Podcast transcripts coming to Spotify. The company announced the move on Tuesday, pitching it as part of a broader accessibility push. They noted that the feature will begin rolling out on select Spotify Original and Exclusive podcasts across both iOS and Android experiences in the coming weeks.
The transcripts, which are auto-generated, will also have a certain element of interactive searchability to it, in the sense that you can tap a specific sentence to play the episode from that point. (Descript or Trint users should be familiar with this flow.)
Spotify doesn’t intend to limit the feature to just its own podcasts, obviously. “This is just the first step for the podcast transcript experience: Our overall ambition is to enable transcripts across all podcasts on Spotify,” the corporate blog post states. It also notes: “As always, we want to know your thoughts, so head over to our Community to share your feedback.”
The other two accessibility changes that were part of this announcement are: (1) adjustments and additions to the visual icons in the app meant to increase readability ease for those visually impaired, and (2) greater text-scaling variability.
Back to the transcript development. This feature, of course, shouldn’t come as a surprise; the possibility of Spotify intending to deploy auto-generated podcast transcripts is baked into the platform’s very terms and conditions. Specifically, here:
5. Spotify for Podcasters Authorized User Content
Authorized User Content. Registered Spotify for Podcasters users (“Podcasters”), including you and any of your Subscribing Entities and Delegates, may, at no cost to Spotify, submit, upload, and/or contribute to the Service authorized audio content (which may include, for example, any associated metadata, images, artwork, texts, photographs, videos, podcast transcripts, marketing materials or other associated materials, as specified by you in the Spotify for Podcasters submission form).
Emphasis mine. It should also come to no surprise that some corners of the podcast community will likely have mixed-to-negative feelings about this. On the one hand, making it easier to mass-generate and normalize podcast transcripts is, truly, a huge step forward for podcast accessibility. And not to mention skimmability, which will be helpful for folks like me who write about and review podcasts. I tell ya, I’ve been doing these reviews for Fresh Air, and I’ve burned so much time trying to find specific sentences when figuring out clips.
On the other hand, there are some trailing concerns. One such concern is the fact that auto-generated transcripts aren’t perfect and may lead to imprecise copy that could cut against the intent of the creator, producer, or writer. (Plus, if one would like to apocalyptically project outward, one could imagine a scenario where this could feed certain forms of misinformation, i.e. “Did you see what Politician X said on Podcast Y?” when in fact the actual word said may have been mistranslated.)
Another concern is linked to more general anxieties about platform power, and the fact that these transcripts — plausibly arguable as derivative creations — will probably drive disproportionate value for the platform compared to the creator. One has to imagine the auto-generated transcripts opens up the possibility for the platform to use those transcripts to improve keyword-searchability matching, and so on. (There’s a possible interesting parallel here between podcast transcripts and text-to-speech services like Audm or Curio; they move in opposite directions, but they play around the same question of how value will ultimately be split between the service/platform and the source of the actual content.)
Then again, this power/value differential at play is table stakes for all sorts of platforms that distribute content from users and content partners, and I reckon there’s a generation of podcast creators looking at this situation and figuring that trade-off is fine — or, perhaps more accurately, that they don’t really have any other choice.
Two more Spotify developments…
(1) The company announced this morning a deal with Storytel, a Swedish audiobook streaming service (think of it as a minor European Audible competitor), that’ll let Storytel subscribers access its offerings through the Spotify platform. That distribution will be built off Spotify’s Open Access Platform, the technology built in tandem with the company’s subscription tool push that gives publishers with a paid subscription base managed on other platforms — say, Patreon or Supporting Cast — the option to deliver paid content to their existing paid audience using Spotify, something that wasn’t previously possible.
Basically, it’s yet another tangible step in the direction towards Spotify’s aspiration to become the all-consuming audio platform, which we’ve been talking about here in the newsletter for years. As Courtney Holt, Spotify’s Global Head of Studios, reiterated in the press release: ““It is Spotify’s goal to be the singular platform for all audio: music, podcasts, live conversations, and now via this partnership, audiobooks.”
This, I should note, isn’t Spotify first visible step around audiobooks exploration. A few months back, the platform published a few shows featuring various famous people doing the narration for select public domain books. I believe it featured David Dobrik doing Frankenstein, which, you know, wild, given the recent controversies.
I should note, at this point, that Spotify isn’t the only multi-faceted organization out there trying to find wiggle room around Audible within the audiobooks world. Slate, as well, recently announced its own foray into audiobooks, using the premium podcast feed platform created by its sister company, Supporting Cast. Pushkin Industries, which also features Slate lineage, similarly has its own small/budding independent audiobooks operation, one that scored itself a Wall Street Journal profile a few weeks back. (My understanding is that the Pushkin push here is led by Brendan Francis Newnam, the former Dinner Party Download co-host, who now holds the titles of VP of Special Projects at the company.)
