Background

Extensive desktop research indicates a significant shift in media consumption patterns, with audio playing an increasingly integral role in daily life. Leading publishers like The Washington Post, News24, Nettavisen, and Forbes Italia have embraced text-to-speech services. Notably, Zetland, a Danish publishing house, has journalists reading articles, revealing that approximately 75% of content is consumed through audio

The momentum towards audio news and spoken word is evident among major publishers and media creators, and we anticipate this trend will persist and grow.

We hypothesized that incorporating Audio Articles would significantly impact news consumption, attract more users to Aftenposten, and boost subscription numbers. To examine this hypothesis, we posed the question:

How do Aftenposten readers respond to audio articles?

1. Can we produce audio articles of high quality to entice new subscribers?
2. Is it possible to cultivate a habit among users to consistently engage with audio content?
3. Can we deliver a superior voice experience compared to other publishing platforms?

Design Sprint

Our objective was to ascertain how users respond to the playback of pre-recorded human-read articles in comparison to conventional text and image-based articles. Drawing insights from the Design Sprint problems and hypotheses, we devised a solution: the implementation of a static and sticky player.

To assess its effectiveness, we tested the prototype using an article and an audio recording of "Gutten som aldri fikk fred," lasting 26 minutes and 57 seconds.

Usability Testing

During the interview:

• Context questions about news consumption
• Test prototype including follow-up questions
• Task questions, related to news and audio consumption.
• Deep-dive questions

Participants:

• 6 Norwegians with the following criteria (3 female and 3 male)
• Ages between 30 and 60 years old, living around Oslo.
• Read Aftenposten at least once a week.

Key takeaways:

• They wanted to listen to news not a feature story.
• Article was too long and they didn’t like the theme of the article.
• Users prefer consuming news through the phone and while commuting.
• Users confuse audio articles with podcast episodes.
• Article content determines whether users will engage with the article.

Taking user feedback on player and usability into account, we crafted a player aligned with the Aftenposten brand.

Our approach involved:

1. Ensuring that typography and icons are prominent and easy to read.
2. Streamlining the player to be less intrusive, providing an uncluttered view for users who prefer reading.
3. Regrettably, we were constrained to implementing only a static player due to technical challenges.

Engagement

How can we predict reader engagement with this article? We opted to integrate an audio version on the Aftenposten website.

"Our initial hypothesis suggested that some users might not be interested in long feature article audio. However, we observed a notable approval for audio features in these articles, indicating a substantial user preference for audio engagement."

While Click-Through Rate (CTR) isn't the most reliable metric, as clicks could be driven by curiosity or accidental actions, the average listen length becomes crucial. Analyzing the results, it's evident that those interested in the audio version listen to approximately 75% of the entire article.

Why is a 3,6% CTR considered good?

In discussions with The Washington Post's Senior Product Manager, it was asserted that anything close to 1% is commendable. Currently, we regard a 3.6% CTR as excellent, acknowledging that it may evolve due to the Novelty Effect. This effect occurs when users interact with a feature simply because it's new. Despite readers' inherent inconsistency in behavior and content consumption, a consistent platform fosters trust and can shape new habits around news consumption.

Looking ahead, we plan to monitor the growth and monthly return to the audio player, identifying users who have developed a habit of listening to articles. Additionally, exploring the context in which users prefer to listen to audio articles promises to provide valuable insights.

Future initiatives

Continuously track the growth and monthly engagement with the audio player to identify users who have cultivated a habit of listening to articles. Exploring the context in which users prefer to listen to audio articles could provide valuable insights for enhancing the overall user experience.

Synthetic vs Human voice

Earlier tests focused on pre-recorded human-read articles, raising questions about the optimal number of human voices needed per article. To address cost considerations, we also explored the financial implications of using human-recorded articles compared to automated voice integration.

To shed light on this, we initiated a test to evaluate the significance between two voices: synthetic versus human-read. Our aim is to discern how listeners respond to each voice type, providing valuable insights for future decision-making.

Results

While listeners exhibit a preference for the human voice, the statistical significance observed in this test is 14%, which may not be considered substantial. Anticipating potential improvements, we expect that the introduction of a custom synthesized voice could narrow the gap between human and robotic voices.

“Fake Door” Test

In the "Fake Door" or "Smoke" test, we aimed to assess user interest in integrating a new feature into the subscription model. The test, designed for non-subscribers behind the Aftenposten paywall, presented a user-friendly experience. Users encountered a player information panel leading to a brief feedback form with three simple questions about the audio, avoiding potential user annoyance.

To mitigate disappointment, we considered providing access to a full audio article version after form submission. Technical difficulties prevented us from completing the test before the internship concluded, but we encourage Aftenposten to explore further opportunities for audio articles. Despite not conducting the "Fake Door" test, insights from the design sprint, engagement tests, and A/B testing consistently point towards a promising possibility.

Audio has a future in Aftenposten!

Summary

After our tests, we talked to the Senior Product Manager of Audio at The Washington Post, who has successfully used Text-to-Speech. Our findings match what we expected from our design planning: to make audio a more central part of reading articles, we should make the experience more like a podcast than traditional news reading.

This includes features like curated playlists based on user preferences, the ability to queue articles, and a front page designed with audio in mind. This shift aims to improve how long users stay engaged, increase subscription value, and provide a better experience.

While working on the design, questions about playlists and the front page suggest that Aftenposten might need to create a whole new way for readers/listeners to use its website and app.

Update on 2024

Aftenposten has introduced an audio experience for lengthy evergreen articles on both their website and app. This integrates a preview of the audio content with a subscription-based model.

Enjoy audio articles on Aftenposten’s A-Magasinet.