> ## Content Index > Fetch the complete content index at: https://abundance.alloconomy.com/llms.txt > Use this file to discover other available public pages before exploring further. # Beyond Accounting: Who Actually Profits from My Mac mini? - URL: https://abundance.alloconomy.com/alloconomy/beyond-accounting-who-profits-from-a-mac-mini/ - Published: 2026-09-11T17:47:33.000Z - Updated: 2026-09-12T09:26:51.000Z - Description: The entries balance, but who makes money? A reconciled model of one computer purchase explains the different profit margins, financing costs, risks and capital demands behind 0% APR. - Author: Sathya Narayanan - Tags: Alloconomy, Accounting, One Purchase Several Businesses, #series-detail Accounting · Two-part series One purchase, several businesses 1. Part 1 [One Mac mini, Five Sets of Books](https://abundance.alloconomy.com/alloconomy/one-mac-mini-five-sets-of-books/) Follow an order through authorization, shipment, settlement, delivery and repayment. 2. Part 2 **Beyond Accounting: Who Actually Profits from My Mac mini?** Current essay Compare contribution, operating profit, net profit and capital with an explicit unit economics model. *Part 2 of 2 · One purchase, several businesses* ## The entries balance. Does the business make money? My Mac mini costs **$1,299**. I pay **$108.25 a month for twelve months at 0% APR**. We use the same assumed completed order as Part 1: Apple ships on October 8, card settlement occurs October 9, and the computer arrives October 12, within the October 8–15 window. I voluntarily pay $108.25 on each billing month-end from October 31, 2026 through September 30, 2027\. Later, we compare the alternative in which I return it on October 22 and receive the hardware credit on October 23. In [Part 1](https://abundance.alloconomy.com/alloconomy/one-mac-mini-five-sets-of-books/), we followed the accounting. Every debit had a credit. The sale, the loan, the payment service, and the shipment each belonged to a different set of books. Now comes the harder question: **which participant actually benefits from doing this business?** The biggest cash receipt need not produce the highest margin. The highest margin need not produce the biggest profit. A loan repaid in full can still lose money. A profitable sale can still destroy value if earning it requires too much capital or displaces a better sale. Our five principal roles remain Apple, Goldman Sachs as card issuer, Mastercard, the unidentified merchant acquirer/processor, and the unidentified carrier. **This analysis covers only the order, payment, financing, and shipment. It does not trace chipmakers, manufacturers, component suppliers, or any other upstream company.** We take Apple's opening product cost as an input. We also show the separately priced annual AppleCare plan as an extension within the Apple group, without pretending it is part of the $1,299 hardware financing. Five roles is our economic map, not an independently verified count of legal entities. UPS and Fiserv appear below only as public benchmarks for the kinds of businesses involved; neither is identified as a counterparty to my purchase. ## First, define what a unit means To Apple, the unit is one computer order. To the issuer, it is a loan that can remain outstanding for a year. To Mastercard and the processor, it is a set of payment services. To the carrier, it is one shipment. We can assign those activities to one purchase and follow their economics until the final installment is paid. Below, a **cohort** means a group of comparable purchases used to estimate risks such as defaults and warranty claims. Each participant still has its own revenue base and service period. | Measure | What it answers | How we will calculate it | | ----------------------------- | --------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------ | | Unit revenue or income | What does this participant earn from serving this purchase? | The participant's own consideration; exclude loan principal and money passed through as agent | | Product gross profit | What remains after the assigned product cost? | Useful for Apple; not automatically comparable with a bank or network | | Contribution | What remains after costs that vary with serving this purchase? | Revenue/income less explicitly defined variable costs | | Operating profit contribution | What remains after the unit carries a share of fixed operating costs? | Contribution less assigned fixed costs and depreciation | | Net profit contribution | What remains after financing below operating profit and tax? | Operating profit less allocated non-operating financing cost and tax; bank funding cost is already above this line | | Economic profit | Did the activity also cover the opportunity cost of capital? | After-tax operating return less a capital charge, on a consistent capital basis | Contribution margin is a management measure, not a universal reported line. A salaried driver may be fixed for tomorrow's extra parcel and variable when planning another delivery route. Manufacturing cost in inventory can contain allocated fixed factory costs. “Variable” requires a decision horizon. Here the horizon runs from the order through its twelve scheduled payments; the dated loan-funding period is 356 days. We assume capacity costs are separately identifiable. The per-order fixed allocations below are planning conventions, not journal entries posted each time I click Buy. One distinction matters before the numbers: **a fully repaid individual loan has no realized principal loss; a lender deciding whether to offer thousands of such loans must still budget for defaults across that group.