Stop guessing which song to promote
The practical song-testing system for independent Hip-Hop & R&B artists
A few years ago, before most people knew who Alessia Cara was, I came across a song called “Here” on SoundCloud.
I listened to it and really liked it. Being an introvert myself, it made sense. Later, I went back looking for the song and it was gone.
Then, months later, Alessia Cara officially emerged with “Here,” and the song became a major part of her breakout.
Now, I have no way of knowing exactly what was happening behind the scenes with that SoundCloud upload. I wasn’t sitting in a label meeting, and I’m not going to pretend I know whether it was part of some formal testing strategy.
But that experience stuck with me. Because it made me think about something independent artists should be thinking about before they spend serious money marketing a song:
What if you could test the record before you go all-in on it?
Instead of guessing, put the music in front of people who don’t know you and pay attention to what they do.
Do they stop? Do they listen? Do they engage? Do they go looking for the song? Do they come back?
That’s the foundation of this entire strategy.
The big idea
Major artists and record labels have something independent artists often don’t have: A&R and marketing teams whose job is to figure out what has potential before putting serious resources behind it.
They can test songs, audiences, creative concepts, messaging and different marketing strategies.
You may not have a team sitting in a conference room analyzing your next single. But you can borrow the mindset.
And today, you have something artists didn’t have in the same way 20 years ago: direct access to audience behavior.
You can put your music in front of strangers and watch what they do. That means you don’t have to make every marketing decision based on a hunch.
Don’t ask “Which song is the best?”
Ask:
“Which song is creating the strongest evidence of demand?”
Those aren’t necessarily the same thing.
Your favorite song may be the most personal. Your producer’s favorite may be the most technically impressive. Your friends may love another song because they know the story behind it.
But your audience doesn’t have that context. They’re strangers. And strangers are honest.
They don’t care how long you spent recording the song. They don’t know how many times you rewrote the verse.
They either stop… or they scroll.
They either listen… or they leave.
They either want more… or they don’t.
That’s valuable information.
First: a song cannot fail if nobody heard it
This is one of the most important concepts in this entire guide. Let’s say you upload a song and it gets:
You don’t have enough information. Two people hearing a song isn’t a market test.
This is where many independent artists accidentally kill good music. They release a song. It doesn’t immediately move. They assume: “People don’t like this.”
But what actually happened? Nobody heard it. Those are two completely different problems.
No audience = no data.
Before you can decide whether a song is working, you need to get it in front of enough relevant people to begin seeing a pattern. That’s your first job.
The testing loop
Think about your marketing like this:
Do this
Creates information.
Not this
Creates expensive guesses.
What are you actually testing?
Here’s where this gets really interesting. You’re not just testing the song. You’re testing:
Imagine you have an R&B song about a breakup. You could present that exact same song as:
Same song. Different presentations. And different people may respond to each version.
That’s why one video can get 300 views while another video using the exact same song gets 30,000. The song didn’t magically change. The way you introduced it changed.
Your first test
Before spending serious money, choose your top 3 songs. For each song, identify:
1. The strongest lyric
What’s the line that makes someone think: “Wait… what did they just say?”
2. The strongest musical moment
The hook. The beat drop. The melody. The vocal run. The section that makes you want to rewind.
3. The most emotional or interesting moment
What’s the part that makes somebody feel something?
You now have three potential moments to test for each song. That’s 9 opportunities to learn something before you spend serious money:
| Strongest lyric | Musical moment | Emotional moment | |
|---|---|---|---|
| Song 1 | Test 1 | Test 2 | Test 3 |
| Song 2 | Test 4 | Test 5 | Test 6 |
| Song 3 | Test 7 | Test 8 | Test 9 |
Don’t make a “promo video.” Make a test.
You have a song. Now imagine putting it in front of 100 people who have never heard of you.
Video 1
“Here’s my new single. It’s out now.”
They scroll.
Video 2 (same song)
“I almost deleted this verse because I thought this bar was too crazy.” Then you play the bar.
People stop. Someone comments “Nah, that was hard.” Someone shares it. Another person goes to your profile.
What changed? The song didn’t. The presentation did. That’s what you’re testing.
Think like an A&R for your own music
Let’s say you’re a rapper with three songs. Instead of arguing about which one is the single, test them. Give each song multiple opportunities. Here’s what the results might look like:
| Song | Who picked it | What happened | Possible next move |
|---|---|---|---|
| Song A | Your personal favorite | Decent views, but nobody finishes the video | Needs a completely different presentation |
| Song B | Your producer’s favorite | Fewer views, but lots of comments | Needs a better hook |
| Song C | The one you weren’t sure about | Fewer initial views, but people watch longer, visit your profile, ask when it drops and share it | Deserves another test |
Now you have something you didn’t have before:
Evidence.
