What you'll be able to do
- Read the difference between a release that's reacting and one that's delivering
- Scale a winner without breaking the learning that made it win
- Run the learnings doc that turns every campaign into label IQ
The short version: Module 3's week-three gate finds the winners; this module spends on them without wrecking them. The tells that separate delivering from reacting: retention-shaped signals (saves, listener repeat, pool-conversion rates) over spike-shaped ones (a big day, a repost bump). Scaling doctrine: raise budgets in steps the learning can absorb (the DAS 100 Module 5 rules, applied with label patience), widen audiences down the warmth ladder before widening geographies, and add lanes (sustain tail → discovery lane → catalog-farm slot) rather than inflating one lane past its efficiency. And every scale decision - win or lose - lands in the learnings doc, the label's compounding IQ: what pattern, what worked, what it cost, what the baseline becomes.
Feeding the fire without smothering it
If this is your first rodeo: "scaling" = adding budget to what's working. It's the moment most campaigns get worse, because money changes the machine that was winning - bigger budgets reset learning, wider audiences dilute the signal, impatience buys regression. The label version has an advantage: no single release HAS to scale (the roster hedges), so scaling happens on evidence, calmly, or not at all.
Reading the tells
- Delivering (scale-worthy): save rates above baseline · listener retention (repeat listens, follower conversion) · pool-conversion efficiency holding as spend grows · the smart-link's service click-through cost stable or improving · organic motion CONTINUING after the initial push (the song moving on its own legs)
- Reacting (not yet): a spike with a cause (the repost, the placement, the algorithm's Tuesday mood) that decays on schedule · engagement without saves · cheap clicks, flat listeners (the click-species warning from DAS 202 Module 4) · numbers that only move when spend moves
- The discipline: reacting releases get WATCHED (another week costs nothing); delivering releases get the ladder below. Scaling a reaction buys a bigger spike and a steeper decay, at markup.
The scaling ladder (in order)
- Extend duration (the sustain tail funded - cheapest scale that exists)
- Step the budget (increments the learning absorbs, per DAS 100 Module 5 - never the double-overnight)
- Descend the warmth ladder (warm exhausted → LBL scene pools → interest-adjacent cold - widen the WHO before the WHERE)
- Add lanes, don't inflate lanes (the discovery lane joins; the catalog-farm slot follows; each lane at its own efficient size beats one bloated lane)
- The $500-shape top-up (Module 3's priority build, arrived at by evidence instead of scheduled by hope)
›Go deeper: the learnings doc (the label's compounding IQ)
The artifact that makes release 30 smarter than release 3: one shared doc, one entry per campaign - the release, the build, the week-three reading, the scale decisions and their results, the creative patterns that worked (format, hook shape, identity), the costs (which become next quarter's baselines), and the one-line lesson. Anonymized patterns feed the roster (Module 2's shared-by-default column); per-artist specifics stay scoped. The doc is what the week-three gate reads its baseline FROM, what the catalog farm consults for rotation candidates, what Module 6's signing conversations quote, and what makes the label's taste falsifiable - the entries where the doc was wrong are the most valuable ones, because a label that can be wrong on paper can get smarter on paper. Fifteen minutes per campaign close. The compounding is the point.
From the field
A label's breakout moment arrived wearing a disguise: a quiet release, week-three numbers just above baseline - saves strong, listeners repeating, nothing viral. The group chat wanted to scale the OTHER release (bigger spike, a repost from a name account, everyone excited). The gate said otherwise: the spike was reacting (decayed on schedule, saves flat), the quiet one was delivering. The ladder ran on the quiet one - sustain tail, stepped budget, warmth descent, discovery lane at month two - and it became the label's biggest artist-development win of the year, at costs the spike release never approached even during its big week. The learnings-doc entry ran one line long: "The loud one was weather. The quiet one was climate. The gate knew first."
Common mistakes
- Scaling the spike (buying a bigger reaction at markup)
- The double-overnight (learning reset, efficiency gone, lesson expensive)
- Widening geography before warmth (the WHERE before the WHO)
- One bloated lane instead of several efficient ones
- No learnings doc - release 30 exactly as smart as release 3
Take-home
The Winner's Ladder Card - the delivering-vs-reacting tells table, the five-rung scaling ladder, the step-budget rules, and the learnings-doc entry template.
Sources & dates
DAS 100 M5 + DAS 202 M4 cross-references as cited · the ladder is Darkroom practice, presented as such
Learn more
Optional extra credit. Nothing below is required for the quiz, the certificate, or the job - the full lesson is above. This is for the sickos who want more.
- DAS 100 Module 5 - the learning-phase mechanics every scaling rule protects