Verified against ChatGPT · 2026-08-14
Tear down per-unit or per-customer economics to find out if growth is actually making money or just making revenue
Breaks a business's per-unit or per-customer economics down into every cost that actually attaches to serving one more customer, distinguishing genuinely variable costs from costs that just feel variable, so growth decisions aren't made on a number that's quietly wrong.
The prompt
Ready to copy — highlighted parts are example details you can swap.
You are tearing down the unit economics of one customer or one unit of the offering described below — the goal is finding out whether each additional customer is genuinely profitable on a fully-loaded basis, not just contributing revenue that looks good until the fixed costs it's actually eating into get counted somewhere else. WHAT ONE "UNIT" IS One subscribing customer on the standard monthly plan. REVENUE PER UNIT $49/month recurring, plus a one-time $99 onboarding fee at signup. COSTS I'VE ALREADY IDENTIFIED AS VARIABLE Cloud hosting cost per active user ($3.20/month), payment processing fees (2.9% + $0.30 per transaction). COSTS THAT MIGHT BE PARTIALLY VARIABLE Customer support staffing — currently 1 support rep per roughly 250 customers; and a shared data pipeline cost that increases in server-cluster increments. GO THROUGH THIS SEQUENCE: 1. Confirm which of the identified variable costs are genuinely variable (scale directly with each additional unit) versus actually step-fixed (only increase in chunks, like needing to hire one more support person per 200 customers rather than a smooth per-customer cost) — a step-fixed cost misclassified as fully variable understates true unit economics at the margin right before the next step. 2. For each ambiguous cost listed, make a specific call: genuinely variable, step-fixed, or fixed, and state the reasoning in one line rather than leaving it unresolved. 3. Calculate contribution margin per unit using only the costs confirmed as genuinely variable, then separately show what the margin looks like if the step-fixed costs are averaged in at current volume — label this second number clearly as volume-dependent, since it will look worse as volume approaches the next step threshold. 4. Flag if the revenue-per-unit figure includes any one-time revenue (a setup fee, an upfront payment) that won't recur — blending one-time and recurring revenue into a single per-unit figure overstates the ongoing economics. Do not conclude whether the business overall is profitable — this is a per-unit teardown, and overall profitability also depends on fixed costs and volume that aren't part of this specific calculation. State that boundary explicitly. OUTPUT FORMAT 1. Cost classification table: cost | genuinely variable / step-fixed / fixed | reasoning. 2. Contribution margin per unit (variable costs only) and the volume-adjusted margin (including step-fixed costs at current volume). 3. One-time revenue flag, if applicable. 4. Explicit statement that this doesn't determine overall business profitability on its own.
Customize
Optional — swap in your own details for the highlighted parts above.
Why this works
The distinction between genuinely variable and step-fixed costs is the entire point of this prompt and the thing most unit-economics summaries get wrong by default — a cost like customer support staffing doesn't smoothly increase with each new customer, it jumps in discrete chunks (one more hire per 250 customers), so treating it as a smooth per-unit cost either overstates margin at 249 customers or understates it right after the 251st hire, and neither treatment reflects what actually happens to cash flow at the margin. Requiring a one-line reasoning for every ambiguous cost classification, rather than letting the model quietly default one way, matters because GPT-5.1 will otherwise often just average an ambiguous cost into a per-unit figure without flagging that the underlying cost structure is lumpy, which produces a contribution margin that looks precise but hides a meaningfully different number depending on exactly how close the business is to its next staffing or infrastructure threshold. Separating the fully-variable margin from the volume-adjusted margin that includes step-fixed costs gives the requester two honestly different numbers instead of one blended figure that obscures which scenario they're actually in — near a threshold, the volume-adjusted number is the one that matters, and collapsing the two into one average would hide exactly the information a growth decision needs. The one-time-revenue flag exists because a business's own booked revenue figures often bundle a one-time onboarding or setup fee into what looks like a steady per-unit number, and that one-time boost makes month-one unit economics look meaningfully better than the ongoing, recurring reality — separating them keeps the growth decision honest about what repeats and what doesn't.
What you get back
Customer support: step-fixed — one additional rep needed per 250 customers, not a smooth per-customer cost; at current volume this adds roughly $0.85/customer/month averaged, but that jumps discontinuously at the next 250-customer threshold. Fully-variable contribution margin (hosting + processing only): $49 - $3.20 - $1.72 = $44.08/month. Volume-adjusted margin including averaged support cost: $43.23/month, worse as volume approaches the next hire threshold. One-time revenue flag: the $99 onboarding fee is not recurring and should be excluded from ongoing per-unit economics.
Verified against
ChatGPT GPT-5.1 · 2026-08-14
Changelog
- 2026-08-14 — Initial publish, verified against ChatGPT GPT-5.1.
Need this built into your business?
If a prompt isn't enough — what Scult builds, built and maintained for you — that's Scult's day job.
EXPLORE WHAT SCULT BUILDS
