HOW TO PURSUE WORLD PEACE THROUGH RELEVANT ARTICLE RECOMMENDATIONS

By March 21, 2016Uncategorized

Well no, this post won’t be about that. But is it relevant? Keep reading…

As a former editor I’m aware of how we felt back in the days when we used to be the gatekeepers to breaking, interesting and relevant articles.

As always when you think back in time the romantic filter tends to level up a notch, but times were good. It felt easier to add value to the site for your group of visitors, the editor’s whole spectra of instinct was to good use, the visitors were more loyal and hey, the screen was bigger.

Engagement was, and still is, The Golden Child that we, in all ends of the production chain, need to nurture and grow together. We pick up our phone approximately 150 times a day – youngsters about 300 times a day, reports claim. Maybe 60 of those times it’s to use other apps, text or calls. Yes, some of us still use our mobile devices to call people. But the rest of the times compete to quench our thirst for immediate information.

The phone is used in quick manners, as we all know. The finger play runs like over a piano, and the pace is unflattering. It’s not easy to gain engagement – information is consumed so fast that we hardly read the whole article before losing interest. It’s a part of the small format. Images and videos are more fun and easy to engage in. Where does that leave us, as publishers?

In this context The Golden Child is not kidnapped by the evil Sardo Numspa, and humankind’s fate does not hang in the balance (sorry, you have to know the movie to get this one). But the struggle is real. Engagement is hard to maintain because of the concept of ”mass media” – that we’ve always used in a certain way. It was supposed to address so many demographic interests that it actually became impossible. How can we keep the interest of a “mass”? Essential tools such as article recommendations are crucial, and after just studying the top five news sites in the Swedish market it’s easy to see that they’re approaching different strategies in their recommendations. A lot of sites are stuck in only recommending related articles, which may have worked before, but it’s time to realize that this has changed.

As a natural reaction to that old way of thinking regarding target groups as big masses with the same profile, the overall trends in all reports point at individualization and personalization in all fields. As a publisher, site owner, editor – how can we address these individual needs and personalize their user experiences? Can we do this without interfering with the editorial master plan of a site? How can we make ”Internet” relevant again?

The dilemma as I see it today (please argue with or against this), lies in how ”relevance” is defined. The visitor often has a predetermined thought when he or she enters a site. They want to be updated on something specific, or probably just followed a link in a certain context (Facebook, Twitter, LinkedIn). After the first click – their mission is accomplished. How can we keep their attention after that fulfilment? Here’s where relevance comes in, and ”related” steps back.

What qualifies as a relevant recommendation? The answer is not as smooth as the question.

In Schibsted’s Future Report 2016, there’s a chapter called ”Serendipity, in search of the human algorithm”, where Google’s new feature (back in 2009) was introduced – search personalization. In an attempt to make the information more relevant to a searcher, Google saw that the most important task was to give different results to different people with the same search. This means that the same key words generated different search results for different people.

So again, what qualifies as ”relevant” content recommendations for a mass of people versus individuals? It’s interesting to put Google’s angle into an editorial environment. To twist our minds just a bit further, important questions regarding ”relevance” and ”related” are; which articles and ads are relevant for a visitor? Should the subjects match the editorial context or the unique visitor’s user data? Is there an intrusion of the brand or identification of an editorial site, if it recommends content that doesn’t match the article as ”related”? I’m throwing out a few questions here that I would love to see some comments on; I won’t debate them myself in this already bursting space.

I believe that we have to recognize the continuously changing environment by focusing on how to add quality, not on what we lose (control is for one already lost). And we’re using recommendation engines wrong if we make them cut opportunities for the visitors to form new connections on individual relevance. We need to design systems that grant the opportunity to further reading in the relevance of the subject, and also the value to discover useful articles that carry us out of our own bubble.

Our mission is to extend the universe for a visitor. Therefore the question stands: How can publishers design a relevant recommendation system for their readers and create killer engagement? …And the answer? It probably changes as regularly as our digital environment.

Ruxandra Bocaciu, Head of Publisher Relations Sweden