My real songwriting process, from a musician who learned it the hard way

Most songwriting advice is backwards.

It starts with craft. Chord progressions, song structure, rhyme schemes, verse chorus verse. All of that is real and all of it matters eventually. But starting there is like teaching someone to cook by explaining the chemistry of heat transfer before they've ever tasted a meal. Technically accurate. Practically useless as a starting point.

The question that actually matters at the beginning of a song is not how. It's what. What are you trying to make someone feel? What is the thing that needs to come out of you that has nowhere else to go? What would you say to a specific person in a specific moment if you were not afraid of what they would think of it?

Start there. The craft is what you use to get there reliably. But the destination has to come first.

I have been writing songs for most of my adult life. I have written them through a spinal condition that kept me in serious pain for months. I have written them for people I was watching head somewhere bad. I have written them in the middle of the night because the thing needed to come out and the guitar was the only container that would hold it. Here is what I have actually learned about how that process works.

FIND THE FEELING BEFORE YOU FIND THE CHORD

Every song I have written that I am genuinely proud of started with a feeling, not a riff.

That might sound obvious but most people who sit down to write music start by playing. They noodle around on the guitar, something catches their attention, they develop it into a progression, and then they try to figure out what to sing over it. Sometimes this produces good songs. More often it produces technically competent songs that don't mean anything in particular because they were built around a sound rather than a purpose.

The approach I use now is to identify the feeling or the situation first, before I pick up the instrument. Sometimes this is explicit. I know I want to write a song about a specific person or a specific moment. Sometimes it's more atmospheric. I have a feeling I've been carrying around and I need to find out what it is by trying to write it.

When I wrote "Tears" I knew exactly what I was trying to capture before I played a single note. The feeling of emotional lockdown. The specific frustration of knowing something needs to come out and being unable to access it. I had been living in that feeling for months during my recovery from the spinal condition. The song was the attempt to finally say the thing.

When I knew the feeling I was after, every decision that followed had a filter. Does this chord progression carry that feeling? Does this melody feel locked or does it feel like it's reaching toward something it can't quite get to? Does this lyric say the actual thing or is it circling around the actual thing and hoping the listener won't notice?

The feeling is the north star. Everything else is navigation.

THE MIRROR TEST

Here is a question I ask myself about every lyric I write before I commit to it.

Would I be embarrassed to say this sentence directly to the person it's about?

If the answer is no, the lyric is probably too safe. Good lyrics say the thing that is slightly too much to say out loud in polite conversation. They go one step further than the comfortable version. They name the thing that is usually left unnamed because naming it makes everyone in the room have to sit with it.

If the answer is yes, the lyric is in the right territory and the question becomes whether I have the courage to leave it in.

Most songwriters, including me for a long time, have a reflex toward softening. You write something raw and true and then the edit brain kicks in and says that's too much, pull it back, give people an out. That reflex is the enemy of meaningful songwriting. The listener cannot connect with the softened version because the softened version is not true. They can feel the distance between what the song is saying and what it actually means.

The version that makes you slightly uncomfortable to listen back to is usually the version worth keeping.

SPECIFICITY IS THE WHOLE GAME

Generic lyrics are the most common failure mode in songwriting and they are almost invisible to the person writing them because they feel meaningful from the inside.

Lines about love and loss and pain and hope feel important when you are in the middle of the emotion. But on the page, without the specific details that make an emotion belong to a particular human being in a particular situation, they dissolve into sentiment. Sentiment is not feeling. Sentiment is the performance of feeling. Listeners can tell the difference even when they can't articulate why one song hits and another one doesn't.

The fix is almost always specificity. Not "I watched you go" but where were you, what time of day was it, what were you wearing, what did the air smell like. Not "I was in pain" but what kind of pain, where in the body, what did you stop being able to do because of it.

You don't necessarily put all of those details in the lyric. But you write from the place of knowing all of them. When a lyric comes from that specific, embodied place it carries the weight of the specificity even when the lyric itself is relatively spare.

