How To Write A Good Song

Most songs fail because people start at the wrong place

I've listened to thousands of home demos. The same three problems show up over and over again. Vague lyrics that could apply to anything. A chorus that adds nothing new compared to the verse. And a structure that feels like it's going somewhere but never quite arrives. Writing a good song isn't about inspiration. It's about making a series of decisions, some of them painful, that narrow the infinite number of choices down to the ones that actually work. Let me be honest about something most tutorials won't tell you: melody usually comes after chord progression, not before. I spent years trying to hum melodies first and then find chords to put under them. That approach produced mediocre results because the chords weren't built to support the melody's rhythm and phrasing. What actually works is building a chord progression that has its own internal logic, then writing a melody that treats those chords as a canvas rather than a cage. The melody should sometimes contradict the harmonic expectation, not always confirm it.

How To Write A Good Song: Start With a Restriction

The single most effective thing I've learned about songwriting is that unlimited freedom is your enemy. When I sit down to write, I impose at least one arbitrary limitation. Maybe it's a single chord tone I have to build the melody around for the whole verse. Maybe it's a lyric topic so specific that it feels almost embarrassing. Maybe it's a time signature that's unusual enough to prevent me from falling back on 4/4 muscle memory. Here's a real example from my own work. I was stuck on a track for what felt like weeks. The verse melody was competent but forgettable, and the chorus didn't grab me either. The problem was that I'd been writing the melody and lyrics in parallel, which meant they ended up with the same rhythmic patterns. The syllables landed on the same beats every time. My workaround was to strip everything back to just the chords and write the vocal melody as pure nonsense syllables, but with one rule: no two consecutive measures could have the same rhythmic pattern. It forced my ear to actually listen to what was different rather than what was similar. That constraint produced a hook that became the entire foundation of the track. I went from about 12 hours of struggle on that song down to roughly 90 minutes once I stopped trying to make everything "work together" from the start.

Lyric Writing That Doesn't Sound Generic

Most people write lyrics the wrong way around. They start with a theme, then try to dress it up with imagery. The better approach starts with concrete details and lets the theme emerge from them. This is why songs about specific moments stick with people. Not because the writer was trying to be universal, but because universality is a side effect of specificity, not the other way around. When I'm working on lyrics, I keep a simple spreadsheet. Columns for the scene, the sensory details, the emotion, and the line I want to land on. If I can't fill in the sensory details column, the scene isn't ready yet. I go back and actually observe something instead of inventing something. This is why writing from personal experience works when it's accurate and fails when it's exaggerated. Listeners can tell the difference between a remembered detail and a imagined one. A remembered detail has weight to it. The biggest pitfall in lyric writing is abstraction without evidence. Writing "I feel so alone" tells me nothing. Writing "I walked past your place and kept walking because I told myself I would" tells me everything. The second line works because it contains a decision, not just a declaration. Every line in a song should either advance the narrative, reveal character, or reframe something that came before. If a line does none of those three things, it's decoration and it's probably unnecessary.

Chord Progressions and Why They Matter More Than You Think

A good chord progression isn't about complexity. Some of the most memorable songs use four chords. The question is whether those chords are doing work beyond simply transitioning between the tonic and the dominant. I think about chord progressions in terms of tension geography. Where does the tension build? Where does it release? Where do you deliberately avoid resolution to keep the listener moving forward? One thing that took me a long time to understand: the pre-chorus exists to create harmonic movement that the verse alone cannot generate. If your verse already climbs to the relative major and back down, there's nowhere left to go in a traditional pre-chorus. That's why so many songs feel like they have no pre-chorus energy. The solution is to plan the harmonic trajectory across the entire section architecture before you write any individual part. Map out where each chord wants to go, then decide which chords you actually let it go to and which ones you sidestep. I use a technique where I write the bridge first. Not because bridges are the most important part of a song, but because they force me to think about what the song is not yet doing. If the verse and chorus stay in the home key, the bridge should move to something that feels genuinely distant. Relative minor, subdominant area, modal interchange with something like a IVmaj7 or bVII. The goal isn't exoticism. It's contrast that makes the return to the chorus feel like coming home.

