[ad_1]
Spotify acquired podcast discovery platform Podz final summer season for roughly $49.4 million, per a Securities and Trade Fee submitting, to assist speed up the streamer’s in depth investments in podcasts. Now, Spotify is testing a characteristic that leverages the startup’s expertise to assist customers discover new podcasts they might like, the corporate confirmed.
Podz had initially tried to resolve the issue of podcast discovery with one thing it known as the “first audio newsfeed.” That’s, it offered customers with 60-second audio clips from varied exhibits that you’d scroll by means of in a vertical feed, just like the format popularized by social apps like TikTok. What made the corporate’s expertise fascinating is that it didn’t depend on podcast creators to provide their very own clips for its feed. As a substitute, it used a machine studying mannequin that had been educated on some 100,000 hours of audio to assist routinely choose clips to showcase.
On the time, Spotify touted the acquisition as a method of constructing out and scaling a greater and extra personalised podcast discovery expertise on its app. With this check, we now have a primary have a look at what such a characteristic may appear to be.
Product designer and tech early adopter Chris Messina tweeted concerning the check and posted a video of the characteristic, which was first uncovered by the Twitter account @SleepwellCap. Right here, you’ll be able to see the brand new expertise in motion.
HUGE!! @Spotify is including a devoted TikTok-style vertical scrolling feed for @spotifypodcasts to its toolbar!
h/t @SleepwellCap #NewSpotify pic.twitter.com/0CWolZGJQy
— Chris Messina 🪬 (@chrismessina) March 26, 2022
A devoted “Podcasts” button takes you into the vertical feed, the place you’ll see the present’s cowl artwork whereas the audio clip performs. The clip can also be transcribed as you pay attention with the phrases within the clips highlighted as they’re spoken. There’s a play button to proceed to hearken to the present and plus “+” button so as to add the episode to the app’s “Your Episodes” checklist.
On condition that that is solely a check, it’s value noting the characteristic may change earlier than a public launch. The characteristic additionally might merely be used to assist Spotify achieve an understanding of how customers would have interaction with such an choice, which may then be used to tell future product developments. In different phrases, there’s no assure you’ll quickly see the identical vertical feed of audio clips in your individual Spotify app.
Spotify confirmed it’s testing Podz’s expertise by means of this expertise however wouldn’t decide to a launch date or plans. (On condition that Podz is machine-learning powered, it might be that Spotify needs to collect extra information to enhance the service earlier than a launch.)
“At Spotify, we routinely conduct plenty of exams in an effort to enhance our person expertise,” a spokesperson stated. “A few of these exams find yourself paving the best way for our broader person expertise and others serve solely as an essential studying. We don’t have any additional information to share right now.”
Podz had been one among a number of startups constructing out new methods to assist individuals discover extra podcasts they might like — a discovery problem that’s been tough to beat since many podcasts are over half-hour in size, making it arduous to get a very good sense of a present’s content material or the hosts’ personalities in a fast and straightforward method.
Podz’s answer was to lean on machine studying expertise, however different startups have pursued completely different concepts. As an example, Moonbeam, a brand new startup from Kayak co-founder Paul English, depends on editorial curation from people blended with machine studying methods for its TikTok-inspired app. The vertical feed format has additionally been utilized in podcast-adjacent apps, like Racket, which provides 99-second audio clips which can be extra like audio tales than exhibits, and with Fb’s Soundbites.
Replace! Right here’s tips on how to attempt the characteristic now:
[ad_2]
Source link