Content recommendations engines – a new way to reach new audiences

A briefing on Outbrain with Simon Edelstyn

All users of retailers are familiar with the Amazon recommendations on books, but personalisation and recommendations services aren’t only relevant to retailers. This interview shows that recommendations engines are also increasingly used online by publishers and bloggers. Publishers are using these tools to increase engagement with their site while also increasing revenue through serving paid links to relevant sites.

I was interested to hear of this tool since it’s not only relevant to publishers, but also provides a way for marketers to increase reach and engage an audience through the type of content consumed based on a cost-per-click (CPC) model. Think of it as working similar to the Google content network. Here's an example, from the footer of Techradar which uses this service.

In the interview that follows I ask about the marketing benefits, users and technology behind Outbrain.

What are content recommendation engines?

Q. Please give us the “elevator pitch” - what category of service do you offer site owners and what types of benefits will they see. Who is using it in action?

Just as Amazon recommends books that are likely to be of ‘interest to readers’, so Outbrain recommends news, blogs and other interesting piece of content that are personalised, and so more relevant to the individual reader.

Outbrain’s content recommendation technology uses a complex set of algorithms to serve up links to content which are more likely to be of particular interest to each individual reader’s reading preferences, taking into account content and other traditional techniques which are generally called “related links”.

The result is higher page CTR’s, more engaged users and an automated system that augments the editorial recommended links.

The Outbrain platform also enables publishers and content owners to distribute their content across the Outbrain publisher network and so increase traffic to their own sites and to generate revenue. For marketers Outbrain offers a way to distribute content alongside publishers own editorial recommendations (on a pay per click basis) – delivering interested and engaged readers to marketers’ high quality editorial online.

To see Outbrain in action, please visit Tech Radar or Daily Express and navigate to a news article. At the bottom, you’ll see a section with a variety of links – both internal and external – these are powered by Outbrain with the internal links helping to increase click through rates and engagement, and the external links generating revenue for the site, on a PPC basis.

Examples

Q. Which sort of companies or websites use your widgets? Can you share benchmark figures which show improvements in terms of increasing engagement or revenue?

Outbrain is the leading content recommendation platform installed on over 72,000 sites including most of the world’s most popular brands including CNN, New York Times, Fox News, AOL, Elle, USA Today, The Daily Express and Tech Radar.

Outbrain can be used by any publisher, content creator or marketer. It is well established in the US working with publishers such as AOL, Elle, MSNBC, USA Today and The New York Post. Brands working with Outbrain include AmEX Open Forum, Ford and British Airways.

In the UK, where we have only been working since the start of 2011, Outbrain is now live on over 70 premium sites, and we work with ten premium publishers including Future Publishing and The Express Group.

Outbrain’s recommendations (both editorial and paid) have become a standard navigational tool following content for millions seeking additional quality content as they browse the web. For publishers as a direct result of helping their users discover more interesting content online, there are clear business benefits such as stickier sites, higher levels of engagement, increased traffic and page views, and a new revenue stream.

A recent study we did of clickthrough rates on headlines shows some techniques for increasing clickthrough which may be interesting to copywriters:

• Length matters: A title with eight words performed best. These titles received a 21% higher clickthrough rate than average.
• Pics get clicks: Outbrain’s data analysis showed readers are attracted to content with images. Thumbnails worked better than logos and increased clickthrough rates by 27%.
• Lists with a twist: Lists and photo galleries typically drive more page views, but they get more clicks as well. Headlines that contained odd numbers had a 20% higher click-through rate than headlines with even numbers.
• Going deeper: A colon or hyphen in the title — indicating a subtitle — performed 9% better than headlines without.
• Questions work: Titles that ended with a question mark had a higher clickthrough rate than those that ended with exclamation marks or periods. If you really feel the urge to use an exclamation mark, our studies have shown that three (!!!) received almost twice as many clicks as all other punctuation marks.

Outbrain technology

Q. So how does your technology work to make recommendations?

Outbrain’s content recommendation platform mixes contextual analysis (articles with similar content), collaborative filtering (people who read this also read), popularity (most visited articles), sharing trends and personalisation (recommendations based on previous behaviour).

Understanding that interesting and related are not always the same thing, our technology works to help readers discover quality content that they may never have otherwise known existed.

The approach to recommended content to users online is often far too simple and generic to be of any real use. For instance, publisher sites listing links to their most popular articles often just shows people what they have already read. Or attempts to provide people with links to “related content” fall flat, because reading about the same subject twice is not always appealing.

At Outbrain, we are focused on re-creating the serendipity of leafing through a magazine or newspaper and falling upon an article of real interest that you didn’t know was there before – meaning using technology that allows users to discover content which is interesting to them, not necessarily simply relevant to the previous article.

Publishers therefore can truly provide their readers a unique reading experience geared at individual likes and reading style. In a world where people are often overwhelmed with unappealing content choices, Outbrain helps readers discover content that is perfect uniquely for them.

For marketers, this approach means they can amplify their own content (either created but them, or earned via PR or positive reviews) to reach a targeted, relevant and engaged audience online.

Implementation best practices

Q. Which best practices in terms of layout, creative and testing do you recommend to site owners who implement your service? I’m particularly interested in rules of thumb that readers could consider for their sites.

Outbrain can be adapted to suit each individual site adopting the same looks and feel so it is fully integrated within the page. Publishers can also choose the number and mix of internal and external recommendation they offer through Outbrain to help meet their own business objectives, be that developing a new revenue stream, or increasing internal engagement within their own sites.

However we recommend that the Outbrain widget is installed at the foot or the article page as readers who are engaged in reading content generally like to finish reading before moving on to further content.

The core Outbrain service widget is free for any type of publisher and can be easily installed.