More Analysis of Spoken (Press)

Back in April I wrote a review of Spoken.press, which is an AI-powered audiobook platform. I experimented with creating an audiobook for Running From Herself, just to learn how Spoken's product performed. The site is still in its beta phase, so while people who are aware of it can listen to audiobooks, the company is hardly well known. I only learned about the site through an article sent to me via one of the book promo companies. 

I found Spoken's process to be extremely cumbersome. I was required to go into my manuscript and add special tags to separate chapters (and the book has 40 chapters!) before uploading. I suspect this was the culprit that ended up transposing some of my text, which required multiple re-uploads to KDP, although I can't definitively prove that. It was, nevertheless, a giant pain in the ass.

I first tried a multiple-character narration. Big mistake. It took me a few days just to assign voices to every character, even those who had one single line in the entire story. Spoken even identified "ATM machine" as a character, since my MC voiced the machine's readout message. I won't rehash my entire review here, but you can read it through the link above. Suffice it to say that in the end, I reverted back to a single narrator, and I never actually listened to the entire audiobook playback.

Well, to its credit, Spoken continues to evolve. I receive email updates about everything it's up to, and finally I decided to give it another whirl. I didn't want to bother with a different story, and I had a thought that if I only converted a few chapters of Running From Herself to audio, I could use it as a free preview on my author website.

So, this time I uploaded the first four chapters only. Thankfully, Spoken no longer requires a user to add those special tags ~ huge improvement. And this time since I only had a small chunk of text to work with, I went with the multi-voice narration, with far fewer character voices to assign. The voice selections haven't changed, but I find them to be sufficient.

I finished assigning voices, then hit "publish".

Well...

When I went back the next day to have a listen, right off the bat the narrator mispronounced "peering". It was more like "pie ring". I assume I'll be able to fix that by using a phonetic spelling of the word. This happens with AI a lot, so that's not a black mark against Spoken. A bigger problem was all the pauses. The narrator would begin a new sentence and then...

...finally vocalize the rest of it. It was almost as if Mel Tillis was narrating my story. (Country fans will get the reference.) I let this go on for a bit to see if it was just a fluke. It wasn't. So I clicked "edit". Spoken had broken up sentences into separate passages for no earthly reason that I can perceive. It just did.

Now I'm in the midst of a complete edit, which requires a ton of backspacing to combine each sentence fragment into a whole. So, no, I haven't even gotten to the playback experience yet.

While mindlessly doing that, I got to thinking about the company's ultimate goal. What is it?

Google tells me:

Spoken Studio is a key part of the monetization strategy.

It utilizes AI to simplify and reduce the cost of creating audiobooks.

Authors can generate high-quality narration using AI clones of voice actors, with voice actors getting paid for each use of their voice.

This significantly reduces the financial barrier for indie authors who might otherwise struggle to afford traditional audiobook production.

Authors can create digitally-narrated audiobooks which can be downloaded and published on the platform.

The pricing for creating AI-narrated audiobooks is notably affordable: $10 per 5,000-word block, including 10,000 narration credits for revisions and changes. 

Spoken operates on a subscription model similar to platforms like Spotify.

A portion of listener subscriptions goes directly to the authors whose work they listen to.

This model incentivizes authors to create content that captivates and retains listeners, as more listening time translates to a larger share of the revenue pool.

One source mentions that 50% of a listener's $10 subscription goes directly to the writers.

This all sounds great ~ in theory. But Audible offers the same service with comparable quality, and it's part of the Amazon empire, and thus an author's audiobooks, if they choose to use the service, are listed on Amazon right next to all the other book's formats. (I went hog wild and converted all my books except Running From Herself to audio after KDP's prompting, but to be honest, I've never listened to them.)

Are readers really going to pay a separate fee to Spoken to access audiobooks (and its narrow selection)? Not to mention Spotify, for which audiobooks are part of a paid member's subscription? 

Then there's the whole frustration factor for authors. Audible did everything for me, and didn't require any manual intervention. It doesn't, though, offer multi-voice narration, but the jury (which is me) is still out on that. I'm still pondering whether multi-voice is more of a distraction than a bonus. 

Once I get done fixing everything that needs to be fixed and can go back and listen to the finished product, I'll return with an update. Maybe it'll be fantastic.

It's going to take a while to find out.