WHY YOU CAN’T AFFORD TO GET YOUR MUSIC METADATA WRONG

Nobody ever finished a mix and thought, right, now for the exciting part: the spreadsheet. But the twenty minutes you spend filling in track titles and credits decides who gets paid for the next seventy years, and most artists give it about four.

I’ve watched a release go out with the drummer’s name spelled two different ways across nine tracks, and it split the album into two artist profiles before anyone noticed. It took months to unpick.

Metadata is just the information attached to your recording: who wrote it, who played on it, when it came out, what it’s called, and the codes that let a computer tell your track apart from every other track ever made. Deeply unglamorous. Also the single most expensive thing to get wrong at this level, because the mistakes are silent and they compound.

Strip away the jargon and metadata is a label on a box. The box holds your audio. The label says what’s inside, who made it, and who to pay. The label travels with your track everywhere it goes, and nobody along that chain ever opens the box to check.

Practical metadata includes the track title, artist name, album title, release date, genre tags, the writers, the performers, the producer, the publisher, the copyright lines, and the identifiers that tie it all together. Some of it sits inside the audio file itself, in the tag formats you’ll have seen labelled ID3v1 and ID3v2 if you’ve ever opened a file’s properties. Most of what matters commercially, though, sits in your distributor’s database rather than in the file, because that’s the version the streaming services read.

Two identifiers do most of the heavy lifting. The ISRC, or International Standard Recording Code, tags a specific sound recording. Your song, your remix and your live version each need their own, because they’re three separate recordings of one musical work. The ISWC does the other job: it identifies the composition, the thing the songwriters and composers actually own, regardless of who records it. If you don’t have your codes sorted yet, we handle ISRC allocation for artists pressing with us, and it takes about a day.

There’s a third layer people forget entirely, which is disc metadata. If you’re pressing physical, the track names that appear when somebody puts your CD into a laptop come from the CDDB Gracenote database, not from the disc. Get that submitted and your album shows up properly named. Skip it and it shows up as Track 01 through Track 11, which looks about as professional as it sounds.

Here’s the part that stings. When metadata is wrong, nothing breaks. No warning appears. The track streams perfectly, the artwork loads, everybody’s happy. The money just goes somewhere else.

Royalty payments are matched by machine against name strings and codes. If your PRS registration says “Jonathan Fraser” and your distributor submitted “Jon Fraser”, those are two people as far as the matching system is concerned. Your publishing share sits in an unmatched pot waiting for somebody to claim it, and a good chunk of that pot is never claimed at all. Estimates of unmatched global royalties run into the hundreds of millions annually, which tells you how normal this problem is.

A misspelled name does something else too. It can put your release on the artist page of another artist entirely, or spawn a duplicate profile so your monthly listeners split across two entries and neither looks like momentum. I’ve seen a band lose a verified profile this way because a compilation submitted their name with a “The” on the front and the platforms treated them as different artists.

Wrong release dates are the other classic. Enter the date in US format when your distributor expects UK format and your carefully planned Friday release lands on a Tuesday in March, months out of sync with your press. Incorrect metadata can lead to a release that technically exists and practically doesn’t.

Then there’s licensing. Music supervisors searching for a track need to clear it fast, and if the publishing metadata is missing or the writer splits aren’t documented, they move on to the next option rather than spending three weeks finding you. Music libraries reject submissions on this basis constantly. Losing money is one thing, but losing the sync placement you’d have got is worse, because that’s the one that pays properly.

Recommendation systems don’t listen to your record. They read the file about your record. Genre tags, mood descriptors, credits, related artists, language, whether it’s explicit: that’s the raw material a recommender system uses to decide when to recommend your music to somebody who’s never heard of you.

Tag a delicate acoustic track as “rock” because it felt closer than the alternatives and you’ll be served to people expecting something loud, who’ll skip in eleven seconds. Skip rate feeds back into whether the algorithm keeps trying. Music discovery at the indie level is largely a metadata problem dressed up as a marketing problem.

Playlist pitching runs on the same fuel. A Spotify editorial curator sees your genre, your mood, your instrumentation and your credits before they hear anything, and platforms like Spotify and Apple both need that submitted properly and early. Miss the window because your release date was entered wrong and you’ve forfeited the pitch entirely. Nothing you do afterwards gets it back.

Credits matter more than people expect here as well. Listing your guitarist and your producer properly connects your record to their other work, which is a genuine discovery route. Somebody who loved a producer’s last project can find yours. That only happens if the names are spelled the same way in both places.

None of this is complicated. It’s just fiddly, and fiddly work gets postponed until the deadline eats it. So here’s the list I’d work through, ideally three weeks before anything is submitted rather than the night before.

Build one source of truth

Open a single sheet. One row per track. Columns for title, exact artist name, all writers with full legal names, performers, producer, ISRC, publisher, PRS numbers, splits as percentages, release date, genre, and whether it’s explicit. Everything you submit anywhere gets copied from that sheet. Nothing gets typed twice from memory.

That sheet is your metadata management system. It sounds grand and it’s a spreadsheet, but the value is that there’s exactly one place to check, and when a collaborator asks what their split was you have an answer rather than a recollection.

Lock your artist name

Decide the exact string, including capitalisation, punctuation and any “The”. Write it at the top of the sheet. Use it everywhere without exception. Inconsistent metadata across platforms is the single most common way indie catalogues fragment, and it’s completely avoidable.

Sort splits before anybody submits anything

Percentages, in writing, signed, covering every writer on every track. Add up to 100 on each one. Do this while the sessions are fresh and everyone’s still pleased with the record, because the conversation gets considerably harder eighteen months later when there’s actual money involved.

Double-check the dates and the codes

Read the release date back to yourself in words. Check every ISRC is unique per recording and that no code has been reused from a previous version. Confirm your cover art meets the spec your distributor asked for, since a rejected image can delay the whole release and nobody tells you until it’s late.

Check what’s live after release day

A week in, look at your release on the DSPs. Spotify, Apple Music, and whatever else you’re on. Check the spelling, check the credits, check the track order. Then look at your analytics for anything odd, like a track with almost no plays, which usually means it’s sitting on a duplicate profile somewhere. It fits naturally into the wider release day checklist we put together.

Pressing physical adds a layer people don’t think about until the proofs arrive. What’s printed on the disc and the sleeve is metadata too, and it’s the version that can’t be corrected with a form submission six weeks later. Once it’s printed, it’s printed.

The spine, the back panel and the disc face should carry the same track titles in the same order as the digital release, with the same spellings. Copyright lines need both the ℗ for the sound recording and the © for the artwork and composition, each with the correct year and owner. Writer credits belong on the back or in the booklet, and if you’re pressing a booklet, that’s where full credits genuinely earn their space. Our CD booklet guide covers how to lay that out without it looking cramped.

From where I sit at CD Unity, the artwork proof stage is the last honest chance to catch this, and it’s the stage people rush hardest because they’re excited and the deadline is close. Read every track title out loud against your sheet. Check the running order twice. We flag what we can, but we can’t know that your bass player’s surname has an “h” in it.

Barcodes matter here too. Your UPC or EAN ties the physical product to the release in retail systems, and it should match the one your distributor holds for the digital version. Different codes for the same album means the sales data lands in two places and neither one shows the full picture.

Josh McKenzie

Hello, I’m Josh, and I’ve been honing my graphic design skills for almost 15 years now, catering to the needs of bands and businesses alike. What really fascinates me is the business aspect of the music industry. In addition to my design work, I also happen to play the Hammond organ, and I strive to share my knowledge through helpful articles that I write exclusively for you all!