


It’s an exciting time in the content discovery space as larger trends in web usage play out in the ways which readers access and engage with content. Though traditional methods like search still reign supreme, we’re keeping a close eye on new trends such as social sharing and the increasing openness of content sites to link freely to one another. From a publisher standpoint, it’s particularly important to recognize and understand these shifts in order to identify the best way to get your content in front of readers. In future reports, we hope to highlight changing trends and track them against our baseline data.

*Results may be skewed towards news and entertainment sites, as they constitute more than 50% of the publishers working with Outbrain.

**Average page views per session is defined as the average number of pages visited by a user during his or her session. Bounce rate is defined as the percentage of sessions that only lasted one page view.

At Outbrain we strive to utilize open source projects wherever possible — we’d like to thank The Apache Software Foundation and Cloudera for their brilliant work on Hadoop and Hive and for helping us handle our “big data” needs, without which our Q1 report might not have surfaced until sometime in Q3.

Hitting "Publish" is only The Beginning!

Download our free A-Z Guide to Brand Publishing to find out how to make the most of your content strategy.

Kelly Reeves

Kelly is Senior Marketing Manager at Outbrain. Previously, she was the founding editor of Urlesque, an AOL-owned website covering internet... Read more

Add A Comment

* Your email address will not be published. All fields are required.

  • Toshiba Satellite L300 Battery| April 15, 2011 at 9:09AM

    Read many similar articles, but you write the best, thank you for sharing, I hope I can see more.

  • TOSHIBA Satellite M70 Battery| April 15, 2011 at 9:09AM

    Generally I never read whole articles but the way you wrote this information is simply amazing and this kept my interest in reading and I enjoyed it. You have got good writing skills.

  • Lenovo 3000 G410 Battery| April 15, 2011 at 9:09AM

    Such a great posts and I find Disqus a very good commenting plug-in. You can always check back your comments because Disqus compiles all your entries…so, it’s easier to see and follow your comment

  • Aisha| April 18, 2011 at 8:20PM

    thanks lads…I very much enjoyed your report and would hope that you do same on a quarterly basis too, eh!

    • Kelly Reeves| April 28, 2011 at 9:21PM

      Thanks Aisha! That’s the plan. Check back after Q2 for an update. Follow us on Twitter to be alerted to the new report: or subscribe to the blog RSS above.

  • Dave Drager| April 19, 2011 at 4:16PM

    Great – love the stats! Hope to see more in the future.

    • Kelly Reeves| April 28, 2011 at 9:21PM

      Thanks, Dave! We’ll be updating this report in particular each quarter, so be sure to come on back.

  • Ed Boudrot| April 25, 2011 at 2:14PM

    Great resource! thank you.

  • R Fleck| April 25, 2011 at 7:19PM

    Fine presentation. Thanks! We’re a content site and will study your report….a lot.

    • Kelly Reeves| April 28, 2011 at 9:21PM

      Great! Let us know if you have any questions at all. Happy we could help.

  • David Evans| April 28, 2011 at 8:20PM

    Would it be possible to publish or somehow send out the Y-axis labels or some form of quantitative data for the last 3 graphs (page views, bounce rates, hyper-engagement)? Without them, I’m having trouble interpreting the size of the effect. Thanks!

  • Joel| May 4, 2011 at 4:16PM

    Awesome article. I want to use this in an article that I’m writing. I’ll shoot you an email

  • Una Grant| May 5, 2011 at 4:16PM

    Really interesting report – I teach consumer insights and market research and this will definitely be course reading for my students in September – looking forward to the Q2 update

    • Lisa LaCour| May 5, 2011 at 6:18PM

      Very cool Una. Thanks!

  • Victor V| May 6, 2011 at 11:23PM

    great article. do you have a country by country report? I am studying how the Filipinos are responding to travel e-commerce business

    • Lisa LaCour| May 10, 2011 at 3:15PM

      Hi Victor, We are working to include data like this in the future. Thanks for reading!

  • Bonjour Tristesse| May 7, 2011 at 12:12PM

    Great information, thank you for sharing this data. It would be nice to see updates at regular intervals.

  • Lasting Rose| June 4, 2011 at 4:16PM

    Very good report and thanks for publishing. Wish there were more like this from the owners of data like Google.

  • Lasting Rose| June 4, 2011 at 4:16PM

    Very nice report and thanks for making it available. 

  • qmlmplnx| July 4, 2011 at 4:04AM

    Im no professed, but I suppose you just created the top point. You unquestionably know very much about what youre referring to, and I can really get behind that. Thanks for getting so straight and so honest in regards to the topic matter. I truly feel like We?¡¥ve a much better reading now.

  • Viola| July 21, 2011 at 11:23PM

    Great report.  Would like to know what the average page view is (graph has no y-axis)

    • Kelly Reeves| July 22, 2011 at 5:17PM

      Hi Viola, Thanks for the feedback! While we recognize the value of the raw numbers, it is our responsibility to respect the sensitive nature of our clients’ data, first and foremost. Even in aggregate, it would reveal baselines, and potentially be unfairly interpreted as indicators as to the performance of our client sites. Instead we are comparing each traffic source to the average. Feel free to email me with any questions you may have:

  • Web Traffic Report| September 9, 2011 at 2:14PM

    Great graphs! it’s really nice to see what’s happening on the web trends. Thanks for sharing this one.

  • Regina| September 24, 2011 at 4:16PM

    I am looking forward to sharing this report. It is nice to have written back up to my pitch. I have always expressed that a complete social media marketing plan should include more than Facebook and Twitter. Part of my initial setup package is to give the business a profile on search engines and directories. I review their website for content and recommend enhancing profiles on professional and service related sites. Thank you, Regina Kempster

  • Ieshoes Com| October 20, 2011 at 1:13PM


  • kimitris aselios| January 7, 2012 at 5:17PM

     I found so many interesting in your blog especially its discussion. 

  • SEO writer| January 11, 2012 at 6:06AM

    Interesting report Kelly. But how did you find out that “Facebook delivers a more diverse audience than Twitter”? I mean what was taken into consideration while concluding this?

    • MCX-Expert-Tips| April 23, 2012 at 5:05AM

      i agree with your facebook delivers a more diverse traffic compare to twitter.

  • MCX Today Tips| April 17, 2012 at 6:06AM

    I am looking forward to sharing this report. you have create a good report. Thanks for sharing this post

  • Curtman40| March 14, 2013 at 12:00AM

    Content Marketing I think it is here to stay those that master it will see benefits from it not just profits but a sense of community which is what we need to be thinking about when we generate our content.


Want to promote your content?