The point isn’t that the numbers automatically tell you what to do. The numbers give you clues. Your job is to figure out what those clues mean.
One song. Multiple tests.
Don’t post one video and decide: “This song isn’t working.”
You haven’t necessarily tested the song. You may have tested one piece of content. For one song, try multiple angles.
Rap: test the bar, the story & the personality
For Hip-Hop artists, try:
The bar test
“I almost took this line out…” Then play the bar.
The debate
“Be honest. Is this actually a hard bar or am I tripping?”
The story
“I wrote this after somebody I trusted did me dirty.” Then play the relevant lyric.
The reaction
Have someone hear the verse for the first time.
The cypher test
Perform the verse directly into your phone. No expensive video, no elaborate production. Just you + the music + the bar.
Then read the data
This is where testing becomes more than just posting. Don’t look at one number. Look at the journey.
- Stranger
- Watches
- Stays
- Engages
- Visits
- Follows
- Listens
- Returns
Every step tells you something.
What the metrics actually mean
| Metric | What it signals | What to take from it |
|---|---|---|
| Views | Exposure | Did anyone see it? Important, but not a success score by itself. |
| Retention | Attention | Did they stay? If people immediately swipe away, your opening may be the problem. |
| Engagement | Response | Likes, comments, shares and saves show whether people did something after watching. |
| Shares | “I want someone else to hear this” | Someone chose to send your content to another person. That means a lot. |
| Profile visits | Curiosity about you | Someone watched your video and then wanted to know who you are. |
| Follows | “I want more” | They want the opportunity to see you again. |
| Saves & repeat listens | Intent | They want access to the music again. That’s a stronger signal than simply being exposed to it. |
Pay attention to the quality of the response, too:
| Comment | What it tells you |
|---|---|
| 🔥🔥🔥 | Nice to see, but easy to leave |
| “Who hurt you?” | More meaningful: the song made them feel something |
| “When does this drop?” | Even more intent: they want to hear it again |
How to diagnose the problem
| If you see… | It may mean… |
|---|---|
| High views + low retention | Your hook or opening may be the problem. Don’t blame the song yet. |
| Low views + strong retention | The content may actually be working. You may have a distribution problem. |
| Strong retention + low engagement | People are watching, but not feeling strongly enough to respond. Try a more emotional, relatable or conversational angle. |
| Strong engagement + low profile visits | The content may be entertaining, but people aren’t connecting it strongly enough to you. |
| High profile visits + low follows | You’ve created curiosity, but your profile isn’t giving people enough reason to stay. |
| High follows + low music consumption | People may like you more than the music they’re being shown. That’s valuable information too. |
| Strong music response + weak social response | Don’t kill the song. You may have a content/distribution problem. |
| Everything is weak | Ask: did enough relevant people actually hear it? If not, you don’t have a conclusion yet. |
The $50 mindset
When you eventually use paid advertising, change the way you think about the money. Don’t think: “I’m spending $50 to make this song successful.”
Think:
“I’m spending $50 to buy information.”
Test two pieces of content. Same song. Similar audience. Similar budget. Now you have a comparison.
If one consistently outperforms the other, you’ve learned something. Then create more variations around the winner.
The goal isn’t to manufacture fake popularity. The goal is to discover what actually works.
Your 7-day song test
- Day 1Choose your top 3 songs.
- Day 2Find the strongest lyric, musical hook and emotional moment in each.
- Day 3Create three pieces of content for Song #1.
- Day 4Test Song #2.
- Day 5Test Song #3.
- Day 6Review the data. Look at:
- Views
- Retention
- Shares
- Comments
- Profile visits
- Follows
- Music activity
- Saves/repeat behavior where available
- Day 7Choose the strongest combination of song + content + audience. Then test it again.
Don’t scale a guess
This is the biggest takeaway.
Don’t spend $2,000 promoting something you haven’t tested when you could spend a fraction of that learning first.
Don’t assume your favorite song is automatically your audience’s favorite.
Don’t assume a song failed because the first post didn’t move.
Don’t assume a million views means you’ve created fans.
And don’t assume low views mean the music is bad. Instead:
Test the record.
Find the response.
Learn the audience.
Then scale.
Because the goal isn’t simply to get lucky. The goal is to become better at making marketing decisions based on evidence instead of guesses.
One more example
I found this video on YouTube of someone who helped a London rapper. I think it’s a great example of this type of thinking in action.
Let me know if this article was helpful and if you tried any of the methods listed.