** Our main profit table is that risk-adjusted decision estimate. We also show the result when this individual buyer repays in full, as assumed in Part 1. ## Use public reporting to anchor the assumptions Company reports are useful anchors. They are not itemized profit statements for my serial-numbered computer. | Participant or benchmark | Public observation | Proper use in this example | | -------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------ | | Apple | Products gross margin was 36.8% in FY2025\. In the June 2026 quarter it was 40.1%; Apple cited product mix and tariff refunds among the drivers. | An order-of-magnitude check, not a disclosed Mac mini margin. Services and specific Mac configurations have different economics. | | Goldman issuer | Goldman continues servicing Apple Card during the announced transition. Its Apple Card portfolio is held for sale, and reporting includes portfolio markdowns. | A reason to model credit, funding and valuation separately; corporate investment-banking profitability does not reveal ACMI profitability. | | Mastercard | FY2025 net revenue was $32.791 billion, operating margin 57.6%, and net income $14.968 billion. Net income/revenue was therefore about 45.6%. | Evidence of the wider network business's profitability, not proof of the fee or margin on this order. | | Acquiring/processing benchmark: Fiserv | Merchant Solutions reported a 30.0% adjusted operating margin in Q2 2026. | A benchmark with an adjusted, segment-specific definition; not Apple's identified processor or its negotiated economics. | | Delivery benchmark: UPS | Its Q2 2026 filing reports a 4.1% consolidated GAAP operating margin, with material operating and restructuring context. | A warning about fixed costs and period effects; not the margin on this parcel or evidence UPS will carry it. | Sources: [Apple FY2025](https://www.sec.gov/Archives/edgar/data/320193/000032019325000079/aapl-20250927.htm?ref=abundance.alloconomy.com), [Apple June 2026 quarter](https://www.sec.gov/Archives/edgar/data/320193/000032019326000020/aapl-20260627.htm?ref=abundance.alloconomy.com), [Apple Card transition](https://www.apple.com/newsroom/2026/01/chase-to-become-new-issuer-of-apple-card/?ref=abundance.alloconomy.com), [Goldman's loan disclosure](https://www.sec.gov/Archives/edgar/data/886982/000088698226000091/R17.htm?ref=abundance.alloconomy.com), [Mastercard FY2025](https://www.sec.gov/Archives/edgar/data/1141391/000114139126000013/ma-20251231.htm?ref=abundance.alloconomy.com), [Fiserv Q2 2026](https://investors.fiserv.com/news-releases/news-release-details/fiserv-reports-second-quarter-2026-results?ref=abundance.alloconomy.com), and [UPS Q2 2026](https://www.sec.gov/Archives/edgar/data/1090727/000162828026053249/ups-20260630.htm?ref=abundance.alloconomy.com). The reports give us context for choosing plausible teaching assumptions. They do not disclose the fee or cost of an individual Mac mini order. We will therefore state each assumed contract term once, calculate its consequences, and test what changes the answer. ## The assumptions behind our completed-order case All figures are U.S. dollars. The $1,299 price, twelve $108.25 installments, and 0% rate come from the order details shown in Part 1\. Fulfillment dates, payment dates, and the following costs are **explicit assumptions for this example**, not reported transaction results. | Input | Base assumption | Why it is included | | --------------------------------------------------------- | ---------------------------------- | ------------------------------------------------------------------------------------------------------------ | | Apple's opening inventory cost | $800.00 | Illustrative product cost; includes $40 of allocated fixed production cost and $760 of variable product cost | | Merchant acceptance fee | $32.48 | The same fee pool as Part 1 | | Issuer interchange / network fee / processor retained fee | $23.38 / $2.60 / $6.50 | These sum to $32.48; the model does not count them again as independent merchant charges | | Freight paid by Apple | $15.00 | The carrier's revenue | | Expected basic warranty cost | $8.00 | Distinct from the separately priced AppleCare service | | Merchant financing support paid by Apple to issuer, S | $65.00 | A scenario variable; no claim that this payment exists in the actual contract | | Product-linked Daily Cash | $38.97 | 3% of $1,299; issuer bears it in this scenario | | Issuer annual funding rate | 5.0% | Applied daily to the dated outstanding principal, from October 9 settlement to the last repayment | | Expected lifetime credit loss | 1.0% of initial principal = $12.99 | Expected loss across comparable loans; zero realized loss for the fully repaid individual case | | Issuer servicing / allocated fixed operating cost | $3.00 / $3.00 | Costs beyond funding, rewards, and credit | | Apple's allocated R&D, selling and administration | $110.00 | Separate from the $40 of fixed production cost already inside the $800 inventory cost | | Illustrative tax rate | 25% on positive pretax profit | A comparison convention; no assumed tax benefit on a model loss | We assume the $65 is the entire separate Apple-to-issuer financing contribution in this case. We do not add a second Apple reimbursement of the same $38.97 reward. If the actual contract funds rewards differently, move the cost to the correct participant and reconcile the offset. Assume Apple pays the $65 financing support on October 9\. We evaluate it over the loan's economic life; that does not make it all accounting income on day one. Sales tax, annual AppleCare, and other card balances stay outside the core profit table. Funding is allocated to gross hardware-loan principal; the servicing budgets absorb small reward and fee cash-timing effects rather than model a second set of daily funding balances. ## Apple: the financing offer consumes part of the hardware profit Start with the visible sale price and the assumed inventory cost: **$1,299 − $800 = $499 of product gross profit, or 38.4%.