Solved: Meta Ads Campaign Not Exiting Learning Phase

Hi there,

Thanks for reaching out! I've had a look at the issues you're facing with your Meta ads campaign for Spotify, and I'm happy to give you some initial thoughts. It's a common set of problems, especially for artists trying to promote their music, but the good news is that they are absolutly fixable.

The core of the problem isn't your audience or some phantom "bot" issue, but a fundamental breakdown in your tracking and campaign structure. We need to sort that out first before you can ever hope to see those costs come down. Let's get into it.

We'll need to look at your tracking... because right now, you don't have any

Okay, let's be brutally honest. The single biggest reason your campaigns are failing is your use of Linktree. When you send ad traffic to a Linktree, you're essentially sending it into a black hole from Meta's perspective. You've placed a wall between your ad and the final destination.

The Meta Pixel works by being placed on a website that you own and control. It watches user behaviour on that site and sends the data back to Meta. When a user takes the action you want (like clicking a "Listen on Spotify" button), the pixel fires a "conversion" event. The algorithm then learns what kind of person takes that action and goes to find more people like them. Without this feedback loop, the algorithm is just guessing. It has no idea who is a good user and who is a bad one.

When a user clicks your ad, goes to Linktree, and then clicks again to go to Spotify, the pixel has no idea what happened after they left your Linktree page. It can't see the action on Spotify's platform. So, as far as Meta is concerned, almost nobody is converting. This is why your costs are sky-high and you're not seeing a "learning phase" status—the campaign isn't learning anything because you're not giving it any data to learn from.

This is a fundamental mistake that wastes a huge amount of money for new advertisers. You're paying to send traffic, but you're not collecting the data that makes the ad platform smart. You're basically running a traffic campaign but telling Meta to optimise for conversions it can't see.

I'd say you need to fix your funnel before anything else

So, what's the solution? You need to create a very simple, single-page website to act as a bridge. This is often called a landing page or a "presell" page. You can build one in a few hours using services like Carrd, Leadpages, or even a simple WordPress install. It doesn't need to be complex.

Here's what goes on this page:

• Your Meta Pixel: This is the most important part. You install the pixel on this page.
• Some info about you/the song: A good photo, a short bio, maybe the album art.
• A clear Call to Action (CTA): A big, obvious button that says "Listen on Spotify" (and maybe other buttons for Apple Music, YouTube, etc.).

Now, when someone clicks your ad, they land on your page. The pixel registers their visit. When they click the "Listen on Spotify" button, you set up the pixel to fire a custom conversion event. You can call it "SpotifyClick" or whatever you like. In your Meta Ads Manager, you then tell your campaign to optimise for that specific event. Suddenly, you're giving the algorithm exactly the data it needs. It will now actively hunt for people who are most likely to click that button, and your cost per conversion will start to drop.

You'll need to understand the 'Learning Phase' (and why your budget is holding you back)

Now let's talk about that missing "learning phase". The learning phase is the period where Meta's system is actively exploring who the best audience is for your ad. To exit this phase and become stable, a campaign's ad set needs to get roughly 50 conversion events within a 7-day period.

Let's do the maths on your current situation. Your daily budget is $10-$25, so let's say an average of $17.50/day. That's $122.50 per week. Your cost per "conversion" (which is probably just a link click in reality) is $4.

$122.50 (weekly spend) / $4 (cost per conversion) = ~30 conversions per week.

You're falling short of the 50-conversion threshold needed to exit the learning phase. The campaign is "stuck" in a state of constant, inefficient learning because it never gets enough data to become confident in who to target. It's often labelled as "Learning Limited" for this reason. This is a very common issue with low budgets. The algorithm is starved of data.

The solution here is twofold:

1. Fix your tracking (as we discussed above). This will give the algorithm real conversion signals to work with, which should lower your cost per conversion.
2. Ensure your budget is sufficient. Once you have a more reasonable cost per conversion (let's say it drops to $2 after fixing the tracking), you'd need a budget of at least ($2 * 50 conversions) / 7 days = ~$14.30 per day to exit the learning phase. Your current budget is right on the edge, but it's completely ineffective with a broken tracking setup.

You probably should reset your cost expectations (but not by that much)

You mentioned you thought $0.50 per conversion would be okay. For Tier 1 countries (US, UK, Canada, Australia etc.), this is very optimistic. To give you a real-world benchmark, I remember one large-scale app growth campaign we managed across Meta and other platforms where we generated over 45,000 signups, keeping the cost just under £2 per signup. While a click to Spotify is a slightly different action, this shows what's possible with a properly structured campaign. The $4 you're seeing now is definitely on the high side, especially because it's for an untracked, low-quality event.

A "Spotify conversion" isn't like a sale or even a solid lead. It's a fairly low-intent action. Someone clicking a link to listen to a song is not as valuable as someone signing up for a trial or buying a product. Therefore, you should aim for a lower cost, but it needs to be realistic.

Once you fix your tracking and get a proper landing page, I'd say a good target to aim for would be in the $1.00 - $2.50 range for a click-to-Spotify event in T1/T2 countries. This is achievable with the right creative and targeting. Getting below $1 would be an excellent result. The key is to stop focusing on the cost from a broken system and start focusing on getting a reliable cost from a working system. Then you can work on optimising it down.

Finally, let's talk about those 'Indian bots'...

I can put your mind at ease here. The old global traffic campaign you ran has not "messed up the meta AI" for your new campaigns. This is a common fear, but it's not how the system works. Each campaign, especially when using a new pixel and targeting a completely different objective (conversions vs. traffic), is assessed on its own merits. The algorithm isn't permanently 'tainted' by past data from unrelated campaigns.

When you told Meta to get you the cheapest traffic possible, it did its job perfectly by showing your ads in countries like India where traffic costs are fractions of a penny. When you create a new campaign and tell it to find people in Tier 1 countries who are likely to convert, it starts a completely seperate and new learning process for that specific goal. The old data is irrelevant.

So, you can stop worrying about that. Your current problem is 100% due to the tracking and budget issues we've discussed, not some ghost in the machine from a past mistake.

This is the main advice I have for you:

To wrap this all up, here is a clear, actionable plan. This is what I would do if I were taking over your account today. Stop spending any more money until you've done steps 1 and 2. There is no point optimising audiences or ads when your foundations are broken.

Following these steps will transform your results. You'll go from blindly throwing money at the wall to running a data-driven campaign that can actually be optimised and scaled. It might seem like a bit of work upfront, but it's the only way to acheive sustainable, low-cost results on the platform.

If you get stuck on any of this, or if you'd prefer to have an expert handle the setup and management to ensure it's done right from the start, that's something we do for clients every day. We could run a proper audit of your setup and build out a strategy that works. You can schedule a free, no-obligation consultation with us to talk through your project in more detail.

Hope that helps!

Regards,

Team @ Lukas Holschuh