One quick thing about Slate: the company has partnerships with the major book publishing companies for its emerging audiobooks operation, including Penguin-Random House, Simon & Schuster, HarperCollins, Hachette, and Blackstone. They’re also working with authors directly — as with the case of Daniel Lavery, which I mentioned briefly in a previous column — and when I reached out to ask for more detail about these publisher arrangements, Slate associate editor Jeff Friedrich tells me: “Publishers have for the most part been extremely easy to work with; clearly they’re hoping to help prop up alternative sales channels to the ones they have today. We’re starting out with just around 100 titles or so, but we expect to grow to many times that size by the end of the year.”
This is also probably a good time to link to my old column about on-demand audio and audiobooks. From 2019: “Audiobooks are not exempt.”
(2) From TechCrunch: “Spotify launches a virtual concert series with The Black Keys and more.”
“And more” being Bleachers, girl in red, Rag’n’Bone Man, and Leon fuckin’ Bridges, c’mon. The splash page can be found here.
Two things that’s interesting to me about this. Firstly, it’s a straight-up virtual concert situation, where users pay $15 per ticket to see the acts perform in a physical location that’s live-streamed to your computer. One wonders how Live Nation Entertainment is feeling about this.
Secondly, this initiative seems separate and apart from Spotify’s efforts around Green Room (née Locker Room), its Clubhouse competitor, though we’ll see.
Virtual music performances are cool and all, but I can’t wait to get back into a crowded dimly-packed room with a beer in hand having my view of the stage obscured by an awkwardly swaying and extremely tall bearded music bro in flannel.
Speaking of events…
Zoom party. Zoom announced that it will be launching an expanded version of its live events platform, Zoom Events, this summer. The platform, which you can find a landing page for here, is generally being marketed as a virtual conference stand-in, but given the fact more than a few podcasts tried to use Zoom to stage live versions of their shows, this might be a germane development to note for those purposes.
Or, you know, there’s always FRQNCY, which I wrote about earlier this month.
Hmm… This morning, Vox Media announced a curious addition to Cafe Studios’ podcast portfolio: a new weekly podcast with historians Heather Cox Richardson and Joanne Freeman called Now & Then, meant to do that thing where historians look at the news and try to draw parallels with the past. (Shouts to This Day in Esoteric Political History, which doesn’t do that exactly, but has largely served the same purposes for me.)
Anyway, two reasons I find this interesting. To begin with, it’s the first addition to Cafe Studios since it was acquired by Vox Media last month. And also: Heather Cox Richardson was — at least at some point in recent history — the “most successful individual author of a paid publication” launched on Substack, according to the New York Times’ Ben Smith. Let’s see if that deep newsletter following translates over to the podcast.
Edison Research on Clubhouse Usage. There’s a considerable amount of endlessly interesting stuff packed into Edison’s survey study of current Clubhouse users, which the firm published earlier this week. Endlessly interesting… but not altogether surprising.
The three big takeaways that stand out to me:
- Only 15% of American social media users over the age of 18 reported having ever used Clubhouse. Compared against a list of other social media platforms, Clubhouse sits at the bottom, along with Gab (15%), and just under Parler (16%). Wild. For broader context, TikTok is at 36%, while YouTube sits at the very top, at 89%.
- There seemed to be some slowing down of Clubhouse usage, with the “Average Weekly Time Spent on Clubhouse” metric dropping to 4:12 in April from 4:48 in March. This broadly tracks with the story that new app downloads have dropped precipitously between February and April.
- Clubhouse users are disproportionately male, whiter, younger, more liberal, and have higher incomes than the average American.
Tom Webster, Edison’s SVP of Research, is spot on with his assessment on what informed the demographics here: “Clubhouse has been an iPhone-only service up until now… and as a result, the audience for the service is not as diverse as we see in the leading social platforms.” (Comparisons to early podcasting would be appropriate here.) But also germane, I think, is the fact that Clubhouse exercises an invite-only approach to user growth and continues to do so. Like begets like, and I imagine the nature of the originating networks played a role in this too.
Keep in mind, though: the interviews for this survey study was conducted between early February and late April. Clubhouse announced that it would be expanding to Android just earlier this month, and the latest word is that the Android app will be fully available worldwide by the end of this week. One imagines that the shape of Clubhouse users would shift considerably as a result of this.
Ambies Viewership. Executive director Michele Cobb claimed that viewership of the event, which took place on Sunday, topped 600,000 unique viewers with about 25,000 concurrent, when I reached out over email asking for numbers. Consumption was spread out across YouTube and Twitch. No choice but to take her at her word on this one, at least at this point.