Somewhere in your headphones right now, a song is doing one of two things: helping you work, or quietly charging you rent for the company. Most of us never check which. We press play out of ritual, out of reflex, and out of something more tender that we rarely say out loud — working is lonely, and silence makes the loneliness audible. The playlist isn't really a productivity tool. It's a companion. Which is exactly why the honest research on music and focus can feel almost rude: it says the answer to "does music help you focus?" is neither yes nor no. It depends on what you're asking the music to do. And most of us are asking it to do the wrong job.

The question isn't "music or silence"

Decades of research on background sound and cognition keep circling the same conclusion: there is no blanket answer, because three variables keep interacting. What kind of task you're doing. What kind of sound you're playing. And what kind of nervous system you brought to the desk.

Change any one of those and the effect flips. The same playlist that carries you through a mindless spreadsheet cleanup can quietly sabotage you an hour later when you're trying to write a difficult email. It's not that music "works" or "doesn't work." It's that music is a cognitive load with benefits, and whether the trade is worth it depends on what you're spending attention on.

Once you see the three variables, you can stop arguing about music in the abstract and start making the trade deliberately.

Lyrics compete for the same channel as your thoughts

Start with the sound itself. Your working memory isn't one undifferentiated pool — in Alan Baddeley's influential model, it includes a component often called the phonological loop, a kind of inner ear and inner voice that holds verbal material while you think with it. Reading, writing, memorizing, planning in words: all of it runs through that loop.

Here's the problem. Lyrics are language, and language is hard to ignore. Research on what psychologists call the irrelevant sound effect has shown that background speech disrupts verbal tasks even when you're genuinely trying to tune it out — the words leak in whether you invite them or not. You don't get to decide that the chorus won't be processed. It gets processed, and it gets processed on the same machinery you were using to hold your own sentence in mind.

It's not only about words, either. Work by researchers including Dylan Jones on the changing-state effect suggests that what disrupts short-term memory most is sound that changes unpredictably — shifting tones, staccato variation, anything your auditory system has to keep re-registering. Steady, stable sound is far less costly than sound full of surprises. This is why a podcast is worse than a lyrical song, a lyrical song is generally worse than dramatic instrumental music, and dramatic instrumental music is worse than steady ambient sound — roughly in order of how much novelty and language they push through your ears.

So when people say "I write better with music," the research suggests a more precise translation: they may write better with certain music — and almost certainly not because the lyrics are helping.

The Mozart effect was never about Mozart

But wait — wasn't there a famous study showing music makes you smarter? Sort of, and its afterlife is one of the great cautionary tales in popular science. In 1993, researchers reported that listening to a Mozart sonata briefly improved performance on a spatial reasoning task. The press turned a small, short-lived effect into "Mozart makes babies smarter," and an industry was born.

The follow-up work told a humbler, more useful story. Psychologists including Glenn Schellenberg and William Forde Thompson showed that the boost had little to do with Mozart and a lot to do with arousal and mood: pleasant, upbeat music temporarily lifts your alertness and emotional state, and an alert brain in a good mood performs better on many tasks. Other enjoyable things produced similar effects. The music wasn't installing intelligence. It was adjusting your state.

This is the most practical finding in the whole literature, because it separates music's two jobs. As a state manager — something that wakes you up, lifts your mood, and gets you to the desk — music is genuinely effective. As a soundtrack running underneath complex thought, it's a cost you're choosing to pay. The trick is to collect the state benefit without paying the interference tax, and the simplest way to do that is timing: use the potent, lyrical, emotional music before and between work, and switch to something steadier during it.

The task and the person set the price

Two more dials decide whether background music is a net gain.

First, the task. Repetitive, mechanical, well-practiced work — sorting files, data entry, chores, workouts — tends to benefit from added stimulation, because the main enemy there is monotony, not overload. Complex, novel, language-heavy work flips the equation: the task already consumes most of your capacity, and every additional stream of input comes out of the same budget.

Second, the person. Personality research, notably by Adrian Furnham and colleagues, has found that introverts tend to pay a higher cognitive price for background music and noise than extraverts do — consistent with the older idea that introverts run closer to their optimal arousal level and are more easily pushed past it. If your extraverted friend swears by studying in loud cafés and you leave the same café feeling scraped thin, neither of you is doing it wrong. You're running different hardware.

Familiarity matters too. New music tugs at attention because your brain orients to novelty — that's what novelty is for. Music you've heard a hundred times fades toward wallpaper. This is why a well-worn album often costs less focus than a fresh algorithmic playlist that keeps serving you interesting surprises. Interesting is precisely the problem.

When music genuinely earns its place

Put it together and music earns its keep in four honest ways. It masks worse sound — unpredictable office chatter is among the most disruptive things you can work near, and steady self-chosen sound trades an uncontrollable distraction for a controllable one. It offsets monotony during mechanical work. It manages your state before hard work begins. And — the most underrated one — it can become a starting cue: play the same track at the start of every session and, through simple association, the music itself starts to mean we're working now, lowering the friction of beginning.

What it doesn't do is make difficult thinking easier while it plays. Nothing does. The work is the work.

Your next moves

  • Run a two-session test today. Do two similar blocks of the same kind of task — one with your usual playlist, one with steady instrumental sound or silence. Immediately after each, jot one line: how often you re-read, how drained you feel. Trust the notes over the vibe.
  • Split today's task list into "language-heavy" and "hands-heavy." Writing, reading, studying, planning go in the first pile; email triage, cleanup, admin in the second. Make a one-line rule: lyrics only for the second pile.
  • Pick one start track. Choose a familiar instrumental piece and play it at the beginning of every focus session this week — same track, every time. You're deliberately building a conditioned cue for starting.
  • If your environment is noisy, mask it with steady-state sound. Rain, brown noise, or a low ambient loop beats both raw chatter and speech-adjacent audio like podcasts, because it changes the least.
  • Move the hype song to before the session. Use the emotional, lyrical music to get yourself to the chair and lift your mood — then switch to your steady work sound when the actual session begins.

Where a timer makes this automatic

That last structure — cue, then state, then work — is really a habit stack wrapped around a focus session, and it's the exact shape Tally is built for. Tally is a premium iOS app that combines habit stacking with a Pomodoro focus timer, so "start track → timer on → steady sound → work" stops being a plan you have to remember and becomes a chain that runs itself, session after session. If you want that scaffolding around your next block of deep work, you can find it at tally.lumenlabs.works.