REAL NOISERECORDS
ArtistOS / Product 02 Decision Lab / Version 1.0

Test before you scale.

Spend on evidence. Not guesses.

RN Signal helps independent artists compare two to four songs, clips, or campaign ideas before putting real money behind one.

Run each idea under similar conditions, enter the results you already have, and get a plain-language next move: advance, retest, or pause. Nothing is uploaded, and every saved test stays on this device.

Compare my ideas
Current status

Live calculation tool. No account required, no upload required, and no data is sent to RN Records.

Version 1.0
01 Decision lab

Free decision tool

Compare your
ideas clearly.

Use results from the same platform, audience, format, budget, and test window. RN Signal compares them against your targets and flags small or uneven sample counts. You still need to check how the test was run; the tool does not validate fairness or statistical significance.

RNSignal

Decision Lab / V1.0

Private on this device
How it works Same playing field. One clear decision.
  1. 01Name the decisionChoose what you are comparing and why.
  2. 02Set the benchmarkSet a goal from comparable past results; RN’s values are examples.
  3. 03Add your resultsUse the same platform and test window for every idea.
  4. 04Get the next moveAdvance, retest, or pause—with the reason shown.
Example loaded The numbers below are a walkthrough—not RN analytics. Replace them with your results or begin with a clean test.
02Choose what success means

These are RN starter targets for the example, not industry averages or platform success rules. Set your targets before looking at this test, using comparable past results or a clearly stated goal. Keep one definition and denominator for each metric across every idea. If you do not have a useful baseline, treat the score as a practice comparison and validate the result in a small next test.

Watched to endPercent of the measured people who completed the same defined clip or listening window. This is not average percentage viewed, average watch time, or YouTube Shorts stayed to watch. Use only a report or test that measures this completion outcome.

ReplayedPercent of the measured people who replayed at least once, counted once per person. Total extra plays or automatic loops are different metrics and should not be entered as this rate.

Saved or sharedDefine one interest outcome before the test, such as saving the track. If using saved OR shared, count each person once even if they did both; do not add raw event counts. Use the same outcome and cohort across ideas. Spotify campaign intent rate uses converted listeners, so do not paste it into an exposed-audience rate without reconciling the denominator.

Took actionPercent of that same cohort who completed the one call to action you chose before the test.

How the score is calculated

Signal Score V1.0: divide each measured rate by its target, cap that ratio at 1, then multiply by its weight: completion 35, replay 20, interest 25, action 20. Add the four components and round to a whole number. Meeting every target earns 100 points; that is not a 100% conversion rate.

RN starter rules: every idea needs at least 100 measured people and count spread no greater than 20% of the mean: (largest count − smallest count) ÷ mean count. Advance also needs a leading score of at least 75 and a lead of at least 10 rounded points. When count rules pass, a leading score below 50 gets Pause; otherwise Retest. These are RN planning choices, not a statistical sample-size calculation or proof of a winner.

03Add each idea’s results

Sample size means unique people in the comparison cohort, not total views or impressions. For this simple model, each rate is people who completed that outcome ÷ people in that cohort × 100. Keep definitions and denominators the same across ideas. If your report cannot supply one of these rates, do not guess or enter zero: compare its original metrics directly instead.

A
B
C
04Review the next move
Example / counts within RN starter limits
AdvanceMeets RN’s starter rules—consider a small next test. RetestPromising, close, or missing stronger evidence. PauseWeighted score below RN’s starter threshold—review before adding spend.
APerformance clip
82Advance
BVideo teaser
64Retest
CLyric visual
39Pause
What to do next

Performance clip leads by 18 rounded score points and meets RN’s starter rules. Consider a small next test if the audience and measurements were comparable. Check actual listening or action results before increasing spend; this score does not establish statistical significance.

Use this as a planning aid. Passing the count rules does not establish comparable audiences, random assignment, statistical significance, or future campaign performance.

Your numbers stay in this browser. RN Records never receives them.

Targets + measured results + count flags + next test.

02 Planned / Hosted testing

One question.
One clean test.

A hosted version with media uploads and public test links is planned. Today’s Decision Lab uses results you enter and saves your work on this device.

  1. 01
    Create the test

    Choose one decision: next single, song section, opening, cover, or campaign creative.

    Artist
  2. 02
    Add 2–4 variants

    Upload comparable audio, video, or image options and name the one variable being tested.

    Artist
  3. 03
    Share one public link

    Qualified visitors receive a randomized variant so they cannot simply choose the one they already recognize.

    Signal
  4. 04
    Capture consented response

    Measure sample size, watched-to-end, replays, saves or shares, and the one action chosen before the test.

    Signal
  5. 05
    Make the next move

    Show the leading score, experiment-quality checks, and an Advance, Retest, or Pause recommendation.

    Artist
Available today

Use the Decision Lab to compare your own campaign results. Its scoring method is explained below, and every recommendation includes the reason behind it.

03 Trust the method

Useful data
has rules.

RN Signal should make a decision clearer—not make weak evidence look scientific.

References reviewed October 1, 2026. These sources explain measurement and experimentation principles. RN sets the weights and recommendation cutoffs; you can edit the targets. The example targets and starter rules have not been validated as predictors of campaign performance.

01 / Ask

Write the question first.

Set one falsifiable idea and the success metric before results arrive. A vague “which is better?” test invites a convenient answer.

Microsoft Research ↗
02 / Fairness

Keep the sample fair.

Random assignment matters. If one idea reaches far more people or a different audience, pause the decision and find out why.

Microsoft Research ↗
03 / Attention

Count attention, not starts.

Completion and replay measurements can add context beyond starts. Replays may reflect interest, looping, or confusion, so check the surrounding behavior. Compare clips of similar length and format using the same measurement definitions, audience and test window.

YouTube Analytics ↗
04 / Interest

Look for a next action.

Choose the action that matches the campaign goal. Saves and playlist adds can show intent to return; clicks and profile visits tell you something different. Track the next listening or business result too, and keep each metric's denominator visible.

Spotify for Artists ↗
05 / Integrity

Reject fake response.

Bots, paid streams, forced traffic, and guaranteed-placement schemes poison the evidence and can damage the release.

Spotify policy ↗
06 / Learn

Verify after release.

A pre-campaign win is a hypothesis. Track whether the chosen idea actually improves qualified listening, action, or revenue.

Blueprint module 12 ↗
04 What’s next

A focused tool.
Room to grow.

Development starts with real artist needs: useful comparisons, clear decisions, and less time spent assembling campaign results.

Now

Decision Lab

Compare two to four ideas, save a test on your device, and copy a plain-language recommendation.

Live here
Next

Hosted pilot

Accounts, secure media upload, randomized public test links, consented event capture, and traffic-quality checks.

Planned
Proof

Real campaigns

Use lessons from RN releases and artist feedback to improve the comparison process.

Planned
Later

Earned intelligence

Explore comparable benchmarks and stronger recommendations as reliable campaign data becomes available.

Exploring

Free operating manual

Need the whole release system?

Open RN Artist Blueprint

Real upcoming release

Want to test the hosted pilot?

Bring RN a real campaign