Hello there,

Let's talk a little about the Eurovision Song Contest. 

In the lead-up to one of the world's largest music shows, the event was already under political tension due to the controversial participation of Israel. But if that wasn’t enough, the Netherlands was disqualified just before the final on Saturday due to an incident involving an event staff member. 

The organisers cited a threat but did not specify what exactly had happened, and disqualification was deemed necessary because the police were conducting an investigation.

Clearly, more information was needed, but neither the Dutch artist Joost Klein nor the implicated staff member have spoken publicly. 

In the information vacuum, traditional media instead started reporting the rumours circulating on social media. A photographer was alleged to have been 'physically' attacked, there was supposed to be a quarrel with the Israeli team - both of which later proved to be untrue. The Dutch public broadcaster responded by reporting a ‘threatening gesture’ but this wasn’t impartial information, and was vague enough to keep fanning the fires of speculation.

Meanwhile, smartocto was on high alert as the servers started smoking. Particularly in the Netherlands, the reach of news sites tripled around the news of the disqualification. Whether each piece could satisfy the news hunger of the visitors is for each title to answer for itself.

Our news needs model lacks a 'Give me all the rumours' category, but perhaps in such situations, there should be a set of protocols in place. For example, a list of questions that an editorial team can use when deciding whether or not to publish, and whether to conduct further investigation:

  • Do we know what happened?
  • Is there a direct impact? (For example, a safety hazard?)
  • Will it affect our audience if we don’t publish immediately?
  • Are the 'what, when, who, where, why' answers factual enough to publish?
  • Are we able to go beyond the speculations made on social media?

In evaluating such large traffic peaks, there’s another, more important question to answer: were we of sufficient value to our audience, such that the trust bond was strengthened rather than damaged?

Subscription boost is a good sign

If you want to answer the last question (whether you have added value), you can naturally look at the conversion rate during the peak days with a subscription model. To be fair, you should consider the conversion ratio, because with three times the traffic, you would need at least three times as many new subscriptions to truly dare to call it a success.

If you are smart, use smartocto notifications to increase conversions. Consider:

  • There is too little premium content on your homepage right now? Does this story qualify?
  • This piece gets a lot of engagement, but not so much traffic. Do a headline test.
  • The number of logged-in visitors is really high for this story. Put it higher up on your homepage.

What’s the deal with Gen Z?

30% of the population is Gen Z, yet 40% of them report that they don’t trust the news. So, what should you be doing to engage this important audience segment? 

Using some interesting data nuggets as a basis, we have some thoughts…


  • Large Language Models are receiving updates, and it seems they are trying to outdo each other, writes the Financial Times. OpenAI has made GPT-4o freely available, and Google announced this week that its search engine will offer new user experiences thanks to its AI program, Gemini, as reported by The Verge.
  • Objectivity is losing popularity. That sounds strange, but it is evident from a survey in the United States. It has never been the best concept for journalism, but Jonathan Stray attempts to propose an alternative for Niemanlab.
  • It is a Dutch initiative, but still interesting to read if you translate it with Google: our (innovative) client Omroep Brabant has employed an AI presenter. She has the appearance and voice of a real presenter but also reads news that is too labour-intensive for the human counterpart to handle.
  • All media face the same problems: once you have chosen a different tool (AI, analytics, or whatever), it is often difficult to get the editorial teams to adopt a new way of working. On the INMA website, we have a blog with tips on how to improve the process.


Hopefully, you enjoyed these reading tips. If you have a reading suggestion that fits our overview, please feel free to share it with us. And to wrap up this Eurovision-heavy newsletter, let’s give a big cheer to our Swiss readers—your song "The Code" by Nemo took the top spot. Well done, Switzerland!

Catch you in two weeks,

Team smartocto