Structure Decisions That Actually Matter

The most common structural mistake I hear is that every section sounds like it could be any other section. Verse sounds like chorus. Chorus sounds like verse. Bridge sounds like everything else. This happens when writers don't define what each section's job is and then give each one a distinct musical identity. Here's how I think about it: the verse is exposition. It introduces characters, situations, and unresolved tension. The chorus is the thesis. It's the point being made, stated clearly and memorably. The pre-chorus is the escalation. It takes the energy from the verse and pushes it toward the chorus. The bridge is the pivot. It forces a recontextualization of everything that came before. The outro is the resolution, whether that's a fade, a hard stop, or a return to the opening motif. Once you define what each section does, you make structural decisions with intent rather than habit. A song doesn't need a pre-chorus. Some of the best pop songs skip it entirely. But if you include one, it should earn its place by doing something the verse couldn't do on its own. Same with the bridge. If your bridge is just a variation of the chorus with different words, it's not a bridge. It's a second chorus wearing a disguise.

The Melody Question

Good melodies share two qualities: memorability and singability. They're memorable because they contain a distinctive interval or rhythmic motif that the ear can latch onto. They're singable because they respect the natural range and breathing patterns of a human voice. These two qualities sometimes pull in opposite directions. A melody that's interesting might jump around too much. A melody that's comfortable might be too predictable. The solution is to make the interesting parts concise and the predictable parts purposeful. I record every melody idea I have, even the bad ones. Voice memo on my phone, right away. The reason is that melodies have an extremely short shelf life in memory. I've lost ideas within 48 hours that I can clearly remember feeling like they were going somewhere good. The act of recording them removes the pressure of having to remember and lets me evaluate them cold later. This alone cut my average songwriting time from about three sessions per track down to one or two. One counter-intuitive technique that works well: sometimes the best melody for a song is not the one that starts in the chorus but the one that appears briefly in the verse and gets repeated in the chorus. This creates a sense of inevitability. The listener feels like the chorus was always there, even though it only showed up as a fragment earlier. That's how you build a hook that feels both fresh and familiar at the same time.

Arrangement and Dynamics

A song isn't just chords, melody, and lyrics. It's an arrangement. And arrangement decisions often determine whether a song feels complete or like a demo. The most impactful arrangement decision is often the simplest: what do you remove? Most amateur tracks are cluttered because every instrument is playing the whole time. The fix is to think in terms of density. Verse is sparse. Chorus adds elements. Pre-chorus builds. Bridge strips back or changes texture. Final chorus adds everything including something new. I follow a rule of thumb where no section should have more than four active elements at once unless there's a deliberate reason for the density. Four elements could be: lead vocal, bass, drums, and one harmonic or melodic instrument. That's already a full arrangement. Everything else is either doubling one of those elements or sitting in the background passively.

What This Approach Won't Do

I should mention where this method breaks down. The restriction-based approach assumes you're starting from scratch or near scratch. If you already have a strong melodic idea or a lyric draft that you're attached to, imposing arbitrary constraints can feel forced and counterproductive. In those cases, the better approach is to audit what you have and identify which element is the weakest link, then strengthen that element rather than rebuilding from zero. Another limitation: this framework doesn't account well for genre-specific conventions that exist outside the Western pop tradition. Microtonal music, complex polyrhythms, and modes that don't map neatly onto major-minor harmony require different tools. If you're working in those areas, the principles still apply but the specifics need to change. A bridge in a drone-based composition serves a different function than a bridge in a standard pop song. The process also assumes access to at least basic recording capabilities. If you're writing purely on paper without any way to hear your ideas, the melody and harmony feedback loop breaks down. You'll spend more time guessing than knowing. A cheap interface and free DAW software will solve this in about ten minutes of setup time.

Practical Workflow

Here's what my actual process looks like when I'm sitting down to write a new song. I start with a chord progression on whatever instrument I'm most comfortable with, usually guitar or keyboard. I play it through until I can hum something over it without thinking about the chords anymore. That hummed melody becomes the seed. I record it immediately. Next I write lyrics using the specificity rule. One concrete image per line. No abstract emotions unless they're grounded in a physical action. I aim for about 80 words total for a standard verse-chorus structure. Not a target to hit exactly but a rough guide to prevent bloating. Then I map out the structure. Verse, pre-chorus if needed, chorus, verse two, pre-chorus, chorus, bridge, chorus, outro. I write each section with its designated job in mind. I don't let the verse try to do chorus work or vice versa.

After that comes arrangement. I build the track section by section, adding elements rather than removing them. Each added element should serve a clear purpose. If an instrument part isn't serving exposition, escalation, contrast, or resolution, I cut it. The entire process, from first chord to final arrangement sketch, usually takes between two and four hours for a complete demo. That's after years of practice. The first time someone tries this, expect it to take longer. The structure reduces the number of decisions you have to make, which speeds things up once you internalize it. This is how I approach it. It's not the only way. Some writers work backward from lyrics. Some work backward from a drum pattern. But the underlying principle is the same: define what each part does, then give each part the tools it needs to do that thing well.