Hi there! I like to listen to the “Slightly something else” podcast. Although it is gaming-related, they have these ad-segments that have some kind of targetting, as it is always in the language from where I’ve downloaded the episone.
Yesterday I was really surprised that the ad seemed very well targeted, because it has been referencing a very specific topic concerning my flatmates and me that doesn’t have anything to do with gaming.
My first theory was that my IP was correlated with search terms my partner googled in my home IP (I search with searxng). But then I realized that I’ve subscribed to another podcast related to the advertised topic. (also: the episode with the advertisements was downloaded via mobile network)
So… Apparently, spotify (where slightly something else is hosted via anchor.fm) analyzes the podcasts I subscribe to via rss feed. Is there any way how I can avoid being profiled? Is this the reason why I’d start using a commercial VPN?
Edit: I’d like to reply to your comments, but my lemmy instance won’t display them correctly. So I’ll try to answer here: I’m already using Antennapod with rss feeds for download. That’s why I was so suprised and pissed that the ad was so obviously targeted.
Ok now that I’ve read your Edit:
Podcasts no longer have RSS feeds that simply link back to an mp3 file that you directly download. Any podcast that is running canned ads (not ad reads that the host reads) is using what’s called Dynamic Ad Injection. This requires the podcaster to supply their distribution vendor with a key file that marks break points in the podcast that the system will use to slice the raw podcast file into parts, and then inject X number of ads into the break, and then reform the mp3 file for delivery. I imagine this can be done dynamically or as part of an automation that cranks out thousands of variations of the podcast file.
The reason this information is important is because it means that the PODCAST FEED itself is betraying you and your data. The feed is controlled by an ad delivery network, which works in partnership with the podcast, to inject dynamic ads based on demographic information the ad delivery network has, and they use metadata about your connection (such as IP address) to match you against a shadow portfolio of your data, or at least a pool of data that hits your demographic.
Now, one other thing here that is betraying you, is your flatmates. It doesn’t matter how much you protect yourself, because you are likely connected to your wifi, along with the other flatmates, and they’re likely not taking the same measures as you are to ensure anonymity online. So you could have received that ad not because of data collected about you, but because of data collected about everyone who uses your ISP-provided address to connect to the internet, even if they’re not listening to this specific podcast.
A VPN can mitigate most of that, as well as using a FOSS Podcast reader like you are. But if you have to listen to the podcast through Spotify, it’s still moot because the IP addresses you get via the VPN will become attached to your data profile from Spotify, which will then find its way back into the existing data on you which includes your apartment’s IP address.
The reality is Podcasts are big businesses and all the feeds for all the popular podcasts are owned by ad networks that collect data on listeners and then pair that data up with other data they collect from data brokers.