Happy face, sad face. Thumbs up and down. Hearts. Just like inventing the alphabet, emojis — the “language of engagement” — created a totally new way of communication.

The emoji was introduced in 1999 when the Japanese designer Shigetaka Kurita made a set of 176 different versions. Kurita was the first to spread emojis on a wider basis. The real breakthrough of the emoji wasn’t until 2010, when they were officially integrated into Unicode, the universal code for all text language. This meant emojis could be seen even if a message was sent from another operating system. When they were launched on iOS devices, they became a natural part of communicating on mobile.

But, already in February 2009, Mark Zuckerberg and his team at Facebook introduced the “thumbs up” emoji as a tool to interact with status updates, comments, and photos to express a positive vibe. Then, in 2016 Facebook extended the possibility to express your feelings with the introduction of “reactions.” Users had the opportunity to press longer on the “like” button to get the option to use five different emojis showing different emotions: love, funny, wow, sad, or angry.

Now, let’s connect the dots between this look in the digital rearview mirror with media companies work on audience engagement in 2021.

To extend the possibility to work in a data-driven way, of course Facebook realised that the more data on user engagement it could get, the sharper its business would be. With more data on the kind of content and subjects users engage with, the more precisely it could expose the right ads for the right audience.

Greater knowledge about behaviour also helped Facebook improve its algorithm, selecting relevant content show its users in the News Feed. The higher engagement with the content, the more likely users are to continue using Facebook and be exposed to more content and ads, creating greater ROI.

Just like Facebook and many other publishers around the world, we at the Swedish media conglomerate NTM use what we learn from data from digital readers to decide what content subscribers should be exposed to. At the moment, we rely on a couple of different data points to run our algorithm, which is called Alva. It decides what to put where on the front pages of our 18 news sites.

Alva’s work with positioning our journalism is based upon knowledge from several different numbers: an articles “news value,” which is on a scale of 1-5 as set by the editor; the decided lifetime for the story; and the number of subscriber pageviews during the latest time period.

In addition to the algorithm, we also use the subscribers’ usage data to understand the audience interest in different subjects. Doing deeper analysis, we get knowledge on what customers prefer and what kind of articles they dislike. What kind of material do we produce a lot of and what do we do too little of? This gives us an understanding on what to change in the newsrooms. For this, we use a KPI on read-through articles, which measures what content actually gets read to an 80% scroll depth.

To get an even deeper understanding of what our audience really engages in, we have been discussing what kind of data points we could add. In close collaboration with our tech department and editorial development team, we decided to try those emojis from the 1990s. If Facebook gets good information from those little cute figures, we can do the same. Copy with pride.

We have also decided to do tests regarding reintroducing the possibility for logged-in subscribers to leave comments on articles. We will begin with one of our news sites but hopefully roll out more throughout the rest of the year.

Both these functionalities are related to our goal of strengthening the relationship between our journalism and subscribers, and also enhance the habit of using our services. The third goal is to enhance the subscribers’ value of our content.

We are only opening up the opportunity to use reactions and article comments for logged-in subscribers. The two new functions will both be placed at the end of articles, so hopefully readers will actually read the article before reacting or commenting, getting more true engagement.

On the front page, visitors will see icons and numbers on how many reactions the article has aroused. Hopefully this will be another driver to increase visitors’ interest in the journalism, besides good headlines and quality pictures.

As you read this in early September, we are about to roll these functionalities out. I’m certain that they will be of great use for us both in developing our algorithm and in our analysis department.

Perhaps these signals of engagement can be taken into account with the Alva algorithm, providing better insight on the journalism that creates a lot of reactions and comments. We will definitely create a visual representation of the data in our reporter dashboard so our reporters can know in real time what kinds of articles cause subscribers to react and comment.

Hopefully we will get an even deeper understanding on what is truly engaging our audience, making it easier for us to increase reader revenue.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .