Part V — Release
Uploading is not releasing
Plan a release as three sequenced stages with a lead time, and name which of them you have never actually done.
The upload is one stage of three
You paid a distributor, filled in a form, chose a Friday, pressed go. Two weeks later the record is on Spotify and nothing has happened.
Nothing was supposed to happen. You did the first third of a release, in its thinnest form.
The most useful framing in our source material breaks a release into three stages that run in order: get the record onto every kind of platform, hand it to human curators, then meet the public (njtuGbE9HCg). Most independent artists execute the first badly, skip the second entirely, and arrive at the third with nothing built. What is left doing the work of a curator is the recommendation system — not chosen, fallen into.
His diagnostic is the part worth keeping. When a record underperforms, two different failures look identical from outside, because both are silence: the songs did not connect, or nobody built the infrastructure for anyone to hear them. You cannot ask the first question honestly until you can rule out the second.
Stage one: the platforms, plural
"Platform" means four intake paths, and only one is what you did.
- Streaming services, through a digital distributor.
- Physical retail, through one-stop distribution or direct to stores.
- Sync agents and music libraries, by email and the platforms supervisors already use.
- Direct-to-consumer — Bandcamp, your own store, your own site.
Which distributor matters, and the axis is not price. A subscription distributor is a landlord: stop paying and the catalogue comes down. A commission distributor earns only when you earn, so it has no reason to remove anything, which is how a back catalogue survives the years it earns nothing (BjBZ2x6oh1I). "Free" is a permanence decision, not a budget one.
Take that source's comparison checklist, not its rankings. Compare on: whether you can bring your own ISRCs, whether distribution is for life, the payout threshold, the fee, whether YouTube Content ID is included and at what percentage, support speed, and metadata and credit flexibility. The terms it quotes are dated and several came from forums — verify every number against the distributor's own current page.
The paperwork that has to be finished before launch, not after
Metadata is a stage-one deliverable, and the argument is about the cost curve rather than tidiness. You fix it now or you never fix it: once the information has propagated across every platform and you have moved to the next project, there is no moment you go back.
The harder version: register the work and open the receiving accounts before the first post goes out (hiETAHqNVUw). A record that takes off leaves you no time for tedious registration, and industry people who want to know who owns a song have to be able to find out.
There is a third reason and it is the one that costs money. A release whose metadata and registrations are wrong still generates royalties — they accumulate with no instruction about where to send them. That is Part II arriving as a release gate, and it is why this chapter has a dependency: The money already exists. If you have not done that chapter's work, this release is not ready.
While you are opening accounts, fix the link. The people arriving at your profile are several audiences — listeners, buyers, bookers, licensors, collaborators — and a bio containing one streaming URL turns most of them away silently, with no error and no signal that anything was lost (z3wzWEopweE). The link block is a routing layer.
Stage two: the curators, and this is the one you skipped
The source names the reason plainly, and it is not time: fear of rejection.
The targets are DJs — internet, club, mix show, college, independent — plus college and community radio, blogs, social curators, playlist curators, sync agents and libraries. Each has a delivery channel, and mismatching it is the failure.
| Who | How you actually reach them |
|---|---|
| DJs | Record pools, clubs, email — and a shared playlist they can take music from easily |
| College and community radio | Email and physical mail |
| Blogs | Email, plus DMs to the specific journalists who write there |
| Social media curators | Email and DMs |
| Playlist curators | Email and submission portals |
| Sync agents and libraries | Email and the library platforms supervisors already work in |
The case for humans is not sentimental. A recommendation carries a relationship a system cannot supply, and human placement puts a record where that music is already being looked for — where it has the best chance of a full listen rather than a skip.
Stage three: the public, and it is a timeline, not a day
Three questions define this stage: who your listeners are, how they like to receive music, and from whom and where. The output is a timeline, because consumption from radio, blogs, stores, streaming and social does not happen on one day.
The ordering is inverted from how it gets executed. Stage three is last in sequence and first in mind: how a listener will experience the record is what decides your stage-one choices.
Lead time, and why it exists
The cheapest test in this chapter. Write down the lead time on your last release — the gap between the record being finished and the record being out.
The source's line is that under six weeks you are not releasing, you are rushing, and his reasoning is what makes it usable: a short lead time is proof you serviced nobody, because there was no window in which you could have. Six weeks is his number, not a measured threshold. Keep the reasoning. Measure your own gap, then ask what you could have handed to a human being inside it.
The cadence argument, and why "a song a month" is contested
Monthly releasing is defended as consistency. Two of our sources attack it from opposite sides.
The first calls it habitual failure wearing the costume of consistency — dropping on schedule into a window you never chose (CDgc3otRb5M). That video also offers dead months, a canonical release day and paydays. All of it is US-shaped, none of it is sourced, and the speaker never says so. Take his three questions instead — when you are moving, why, and how — plus his one crisp sub-question: decide in advance what fraction of the goal still counts as enough to keep going.
The second attacks the labour (Xalr555OepY). A release is four jobs in parallel — finishing the record while chasing mixes on another, the asset pile (artwork, canvases, visualisers, sequenced so they do not collide), the admin (ISRCs, split sheets, producer agreements, clearances), and the promotional run for the song you put out three weeks ago. Thirty days does not contain four of those. His replacement is a longer gap — six to eight weeks — with most of the promotional weight after the release. That interval is one speaker's and is corroborated nowhere else here; carry it as a proposal, not a standard.
Two ideas from it survive independently of the number.
Stack the back, not the front. A team with no marketing budget cannot buy a launch spike, but it can keep pushing a record for as long as the platform still treats it as new. Week one is the music, release week is the story, and the weeks after are milestones celebrated with the people who caused them.
One cycle is longer than the gap between releases. Asset work and admin start before promotion, so cycles overlap — you read song A's numbers while song B is already in testing. Plan for the gap, discover the cycle, and you will conclude the plan failed. The same idea appears elsewhere as a scheduling rule: be several campaigns ahead, because the failure is exact — everything goes into one single, and two and a half weeks later nothing is queued (hiETAHqNVUw).
⚠️ That video also gives a campaign length and a momentum-decay window, and both are unusable. The length is stated two ways in one breath, the halves transposable. The decay figure contradicts itself inside one sentence — "more than 30 days, or really two or three weeks." Plan against neither. The shape is unambiguous: promotion lingers after the run and has to hand over to the next thing before the audience goes cold.
All of which argues for stacking a catalogue rather than front-loading one drop. A finite work closes a loop; an endless run of singles teaches an audience you will never deliver a finished thing (BC7TiuVATvM). A catalogue still distributed in five years is an asset. A single you rushed out to stay visible is not.
Serve everyone on release day, not later
A release that only puts music on streaming is built for the least committed third of your audience (47sXzoG6SN4). The named failure is serving the casual listener first and promising the rest "later down the line." The rule that goes with it: whatever a higher tier gets must be something a lower tier genuinely cannot reach, or the ladder collapses into one price. See Followers are not fans for the sorting exercise underneath it.
⚠️ That video's title promises a 2000% boost. The number is never said aloud in the video and nothing in it is measured. We name it so you do not meet it elsewhere and believe it.
What to do this week
- Write down the lead time on your last release, then name one human curator you could have
serviced inside that window.
- Pick three curators and find their real delivery channel — inbox, portal, pool or postal
address, not their platform. Three, not thirty.
- Confirm the metadata, registrations and receiving accounts are done for the record you are
about to put out. If they are not, that is this week's release work and the date moves.
- Look at your link in bio and name who it turns away. If a booker, a buyer or a supervisor
arrived today, is anything there for them?