The Money You Actually Keep
A stage-by-stage walk down the margin waterfall: what happens to a single dollar between the price on the page and the money in your account. The same seven stages for ecommerce, services, subscriptions, and lead gen. Only the words change.
The one idea underneath all of it
Ask a founder how much money they make per sale and most can't tell you. Not the revenue. Not the gross profit. The real number, after the processor takes its cut, after the freight, after the return that came back unsellable, after the ad spend. The number that decides whether scaling makes you richer or just busier.
The reason it's hard isn't that the math is complex. It's that almost every calculator and spreadsheet gets it structurally wrong: applying refund rates to revenue, treating shipping as recoverable, stopping at gross margin and calling it done. The errors are small per order and ruinous at scale. Half a margin point, multiplied across a year of ad-spend decisions, is real money built on a phantom.
So here's the whole thing, stage by stage, in the order a dollar actually flows. The math is identical for every business type. An ecommerce brand, a consultant, a SaaS app and a lead-gen agency all resolve to the same waterfall. The only thing that changes is what you call each line. The left column is the mechanism (the formula, worked). The right is the why (what it means, the lie most people believe, and how the words change by business type).
Gross margin is not the number
The real figure is contribution margin per order, after every variable cost and the true cost of refunds.
Percentage costs and fixed costs behave differently
Raise price 10% and your margin doesn't rise 10%. Percentage fees eat in; fixed fees shrink as a share. That's operating leverage.
Refunds are not a revenue haircut
A return loses the sale, may return the goods, and never returns the shipping or the processing fee. Three rules, not one.
Break-even is a boundary, not a target
Your margin sets the most you can pay to acquire a customer. Past it, every extra sale loses money. Faster.
The four things that decide whether the number is real
These aren't stages, they're the conditions that make the whole waterfall trustworthy. Get them wrong and every figure downstream is confidently false.
- Per order, not totals Profitability is decided one transaction at a time. Blended monthly P&L hides the products and channels that lose money on every sale. Work at the unit.
- Money is not a float 0.1 + 0.2 does not equal 0.3 in most software. Tiny rounding errors compound into real discrepancies at volume. Calculate money in exact decimals, never floating point.
- Cash flow, not accounting fiction The question is what actually leaves and enters your account, and when. A "5% refund rate" is meaningless until you know what comes back sellable and what's gone for good.
- Per-product, not global averages A 10% return rate with 90% resellable is a different business from 10% with 0% resellable. Aggregate rates lie. Recovery is a per-product fact.
From the price on the page to revenue you keep
The number a customer sees is rarely the number you bank. Tax and discounts come out before a single cost is counted. Start the waterfall honest, or everything below inherits the error.
The first ten dollars were never your money.
If your price includes tax, a chunk of every "sale" is collected on the government's behalf and passes straight through. Counting it as revenue inflates every margin downstream.
Treating the sticker price as revenue. It feels like $110 came in; only $100 did.
| Model | The "price" |
|---|---|
| E-com | Retail price |
| Service | Fee / day rate |
| Subscription | Monthly plan |
| Lead gen | Deal value |
What an average transaction is really worth.
Most businesses sell more than one unit per order and discount more than they admit. Both move the real figure before costs even enter.
Modelling a single unit at full price. The discount you "rarely" give is averaged across everyone and quietly lowers the line you build everything on.
E-com: items per cart. Service: scope / hours per engagement. Subscription: usually Q = 1, the plan. Lead gen: contract size.
The two kinds of cost, and why the difference is everything
Costs split into the cost of the thing itself, and the cost of selling it. Within selling costs, the split between percentage-based and fixed is the single most misunderstood lever in unit economics.
The cost of the thing, fully loaded.
Every cost to land the product in a customer's hands, not just the supplier invoice. Inbound freight and packaging are real and routinely forgotten.
Counting only the supplier price. The freight to get it to your warehouse and the box it ships in are COGS too, and they're often 15-20% of the "true" unit cost.
E-com: supplier + freight + packaging. Service: labour + travel. Subscription: infrastructure / hosting. Lead gen: fulfilment cost.
Where raising your price doesn't do what you think.
Percentage fees (processing, marketplace) scale with price. Fixed fees (a flat shipping label, a $0.30 transaction fee) don't. This split is operating leverage, and it's why a 10% price rise never gives you a clean 10% more margin.
Lumping all fees into one "fees" percentage. Do that and your scenario modelling is wrong every time you change price, because half your fees didn't move and half did.
E-com: processing + marketplace + shipping + pick/pack. Service: processing + booking platform fee. Subscription: processing + dunning. Lead gen: platform / CRM per-deal fees.
Marketplace sellers who forget the 15% take rate is a percentage cost: every price rise hands a cut straight back to the platform.
Contribution margin, and the truth about refunds
This is where most calculators stop too early, and where the single most expensive error in unit economics lives: treating a refund as a simple revenue haircut.
What each sale contributes, before returns.
