The record-label carousel maps a production workflow in which agents prepare work and people make the calls. Taste leads to a brief; music and visuals develop in parallel; quality checks route revisions; a person approves the master, release and spend. This guide turns that map into a planning worksheet. It describes a workflow to scope and test, not a promise of an autonomous profitable label.
Before you start
Choose one test release and name a person accountable for it. Keep a release record with a unique ID, brief version, asset versions, owners, review status and decision history. Define the tools, access and budget available for this pilot. Store input permissions and source information alongside the assets so reviewers can assess what is ready and what still needs checking.
The workflow
- Define taste and the brief: record the audience, musical direction, mood, references you are allowed to use, exclusions and acceptance criteria. A&R converts those choices into clear music and visual assignments. Version the brief so later changes are visible.
- Create music and visuals in parallel: give each branch the same release ID and approved brief. Track drafts separately from accepted versions. Require a responsible owner for each output and prevent a draft asset from being mistaken for the final release package.
- Compose, mix and check: run the agreed technical checks and a listening review. Log the check, result and asset version. Failed engineering checks return to the engineer. Creative revisions return to the composer with specific feedback and the current brief.
- Approve the master: assemble the proposed master, review notes and unresolved issues. A named person accepts or requests revisions. Tie the decision to an exact version; replacing the audio invalidates that approval until the replacement is reviewed.
- Approve release and spend: assemble the selected master, artwork, credits, metadata, permissions evidence, release plan and proposed budget. Keep the package blocked until the responsible person has reviewed it. Approval of an earlier draft or a different budget is not approval of this package.
- Distribute and observe: submit only the approved package through the chosen distribution process. Record the provider receipt and actual status. If submission fails, preserve the attempt and investigate before retrying so the same release is not submitted twice.
- Approve learning: collect available release signals with their dates and sources. Separate observations from interpretations. Propose changes to the next brief, then let a person decide which lessons enter the shared working guidance.
Worked example
Illustrative example: release R-001 has master version 3 approved. An artwork revision arrives while a revised mix creates master version 4. The workflow holds the package because version 4 has no master approval. After review, a person approves the exact audio and artwork versions and the stated spending limit. A later suggestion to increase spend needs a new decision. No revenue or audience outcome is assumed.
Your worksheet
- Release ID, accountable owner and brief version: ___
- Music and visual owners: ___
- Required checks and revision routes: ___
- Approved master version and decision record: ___
- Artwork, metadata and permissions review status: ___
- Release and spend approver, scope and limit: ___
- Distribution receipt and failure route: ___
- Learning proposal, evidence and human decision: ___
Test it
- Replace an approved master with a new version; confirm release is blocked until that version is reviewed.
- Remove a required package check or change the budget; confirm the previous approval cannot authorize the changed package.
- Simulate a failed submission and an unapproved learning suggestion; confirm neither becomes a successful release or accepted guidance automatically.
What to measure
Track revision cycles, blocked handoffs, failed submissions and the time people spend reviewing. Keep production reliability separate from commercial results. Label streams, merchandise or sync as possible business channels rather than guaranteed revenue, and make future decisions from observed evidence.