My Results & Data

[melancholyflower]

Hi all! Just ran my first ad campaign for a new release and it was somewhat successful! I spent a total of $235, got a substantial Release Radar push, and currently have 10,814 streams. I'm attaching a couple screenshots from Spotify Insights, and I'm also providing a link to the spreadsheet I used to track my data in-depth as the campaign progressed.



Main positives:
- Got a substantial Release Radar push
- Very good save rate prior to Release Radar (56%)

Main negatives:
- Release Radar push only lasted one week
- Didn't make it onto Discover Weekly (yet)
- Save rate after Release Radar is significantly less (11.6%)

I'm wondering if the decrease in save rate is an indication that it performed not-so-great during the big Release Radar push, so the algorithm didn't renew the push for a second week and didn't add it to Discover Weekly. Overall this seems like a good start and I'm excited to iterate on it!

Link to spreadsheet containing lots of data: https://tinyurl.com/y6hjcds5

Hope this is useful to some of you!

[georgik]

Thank you for sharing! What countries did you target?

[melancholyflower]

I targeted all Spotify countries except for India, Algeria, Morocco, Tunisia, Nicaragua, Egypt, the Philippines, and El Salvador. This song was my first time running a campaign for a new release, but I practiced by running a campaign for an old release first, and these were countries that did not so well for me. Specifically, they had poor ratios of conversions to Spotify listeners. For example, I was getting a lot of clicks (and conversions) from Algeria, but no Spotify streams. Not sure why (shouldn't the landing page get rid of bots?) but removing those countries has improved my Spotify results.

I also focused on targeting Brazil, Mexico, since those two have had the best conversion to Spotify listeners ratio in the past. Specifically, I ran a few ad sets that targeted all the countries, but I also did one ad set specifically targeting Brazil and Mexico.

Hope this helps!

[Regret Regret]

That’s awesome. Can you tell us how many people started following you on Spotify? I’m curious how many you would get in a day before vs during your release radar week.

[melancholyflower]

Thanks! Before release I had 101 followers (now 184). I was getting about 70 streams a day from all my previous releases combined.

[joh]

Nice! I've had the same experience with Algeria, Morocco etc. I needed around 20k streams to get on Discover Weekly, but I guess it depends on save rates, how long they listened to the song etc.

[melancholyflower]

Okay good to know! I remember someone else also mentioning that their songs tend to get on Discover Weekly after about 20k streams, so that's interesting.

[SHK]

Hey Melancholy Flower! Congrats on your success! It's awesome and inspiring.

I have started to run some tests with ads, but with no success so far. I have a question about your data: On your spreadsheet you show that on the first day you spent $25 and got 51 streams, and on the second day the $25 spend got you a whopping 405 streams. How did this huge jump in streams occur? Just due to the fact that your cost per conversion decreased dramatically from day 1 to day 2?

I hope you don't mind the question, but I really want to learn more before I end up spending a ton of money! Thanks a lot for your help!

[melancholyflower]

Hi SHK! Glad you found it helpful

On a high level, I'm pretty sure the 51 --> 405 jump is because the "days" for my campaign's daily budget are not aligned with the "days" in Spotify for Artists. To elaborate:

1. My ads didn't start running immediately when the song came out because it took about 18 hours for the ads to be approved. So it was really only running for about 6 hours on that first day (and it only spent about a quarter of the daily budget).

2. The way Spotify reports daily streams seems to be fairly complicated. That is, when someone streams your song, which "day" is that stream counted in? I think there may be some lag but I think it also depends on the listener's timezone. On a possibly related note, you'll notice that the spreadsheet lists 51 streams and 0 listeners after day 1: that's not a typo, it was showing a dash " - " for the number of listeners after day 1. So the Spotify stream/listener reporting is not always perfect.

My conversion cost was mostly constant, by the way.

Out of curiosity, when you say you haven't had success so far, what do you mean? Like high conversion cost? Or not getting streams on Spotify for artists? Definitely agree regarding learning more before spending lots of money; when I was testing, I just used a daily budget of $5-$10 and that was fine. I believe Andrew said somewhere that $5 is the minimum daily budget he recommends.

Hope this helps!

[OnyxX]

FYI, the data for each day from Spotify comes from the previous day up until some time during the afternoon EST (maybe 5pm EST?), meanwhile, Faceback ads are based on PST. So, your conversions from midnight PST to around around 1 or 2pm PST will show up a day later, and any conversions after until the next day will show up the day after. Yeah, I spent time trying to figure that out when I was tracking conversions from India and a few other countries to see if they were actually resulting in streams (only about 25% for India, so I don't target them anymore). Also, people in poor countries are more likely to have free accounts and can only shuffle on mobile. If you check your devices in the breakdown, most conversions come from mobile.

[georgik]

From Spotify: "Your audience, song, and playlist stats update once a day at approximately 3 PM EST. Days in Spotify for Artists are based on the UTC timezone (e.g. if you release a track at 7 PM EST, there won't be any data for that release day in Spotify for Artists as it's 00:00 UTC)."

https://artists.spotify.com/faq/stat...fy-for-artists

So if I understand correctly, your stats update around 3PM EST, however the days you are seeing in you Spotify For Artists page are strictly based on 24 hour UTC cycles

[OnyxX]

Yes, correct. The daily FB conversions won't ever match the "daily" Spotify data because Spotify cuts off in the middle of the day. A conversion in the morning will be in the next day's stats and a conversion in the evening will show up two days later.

[georgik]

you can create a new ad account based on UTC

[OnyxX]

Yes, that would be an approach to get them to match. I personally don't care that much. I just account for the lag when tracking results. It shouldn't matter if your ad spend is consistent, anyway, but it's good to be aware of it.