** That is a deliberately narrow, pre-order-cost product margin. It is not a reproduction of Apple's reported gross margin: reported cost of sales and our management cost buckets may classify freight, warranty, and other items differently. For contribution, the $40 of fixed production cost comes out of variable product cost and is added back as a fixed allocation afterward: | Apple: one hardware order | Amount | | --------------------------------------------------------------------------- | ------------------- | | Revenue, ignoring the bundled-service timing allocation discussed in Part 1 | $1,299.00 | | Variable product cost | ($760.00) | | Payment acceptance | ($32.48) | | Delivery | ($15.00) | | Expected basic warranty service | ($8.00) | | Assumed financing support | ($65.00) | | **Contribution** | **$418.52 / 32.2%** | | Fixed production allocation | ($40.00) | | Assigned R&D, selling and administration | ($110.00) | | **Operating profit contribution** | **$268.52 / 20.7%** | | Allocated non-operating financing cost | ($2.00) | | Illustrative income tax | ($66.63) | | **Net profit contribution** | **$199.89 / 15.4%** | We have not deducted $800 and then deducted its $40 fixed component a second time. The route through contribution and the route through product gross profit reconcile to the same $268.52 operating result. My base-case judgment is that Apple can plausibly earn the largest absolute dollar contribution among these five roles, because its consideration is the price of the machine. The exact $199.89 is a scenario output, not a forecast with that precision. A less profitable configuration, a promotion, different tariffs, or a more expensive financing contract could change it substantially. The commercially interesting question is whether the financing offer caused the sale. If I would have bought the identical Mac for cash anyway, the financing subsidy mostly reduces Apple's profit on that sale. If the installment offer brings in a profitable order that otherwise would not exist, paying for it can make sense. With our assumptions, contribution before the hypothetical $65 support is $483.52\. Compare 100 otherwise identical card-funded orders without that extra support against supported orders earning $418.52 each. Apple needs about **115.5 supported orders** to match the same contribution dollars: $48,352 ÷ $418.52\. That is roughly a **15.5% volume uplift**, before any change in capacity costs, returns or customer mix. This comparison holds the other payment economics constant; a true cash-payment comparison would change those too. “Installments increase conversion” is therefore incomplete. The useful measurement is incremental contribution after financing costs and after allowing for customers who merely switched payment methods. ## Goldman: the price of waiting can exceed the fee for lending The $1,299 principal is not revenue. We model the issuer's gross income from interchange and hypothetical merchant support as: **$23.38 + $65 = $88.38.** This is an analytical income measure before rewards and funding costs, not a claim about Goldman's reported net-revenue presentation. Reward costs can be presented net of related income. Changing presentation changes a percentage denominator without changing economic dollars. The bank does not fund $1,299 for a whole year: repayments progressively return the principal. Use the actual assumed dates rather than twelve identical monthly periods. It funds $1,299 for the 22 days from October 9 to October 31; then $1,190.75 for the 30 days to November 30; then $1,082.50 for the 31 days to December 31; and so on until September 30, 2027\. Multiply each outstanding balance by the days it remains outstanding and add the results: **Total balance-days = 245,078 dollar-days.** At our assumed 5% annual rate and an actual/365 convention, **funding cost = $245,078 × 5% ÷ 365 = $33.57**. The time-weighted average principal over the 356-day funding period is about **$688.42**. A later first payment or a delayed final payment would increase that cost. | Issuer: one twelve-month loan cohort | Amount | | --------------------------------------------- | ------------------ | | Assumed interchange | $23.38 | | Hypothetical merchant financing support | $65.00 | | **Gross model income** | **$88.38** | | Funding the declining balance | ($33.57) | | Expected lifetime credit loss | ($12.99) | | Product-linked Daily Cash borne by issuer | ($38.97) | | Variable servicing | ($3.00) | | **Contribution after funding and risk costs** | **−$0.15 / −0.2%** | | Assigned fixed operating cost | ($3.00) | | **Pretax operating contribution** | **−$3.15 / −3.6%** | | Assumed current tax benefit | $0.00 | | **Net result** | **−$3.15** | This is the surprising part: **even $65 of merchant support does not quite cover the costs in this particular scenario.