The money a single order throws off to cover your fixed overheads and profit. This, not gross margin, is the figure that decides whether you can afford to acquire a customer.
Calling this the final margin. It isn't. It ignores the orders that come back, and for many businesses that's several points of margin still to lose.
Identical across all four. Contribution is contribution. This is the clearest proof that the math doesn't care what you sell.
A return is three different losses, not one.
When an order comes back: the revenue is gone (you refund it), the goods may return to sellable stock (recoverable, set by ρ), and the shipping and processing fees are sunk forever. Three rules. The recovery rate ρ is per-product: a melted candle is ρ=0, refurbished electronics maybe ρ=0.9.
"A 5% refund rate just costs me 5% of my margin." That treats a refund as a neutral non-sale worth $0. It isn't, it's a negative order (here, −$61): you refund the revenue, recover only some of the goods, and the shipping and processing are gone. The naive haircut shows $39.18; the truth is $36.13. On a $10M business that ~$3/order gap is $200–300K of profit that was never there.
This holds all variable selling costs as sunk on a refund. That's the conservative choice and it's right for shipping and processing. Marketplace commission is often refunded with the order (would lift the number); return shipping is often an added cost (would lower it). Model those two at the fee level if they're material to you; the direction of each is noted so you're not misled.
E-com: returns + restocking. Service: cancellations + rebooking rate. Subscription: churn (ρ rarely applies). Lead gen: deal fall-through.
A global recovery rate across all products. A business with 10% returns at 90% recovery loses ~1% of margin; the same return rate at 0% recovery loses ~10%. Averaging them hides which products are quietly bleeding.
Break-even, and the boundary you cannot cross
Once you know the real margin per order, acquisition becomes arithmetic. Your margin sets the ceiling on what you can pay for a customer. Most overspending is just not knowing where that ceiling is.
Your margin is the speed limit on ad spend.
You can spend up to your net contribution margin to acquire a customer and break even. Your required ROAS falls straight out of it. These aren't goals to beat, they're walls: cross them and every additional sale loses money faster the more you scale.
Chasing a ROAS target borrowed from a podcast ("aim for 4x"). The number isn't wrong, it's just not yours. The right ROAS is your AOV divided by your net margin. A high-margin product breaks even at 2x; a thin one needs 6x. A benchmark is a prior, useless until you put your own margin into it.
E-com: CPC → conversion → order. Service: cost per lead → booking rate → client. Subscription: CAC → signup → subscriber (then LTV over churn). Lead gen: cost per lead → qualify → close (two-stage funnel).
Scaling budget when CPA already exceeds Net_CM. More spend doesn't fix a negative unit; it multiplies the loss. Fix the margin first, then scale.
The four levers, and which one to actually pull
Once the waterfall is correct, hitting a profit target stops being guesswork. There are only four levers, and they cost wildly different amounts of effort to move.
- Conversion rate Usually the cheapest lever and the first to try. A small lift in CVR drops straight to orders with no price resistance and no cost increase. Landing pages, friction, clarity (see the conversion-journey companion).
- Price The most powerful lever per unit of effort, and the scariest. A 3% rise often beats a 25% cost cut, but it meets demand elasticity, so it's never a free 3%. Different channels feel it differently: search shoppers flinch, loyal customers barely notice.
- Costs Real but slow and finite. You can negotiate suppliers and trim fees once; you can't do it every quarter. High-COGS businesses get more from this lever than high-fee ones.
- Budget / scale The most tempting and most dangerous. Only profitable if CPA < Net_CM. Below that line it prints money; above it, it prints losses faster. Never the first lever for a unit that isn't already profitable.
The margin waterfall, made interactive
Everything above, running live, on the same engine as The Universal Margin. Switch between e-commerce, services, subscriptions, lead gen and digital products. The maths never changes, only the words. Watch the dollar fall, size up your acquisition break-even, and let the goal solver work backwards from a target margin or monthly profit to the smallest change on each lever. Free, and embeddable on your own site.
Physical products, shipping, returns.
- Cost to acquire (CPA)
- –
- Max bid per click
- –
- New orders / mo
- –
- Contribution / mo
- –
- Profit after ad spend / mo
- –
- Avg lifetime
- –
- Lifetime value (LTV)
- –
- LTV : CAC
- –
- Payback period
- –
Set a target and Solve. The calculator works backwards to the smallest change on each lever, then one tap applies it.
- Net revenue (AOV, ex-tax)
- –
- Cost of goods
- –
- Variable selling cost
- –
- – of which percentage fees
- –
- – of which fixed fees
- –
- Contribution margin (pre-refund)
- –
- Refund leak
- –
- True contribution margin
- –
What most calculators get wrong
- Refunds taken off revenue, not margin. A refund doesn’t just hand back the sale price; it removes the contribution you would have kept, and you usually eat the selling costs anyway. Applying the refund rate to revenue understates the real damage.
- Returned goods treated as fully recoverable. A returned unit is rarely 100% resellable. Counting full recovery quietly flatters cost of goods on every refunded order.
