In a world that is seeing an incredibly fast paced evolution of AI tools, no matter the industry, the role of the human in this evolution is more important than ever. In digital publishing, editors are making day to day decisions on which content to create and how to distribute the content in the most effective way. AI technology such as audience and content recommendations can help them in getting their job done, but newsrooms require actionable insights and controls to trust these technologies.

In order to understand the needs of these editors, we spoke with our digital publishing customers to get feedback on their day to day challenges and explore how technology can help to provide them the solutions they need.

‍Day to Day Challenges

Across the editorial teams that we work with, there were some common challenges that arose in their day to day duties. Primarily:

‍

Which stories should be written

When deciding which articles could be most relevant, editors need to balance a variety of factors including: the value the story has, the target audience, the editorial mission of the company, available resources and competition. 

Engaging and eventually converting fly-by users

On average 50% of visitors to news websites spend less than 15 seconds on the site, indicating that these fly-by users are not truly engaging with the content in front of them. Knowing that subscriber revenue is an increasingly important revenue stream, editors are addressing how to engage these users from the first interaction and how to personalize their experience even though they are not logged in. 

Choosing the right stories to show in a limited space from a huge catalog of content

Newspapers are churning out new articles every hour, making for an incredibly large body of content to work from when choosing what to display. However, not every article is relevant to every user, making the decision on what to show all the more important.

Newsroom concerns about trusting AI to make 100% of those decisions on what to show to a user

Choosing what content to show is always a balancing act, especially when every journalist wants to see their work showcased. However, when leaving this decision to AI over the editor, many do not trust that the AI will show the right article, while also being concerned about perceived lack of transparency around why a certain article was shown.

Zooming in to the challenges of 2 of customers, Mediafin & Hello!, we see some other interesting challenges we are sure other publishers are facing. 

“For us, our content challenges go beyond just choosing the right content from such a wide selection. Choosing which articles to promote or place over the weekend is a particular challenge due to less staff. We also need to consider that we have a variety of evergreen content that should be reactivated to the right audiences. Insights into how all this content is performing online is key to decide what to do for future publications and distribution.” - Mediafin 

“At the heart of what we do is engagement, and how we can increase it, especially as it relates to engagement on pages where our advertisers are present. The variety of engagement times is also very dependent on the type of content, which makes showing the right content to the right person crucial and difficult to do manually.” - Hello!

Solution - AI for Editors

These challenges are nothing new, but they are ever present and a driving force in decision making of new technologies. However, technologies exist to make the lives of editors easier, and understanding exactly how they do this is crucial. Known as editors' insights, these can be a wide variety of things including: how content is performing, how recommendations are being shown and why, audience segmentation and behavioral information. These insights provided by technology can help the human editor to make data driven decisions, but also gives insights that the editorial team and newsroom would not have had otherwise. 

‍

Curated content Recommendations

By now, it is nothing new that AI can be used to choose the right piece of content for the individual user that is browsing a website or mobile application. However, leaving this choice 100% to AI can be scary, which is why a hybrid solution of AI and humans working together can be incredibly powerful. Giving editorial teams the mechanisms to create filters, or to blacklist or boost certain kinds of content based on a variety of factors like age of article or topic allows tailoring the AI to the business needs of the company.

A great example is our “Editors Pick” module. This module allows for an editor to create a curated content pool of articles for the AI to recommend for that specific position on the website. This allows for the editorial team to remain in control of the content being shown, but leaves the tediousness of choosing what to show and when to the AI. 

Next best action: real-time insights

Automating content recommendations are just one element of your distribution strategy, and don’t always help you in deciding which news trends to invest time in or which articles to shortlist as headline, newsletter or push notification. The data sources for editors are endless but consolidating them effectively is all but impossible. An AI powered notification system dedicated to giving real-time, actionable insights to editors so that they can make the right decisions effectively, could help a lot:

• Which articles would perform well on social media, or in my evening newsletter?
• What articles have potential to do well in the home page headlines?
• Which article needs an update, or a better tagline?
• Which audience segments are getting less engaged?
• …

Audience and content insights

One of great contributions of AI for newsrooms is additional insights about an audience that you may already have a solid understanding of. Thinking about content, AI is able to look at all that content your newsroom creates, in combination with what recommendations are being shown and engaged with. This leads to the AI being able to recommend content to write based on the audience that it believes will lead to high engagement. These topics can then be taken by editors and further explored within their existing framework to evaluate ideas. 

AI can also help in segmenting audiences in ways that may not have been considered based on what recommendations are being shown. It can help to identify niche audiences, what content your fly-by users are engaging with, or even habits of these audiences, all of which can help provide more insights for your editorial team. 

Conclusion

In the day to day of a newsroom, editorial decisions are key, and whether they are curated by AI or not,  Editorial insights are crucial in helping to make them. Even with an increase in AI technology, the relationship with the newsroom will always be stronger when used as a tool by the human editor. At Froomle, our mission is to support newsrooms and make the nuisances around using artificial intelligence for content recommendations clear and understandable for all. 

We know that in order to do this and build trust, providing insights into the recommendations being shown, the audiences seeing these recommendations and how your content performs is important. 

So, we want to hear from you! What do you need to create trust in your newsroom, make better decisions and ultimately increase user experience & channel performance? Leave your feedback on our ideas here and help shape the future of smart news!