Another week zoomed by. Yeesh. I've been a lot more focused on business development (with the major labels) this week, as well as a few marketing initiatives we're going to undergo. Still, a good bit of code has also been smashed :)
Earlier this week I spent some time overhauling the way we present Spotify playlist data in the list of curators. You can now easily view playlist-specific statistics or listen to the playlist (whereas formerly this was a bit harder to navigate).
Another cool new feature you'll see rolling out over the next week or so is geographic data associated with each playlist. This one's a bit harder to pin down, but we've got a cool partnership Is It A Good Playlist that allows us to loosely piece together where a playlist's listeners are based -- and then back that into an overall estimate for the playlist itself. As with all things, it's difficult to know with 100% certainty that the analysis being done is accurate (Spotify won't give us much), but... I think it'll be helpful regardless, and preliminary data looks really enlightening.
I changed some behind-the-scenes stuff for curators, giving the Earnings and Playlist tabs sub-menus that are easier to navigate. As time marches on, more and more data gets collected, which meant these pages had a *lot* of scrolling going on. Should be much better now.
We also did a bunch of work on our back-end ticketing system, but given that pretty much none of you can see that, I'll spare you the detail :-D
We've started work on two new behind-the-scenes things, both of which I'll pick up on next week.
The first is analyzing Links data and trying to extract meaningful clues into a playlist's average listeners. This is proving pretty difficult, as the number of clicks into a playlist *do not* seem to have a strong correlation to the number of listeners a playlist actually manages to generate for each individual song. More digging needed!
The second subject relates to the new TikTok chart we launched last week. We're going to experiment re-investing in those songs - that is, pumping our own money into submitting them to influencers. The logic is that this will 1) deliver a bit of "virality" for artists who are leveraging TikTok submissions and 2) keep Influencers engaged with SubmitHub (if you're only receiving 1-2 submissions a day you tend to stop checking in - and we want to avoid that).
Ugh. This again. Basically Facebook/Instagram and TikTok are making it harder and harder to get public information from their profiles using a script. Something to do with you needing to prove you're a real human, yada yada. Long story short, I switched everything to a new scraping API service called ScrapingBee. Seems to be having a much higher success rate than ScraperAPI was. Fingers crossed!