Understanding Open Podcast Prefix Project (OP3) with John Spurlock | Hacker News Reader

Understanding Open Podcast Prefix Project (OP3) with John Spurlock

blog.castopod.org

24 pointswalterbell21 comments

Questioning the premise... is there a need for this? Chasing analytics is famously poisonous in other fields of content-creation.

I guess advertisers want to know exactly who's listening. The point of centralizing it like this would then seem to be to make it possible to validate your traffic through a theoretically impartial third-party.

(I build Pinecast.com)

One of the biggest things is knowing who's an actual human downloading your podcast. If it's all bots, that tells you very little about growth or what's working and what's not. Even if you're not running ads, having an inaccurate view of how many people are listening makes it hard to know what to do; it's very much not "if you build it they will listen".

Having a third party is nice even if you don't have an advertiser in the loop. Being able to compare (between 2 episodes, 2 podcasts, 2 countries, 2 players…) is a real plus. And at least OP3 is open data.

Podcasting is open in many ways, but RSS-based podcast apps have no idea what shows are "hot in cleveland" or have related listener bases, etc outside of their own app.

When shows start using OP3, they open up these (listener privacy-preserving) stats to _any_ app, and set themselves up for future discovery / inbound opportunities on this basis.

> Similarly, data should be open to stimulate innovation... At OP3, we apply this philosophy to both our podcasts and our data. Currently, OP3 focuses mainly on providing data to podcasters, then to podcast applications, and finally to the industry as a whole. The idea is not to allow the download of a complete database like PodcastIndex.org does, but to provide specific statistics and analyses. Users can get data on specific podcasts if they know their URLs. We may consider making the data more widely available on a paid basis in the future to support the project.

That is not open data as I understand it. How can data be called open when it's locked on someone's server and I have to pay them to access it? What's the difference between this open data standard and collecting data about users and selling it to marketing companies?

Yea that part got lost in translation.

The OP3 API is completely free to use: https://op3.dev/api/docs

> OP3 API is completely free to use

Is there a license for the data provided by the API?

> The idea is not to allow the download of a complete database like PodcastIndex.org

Can an API data client publish a private or public mirror of the entire database?

It's freely available to anyone - the data (after securely hashing IPs) is provided as a service for everyone in the open podcast system.

There is a bit about the data api in the privacy policy [1]. Feel free to engage over on the github repo if you are blocked on anything or to chat about it [2].

[1] https://op3.dev/privacy [2] https://github.com/skymethod/op3/discussions

  The Open Podcast Prefix Project (OP3) is a free and open-source podcast prefix analytics service committed to open data and listener privacy.

It would be helpful to link "open data" to an op3.dev/opendata page that either defines data usage rights or states something like "podcast analytics data is intended for use by individual podcast publishers. For usage of aggregated data, please contact []". Maybe an FAQ with examples of acceptable commercial usage.

The "listener privacy" text could be linked to op3.dev/privacy.

OP3 is a great tool and a great idea, but unfortunately the reason most existing prefixes have so much market share is because they are tied in to advertising. Podtrac is the biggest player despite being a crummy analytics tool because podcasters use their Podtrac ranking to signal their reach to advertisers. Everyone uses Podtrac so the numbers are apples to apples. I sincerely hope OP3 gets enough critical mass that it can fill the same need.

I hope it gets enough critical mass too!

Is Wordpress suitable for hosting a paid podcast with per-subscriber feeds, without using a commercial delivery network like Castos that costs $500+ monthly for a few thousand subscribers?

It looks like open-source Castopod supports OP3 and could run alongside existing Wordpress sites, https://castopod.org/

What's wrong with RSS?

Nothing

You stick this prefix in front the audio enclosure file urls inside your rss feed

Why should I do that? It only introduces a single point of failure and destroys my listeners' privacy.

Is a single point of failure with the URL prefixes not a terrible idea?

From the homepage [1]

> We've been running smoothly since Sept 2022, now measuring over 15 million podcast downloads every month across more than 1600 shows

Knock wood, but OP3 podcast redirecting has never had an single outage since it launched in Sept 2022.

Turns out Cloudflare Workers do a great job at simple http redirecting via their hundreds of edge locations, even during extended outages of the other parts of their stack.

[1] https://op3.dev

Indeed, you have to be really careful and monitor your prefix provider very closely. But it's the best (only?) way to get podcast analytics. Other technologies such as RAD would have been much more resilient, but no one wanted them. :-/

At some point I wanted to embed a monitoring feature within Castopod that would disable op3 in case it fails to respond… but it would add some much plumbing that the monitoring mecanism itself would probably fail long before Op3…

If you look at any major podcast, they likely already have 2 or 3 of these. They're rarely a problem. OP3 in particular is Cloudflare workers IIRC so it's about as reliable as Cloudflare.

Apple will stop following 302 redirections after… 25 of them!

Reply on news.ycombinator.com