"Going Down" works because I knew exactly who I was writing it to and exactly what I was watching them do. The lyric is not full of identifying details but it carries the weight of a real situation because that's where it came from. Listeners hear the difference. They feel addressed by something that knows what it's talking about.

STRUCTURE IS A TOOL NOT A RULE

Verse chorus verse chorus bridge chorus is a template. It is not a law.

I use conventional song structures most of the time because they work and because they create expectations in the listener that you can use. The bridge is powerful partly because the listener has been conditioned to expect a shift there. The final chorus hits harder because the structure has been building toward it.

But the structure should always serve the emotional journey of the song and not the other way around. If a song needs to be two verses and a chorus and then stop, it should stop. If a bridge is going to deflate the momentum rather than deepen it, cut the bridge. If the song needs to repeat the chorus four times at the end because that's what it takes to let the feeling land, repeat it four times.

The question I ask about structure is the same question I ask about everything else. Does this serve what the song is trying to do? If yes, keep it. If no, it doesn't matter that it's what songs are supposed to do.

My song "Mistakes" has a structure that doesn't follow the conventional template in a couple of places. It does what it does because the emotional content required it. I tried the conventional version. It felt resolved in a place where the song should still feel unresolved. So I changed it. The song is better for not following the rule.

THE PRODUCTION IS PART OF THE MESSAGE

This is the part of songwriting that most guides treat as separate and it is not separate.

The way a song is produced, the arrangement, the tempo, the density of the instrumentation, the space between the notes, is part of what the song is saying. A lyric about vulnerability over a dense, aggressive production is saying something different from the same lyric over a sparse, quiet arrangement. The production is not decoration. It is argument.

When I recorded the guitar parts for "Tears" I made deliberate choices about how much space to leave. The song is about being emotionally locked, unable to access what needs to come out. A dense production would have contradicted that. The space in the arrangement is doing some of the emotional work of the lyric.

My Telecaster's clean tone for certain passages and the break into grit at others are not just sonic choices. They are structural decisions that support what the song is trying to communicate. The electrical engineering background helps here. I think about the signal chain as a system. Every element of the production is either carrying the signal forward or introducing noise. Strip out the noise. Keep what carries.

THE PART NOBODY TALKS ABOUT

The best thing you can do for your songwriting is finish songs you know are not good.

This runs against most people's instincts. You start a song, it's not going anywhere, you put it down and start something else. That feels like good judgment. It is actually the habit that keeps you from developing.

Finishing a bad song teaches you things about structure and craft and your own tendencies that abandoning it does not. You have to solve problems that arise in the second half of a song. You have to find an ending. You have to make a bridge work when the bridge is clearly the weakest part. All of that work, on a song you already know is not your best, is where the craft develops.

I have songs in my catalog that I would not play for anyone. They were necessary. They were the practice that made the songs I'm proud of possible.

Write the bad songs. Finish them. Learn from them. Move to the next one.

WHAT THIS ACTUALLY LOOKS LIKE IN PRACTICE

When I sit down to write now, here is roughly what happens.

I start by identifying what I am actually trying to say. Sometimes I write it out in plain language before I touch the guitar. Just the sentence. What is this song about, said directly, with no poetic language and no metaphor. That plain sentence becomes the anchor. When the lyric drifts away from that sentence I know it has drifted.

Then I find the feeling on the instrument. I play until something matches the emotional temperature of what I am trying to say. Not a finished progression. Just a direction. A key, a tempo, an energy.

Then I start putting words to it, starting with the chorus because the chorus is where the song's thesis lives. If I cannot write a chorus that says the thing directly and memorably I do not yet know enough about what the song is trying to do to write the verses. The verses are the argument. The chorus is the conclusion. Work out the conclusion first.

Then I apply the mirror test. Then I look for the soft spots in the lyric where I have circled around the thing instead of saying the thing. Then I cut those and replace them with the more uncomfortable version.

Then I finish it. Even if it is not working. Especially if it is not working.

Then I start the next one.

Tony Oso

Stream Tony Oso at tonyosomusic.com/music

If this was useful, share it with someone who is trying to write something real. Subscribe at tonyosomusic.com to stay in the loop on new music and new posts.