- Stopping at gross margin. Gross margin ignores percentage and fixed selling fees and the refund leak. The number that decides what you can spend to acquire a customer is the contribution you actually keep.
This one is cash-flow-correct instead. Cash-flow-correct margin: a refund loses the sale and the selling costs, and returns goods only to the resellable extent. The Universal Margin (research), Christopher White
Two lines of code. It stays in sync, costs nothing, asks nothing.
Paste the URL of the page you'll embed the tool on. We'll read its colours and fonts and show you what the calculator would look like in your brand.
Extracted colours
Click an area below, then tap colours above to paint it. The area stays selected, so try a few quickly. The green (money kept) and red (leaks) are generated automatically so the chart always reads right.
Fine-tune surfaces (cards, fields, buttons, active tab)
Pick a colour from the strip above
How the collaboration badge will appear:
× <div data-cwc-tool="margin-calculator" data-cwc-title="Margin & Break-even Calculator by Christopher White Consulting" data-cwc-reserve-narrow="2200" data-cwc-reserve-wide="1320" style="min-height:1320px"></div>
<script src="https://christopherwhite.com.au/embed/loader.js" data-cwc-tool="margin-calculator" async></script>Five ways the number lies to you
Each of these inflates your apparent margin, which is the dangerous direction: it makes you spend money you don't have on customers you can't afford. All five are common in tools people trust.
Refund rate × revenue
Treats a return as a revenue haircut. Ignores recovered goods and sunk fees. Overstates margin by 2-3 points on returns-heavy lines.
Stopping at gross margin
Revenue minus COGS feels like the answer. It ignores every selling cost and every refund. It is roughly twice the real number.
One blended "fees" %
Collapsing percentage and fixed fees makes every price-change scenario wrong, because only half the fees actually move with price.
Borrowed ROAS targets
"Aim for 4x" is someone else's margin structure. Not wrong, just not yours. Your break-even ROAS is your AOV over your net margin. Nobody else's number applies.
Floating-point money
Rounding errors invisible per order, material at volume. Calculate money in exact decimals.
Blended P&L, no unit view
A healthy monthly total can hide products and channels losing money on every sale. The unit is where truth lives.
Trust the math, verify the inputs
- Find your worst unit Calculate net CM per product and per channel, not blended. The average always hides a loser.
- Stress the refund line Set recovery ρ honestly, per product. The candle and the laptop are not the same business.
- Separate your fees Tag every cost as percentage or fixed before modelling any price change, or the model will lie.
- Find your boundary first Know your break-even CPA and ROAS before you set a spend target, not after the quarter.
- Re-run when inputs move Supplier prices, processor rates, and shipping all drift. A margin model is maintained, not finished.
Method & basis
Most of this is not novel, and shouldn't be. The individual stages are standard managerial accounting: contribution margin, variable vs fixed cost behaviour, break-even, and LTV as margin over churn are textbook. What's assembled here is (a) running them in the correct cash-flow order, (b) the refund model, which most calculators get wrong, and (c) the claim that one engine serves every business model. Those three are the parts to scrutinise.
- The waterfall stages: tax, AOV, COGS, variable selling costs, contribution margin, break-even. Standard managerial accounting; the contribution here is sequence and cash-flow correctness, not the formulas themselves.
- The cash-flow refund model: a return loses revenue, may recover goods (per-product rate ρ), and sinks variable fees. COGS_chargeable = 1 − (R × ρ). A refunded order is a negative order, not a $0 non-sale. This is the part most tools get structurally wrong, and the part worth checking hardest.
- Unit operating leverage: percentage-based costs scale with price; fixed costs don't. Required for any correct price-change scenario. Textbook, routinely collapsed in practice.
- Break-even as boundary: max CPA = net contribution margin; required ROAS = AOV ÷ net CM, on a stated revenue basis. Channel-specific, never a universal target.
- Universality: a service priced with zero shipping and zero pick/pack yields identical unit economics to a physical product; the vocabulary changes, the math doesn't. This is a sufficiency claim, demonstrated across four models (e-com, service, subscription, lead gen), not a proof that a fifth (e.g. two-sided marketplace economics, where both sides take a fee) couldn't strain it. Treat it as a strong working hypothesis, not a theorem.
- Precision: money calculated in exact decimals, never floating point, to prevent compounding rounding error at volume.
- Provenance: the full derivation, the six failed versions, and the test methodology are documented in C. White, "The Universal Margin" (2026). These formulas run in production in ProfitOS, verified against a Python oracle and 10,000 randomised scenarios per invariant. This paper is the reasoning, not a sales page; if you want the engine that does it for you, that's what ProfitOS is, and it's reasonable to be skeptical of a method whose author also sells the tool. The defence is that the math is checkable, so check it.
White, C. (2026). The Money You Actually Keep. Christopher White Consulting. https://christopherwhite.com.au/playbooks/how-to-unfuck-your-margins/