How can you build a sustainable local news app for immigrants in Germany?
With Newsbriefs, Krishna Prasad, Hari Prasad and their team set out to develop an app with local news for English-speaking immigrants in Germany. At the end of the MIZ funding, we spoke to Krishna about why it is so important to make local news better accesible for immigrants, what challenges they encountered with their project along the way and what tips they can give to other founders of journalistic start-ups.
„We accelerate the social inclusion of immigrants by providing them with local news in the language they know.“
Hi Krishna, your team has developed an app that provides journalistic content for expats in Europe. Why does this target group need their own media product?
Krishna Prasad: Imagine moving to a new country. The language is different, culture is different, food is different. Everything around you is new and different. On top of that there is a pandemic, a Ukraine crisis, energy crisis etc. and you are unable to know what is happening around you because all the major news sources are in the language you don’t know. This is the story of millions of immigrants who move to Europe, filled with hopes and aspirations. We accelerate the social inclusion of these immigrants by providing them with local news in the language they know.
"The topics of interest for immigrants are different from the local population."
Are there topics that are particularly relevant to your target group and, if so, how did you identify these topics?
Krishna: Yes, the topics of interest for immigrants are different from the local population. For example – during elections we are more concerned about the possible changes to immigration policies, during Brexit unlike the local population’s concern about the future of the EU, we were more concerned about what will happen to the Brits who were working in Germany and so on. We use journalistic expertise, but also data to identify the news stories that our readers are interested in and provide extensive coverage on such topics.
Krishna presenting Newsbriefs at the MIZ Role Model Session #10
One of your main goals was to increase user engagement with various new app features. What kind of features did you develop and which of them proved to be most successful?
Krishna: We used a variety of measures to increase user engagement, some of the major ones are:
- We built a mechanism to send notifications to specific user segments and increased the user engagement by 30%.
- We created explainers and timelines to help readers deep dive into topics they want. This increased user engagement by more than 50%.
- We saw very high intent of users to talk about current news events. This encouraged us to include comments section on the app but we are seeing very little user engagement with comments. We are currently talking to our users to identify the reason why they do not use the feature.
- We created data driven content feedback loops to identify relevant news stories. On average every user ended up reading 60% of the stories that we produce, thereby increasing the user engagement.
„We believe AI can help journalists but cannot completely replace human journalistic expertise.“
Another goal was to make content creation more efficient, also using Machine Learning Algorithms. What are your experiences?
Krishna: Our experiments started with identifying the major processes included in content creation:
News identification: We created an automated system that would go through various sources and club the content on the basis of topic, emotion and similarity with each other. This was then passed on to the editorial team. We were able to reduce the time required to identify news by 50%.
News translation: We created a Machine Learning model that works for news. Our accuracy of translation was greater than that of DeepL and Google. We use the translation system together with the team of journalists. We believe with time we will be able to optimize the translation system more and make it relevant to journalistic needs.
News summarisation: We identified Google Pegasus to be the best model for us and trained it on data from European Parliament and other sources. Though we are able to summarise the articles, we are not able to get accuracy because the training data is old and is not able to cope up with the changing global scenarios.
Due to such experiences, we believe AI can help the journalists but cannot completely replace human journalistic expertise.
Screenshot of the Newsbriefs app for Android and iOS
What was your biggest challenge that you didn't expect before you started the project?
Krishna: The biggest challenge that we thought initially was whether we would be able to get users to our platform. But as we went further in the journey, we identified the lack of a sustainable revenue model for journalism. We conducted many experiments around ads, sponsored content and subscription but are still struggling with covering our basic operational costs.
"If we were to bet on the future, we believe quality news with subscription-based revenue is the future to make a journalistic app profitable."
How can a journalistic app become profitable? What did you and your team do to make the app more profitable?
Krishna: It is difficult for journalistic apps to be profitable. We conducted various experiments to reach profitability. These included: subscription (we experimented with various models of subscription and a metered paywall approach is working best for us), Ad revenue (we realized that we compete with entertainment media too, but we did not want to bombard the users with the content just to increase their time spent) and Sponsored Content (this method is better in terms of revenue predictability and user experience control, but too much-sponsored content would dilute the overall news quality).
Until there is a new groundbreaking method to generate revenue, journalistic apps will have to experiment with these various channels and optimize the revenue generation. We also believe there is a positive trend in user’s propensity to pay for news. This is a hopeful situation and if we were to bet on the future, we believe quality news with subscription-based revenue is the future to make a journalistic app profitable.
What are your top three tips for founders of journalistic startups?
1. Focus on the revenue model and Total Addressable Market, it is important to look at revenue from day 1 and optimize your product.
2. In a world where the media supply chain is costly, technology is quintessential.
3. Focus on content, quality, and relevant content will build trust among the users and boost your user engagement.
Marion verantwortet den Bereich Innovationsförderung. Sie ist Ansprechpartnerin für alle Fragen zu den Förderbedingungen, der Antragstellung und verantwortlich für die Betreuung der Projekte.
Zum Abschluss der MIZ-Förderung haben wir mit Victoria Waldersee und Ba Linh Le vom Team LanguageLens über ihr ihr KI-Check-up-Tool und ihre Studie zur Berichterstattung über häusliche Gewalt gesprochen.
Zum Abschluss der MIZ-Förderung haben wir Torben Luckow vom Team Media-Check gefragt, wie man mit einem technisch innovativen Produkt Schüler:innen für Medienkompetenz begeistert.
Wir suchen eure innovativen Ideen für den Journalismus und die Medienbranche – ab sofort könnt ihr euch für die nächste Runde der MIZ-Innovationsförderung bewerben.
Zum Abschluss der MIZ-Förderung haben wir mit Frederik Görtelmeyer vom Team Cultway über ihre KI-Software, Nutzer:innenfeedback und die eigene Rolle als Gründer gesprochen.
In Praxis- und Theorieeinheiten sowie Einzelcoachings könnt ihr vom 4. bis 8. September 2023 wieder die Grundlagen der Radioproduktion lernen.
Wie können Online-Audioformate dynamisch auf Daten und Interaktion reagieren? Wir haben zum Abschluss der MIZ-Förderung mit Vinzenz Aubry über seine Open-Source-Software Gencaster gesprochen.
In Kooperation mit dem MediaTech Hub Accelerator fördern wir wieder innovative Projekte von Mediengründer:innen – von der Entwicklung eines Prototyps bis zur Unternehmensentwicklung.
Zum Abschluss der MIZ-Förderung haben wir mit David Streit vom Team St. Audio darüber gesprochen, warum das Thema Distribution bei der Podcast-Produktion oft vernachlässigt wird und wie ihr KI-gestütztes Tool dabei unterstützen kann.
Wir suchen Verstärkung für unser Team – aktuell eine:n Werkstudent:in im Bereich Innovationsförderung sowie eine Person für den Bereich Office Management.
In fünf Tagen kostenlos die Grundlagen der Radio- und Audioproduktion lernen – vom 20. bis 24. März 2023 geht die MIZ RADIO SCHOOL in die nächste Runde.