What do audiobook listeners want from narration?

Last week we covered an interesting research finding from Edison Research. We learned that fiction audiobook consumers prefer AI to human narration … some of the time, in some ways.

This week we attended a deep-dive webinar produced by Edison and sponsored by Spoken: The AI Audiobook Company. Hosted by Edison VP Megan Lazovick and Spoken founder/CEO Philip Marshall, It provided details of the underlying research conducted by Edison.

The session was titled “AI vs. Human Audiobook Narration: A Blinded Survey of 1,000 Listeners.” The 30-minute session presented a research study comparing single-narrator (human) audiobooks to AI-generated multicast audiobooks. (Spoken is a producer of AI-generated multicast audiobooks.)

We learned that three key findings derived from this survey:

• Most listeners could not identify AI narration in blind listening tests
• Purchase intent between the two production types was tied at the start
• Acceptance of AI audiobooks (measured as “willingness to listen”) rose after exposure to AI audiobooks.

That result on its own is fascinating, revealing that initial hesitation about AI audiobooks might be less about an objection to AI-produced audio generally, and more about unfamiliarity.

Supporting that idea, we learned that 70% of respondents like hearing distinct voices per character — that number rises to 81% among more experienced (“frequent”) listeners. That data point leans toward AI presentations, which can deliver a cast of multiple voices more easily than hiring multiple voice actors.

But moving beyond the good-and-bad of audiobook narrating, another indicator of listener values and priorities is this finding: “Quality and immersion outrank cost (44%) and celebrity voices (38%) as drivers of AI interest.” 

Edison’s research captures these voicing issues in a concise graphic:

Well then … what DO listeners want from audiobook narration effects? Edison discovered five priorities as shown below; we note that the top three are about voice, and the bottom two are about background audio.

Another key finding, and bad news for single-voice human narrators, is this: Frequent audiobook listeners gravitate to multi-cast narration. That trend leans away from the traditional human single-voice style … and trends toward AI-generated multi-voice performances as produced by Spoken.

Whatever the narration source, we learned that survey participants value narration quality and immersiveness more than low cost-to-listen or celebrity narrators. There are many motivators which pull listeners to AI narrations, as the chart below indicates, and the top two are production quality and immersiveness.

Several other detailed findings are in Edison’s report (which is HERE).

---