** The exact hypothetical break-even support is $68.15\. That is not evidence Apple actually pays $68.15, or pays anything at all. It identifies what this combination of assumptions requires. For the **individual completed order in Part 1**, set realized principal loss to zero. Its result improves by $12.99, from a $3.15 expected loss to **$9.84 pretax profit**; at our illustrative 25% tax rate that is **$7.38 net**. That difference is the release of a risk budget in the economic analysis, not a claim about the line on which Goldman reverses an actual accounting valuation. Repayment still leaves funding, rewards, and servicing to pay for. Nor should 0% be justified by casually assuming the bank will charge this same installment balance ordinary revolving interest. The product is explicitly interest-free under its terms. Economics on other card purchases belong to a broader customer analysis and are not smuggled into this order. Goldman's portfolio-sale markdowns also cannot simply be allocated to this one order using the $12.99 expected-loss assumption. A sale valuation can reflect funding, contractual economics and market pricing as well as credit. The [July 2026 earnings filing](https://www.goldmansachs.com/investor-relations/financials/8k/2026/8k-07-14-26.pdf?ref=abundance.alloconomy.com) concerns portfolio reporting; the table above is a lifetime economic model. ### How easily does the bank's answer flip? | Merchant financing support S | Income including $23.38 interchange | Pretax result, other base assumptions unchanged | | ---------------------------- | ----------------------------------- | ----------------------------------------------- | | $0 | $23.38 | −$68.15 | | $40 | $63.38 | −$28.15 | | $65 | $88.38 | −$3.15 | | $100 | $123.38 | $31.85 | A one-percentage-point increase in the annual funding rate costs about **$6.71** on this repayment schedule. A one-percentage-point increase in lifetime principal losses costs **$12.99**. Moving the $38.97 reward cost to another party helps the issuer by exactly that amount but burdens someone else; it does not make the cost disappear. This is why my judgment on the issuer is a range of possible outcomes, not a confident claim of positive unit profit. The missing commercial terms can dominate the answer. ## Mastercard: a small fee can carry a high percentage margin Mastercard earns the assumed $2.60 network fee, not the $1,299 principal and not the $23.38 interchange. A powerful global network can spread much of its infrastructure and development cost across many transactions, while a particular message has a small incremental cost. For our purchase, suppose $0.25 is variable network cost and $0.85 is an assigned share of fixed operations, technology, depreciation and administration: | Network: this payment service | Amount | | -------------------------------------- | ----------------- | | Assumed net fee revenue | $2.60 | | Variable service cost | ($0.25) | | **Contribution** | **$2.35 / 90.4%** | | Fixed operating allocation | ($0.85) | | **Operating profit contribution** | **$1.50 / 57.7%** | | Allocated non-operating financing cost | ($0.02) | | Illustrative tax | ($0.37) | | **Net profit contribution** | **$1.11 / 42.7%** | The operating margin is deliberately in the vicinity of Mastercard's reported company-level margin, to create a plausible teaching scenario. That is a calibration assumption, not independent evidence that this transaction earns 57.7%. Actual fee schedules, customer incentives, service mix and cost allocation could make the result different. A 42.7% margin on $2.60 produces $1.11\. A much lower margin on Apple's $1,299 produces far more dollars. If we instead apply Mastercard's corporate margin to the entire computer price, we are assigning it profit on revenue it never earned. For the network, the important business questions include net fees after incentives, transaction growth, resilience, regulation and the cost of maintaining acceptance. None requires it to bear the buyer's twelve-month consumer loan. ## The merchant processor: keep onward payments out of the denominator Our acquirer retains **$6.50** after the modeled interchange and network charges. Assume $1.60 of variable processing, support, and expected incident cost, plus $2.90 of fixed operating allocation. | Acquirer/processor: this merchant service | Amount | | -------------------------------------------- | ----------------- | | Retained fee revenue in our net presentation | $6.50 | | Variable cost | ($1.60) | | **Contribution** | **$4.90 / 75.4%** | | Fixed operating allocation | ($2.90) | | **Operating profit contribution** | **$2.00 / 30.8%** | | Allocated non-operating funding cost | ($0.25) | | Illustrative tax | ($0.44) | | **Net profit contribution** | **$1.31 / 20.2%** | That 30.8% scenario margin is not a disclosure about Apple's processor. Fiserv's 30.0% adjusted Merchant Solutions margin is only a contextual benchmark and uses its own segment and adjustment definitions. The provider serving a large enterprise merchant can have very different pricing from a small-business terminal or e-commerce bundle. The accounting presentation also matters. Reporting pass-through fees gross would enlarge both revenue and cost and lower the apparent margin, even with the same retained dollars. Before comparing two payments companies, reconcile their definitions of revenue. Operationally, a processor can improve this order's economics by reducing failed authorizations, support work, disputes, or settlement delay. It can also be forced to accept a thin fee to win a large merchant. High throughput alone proves neither pricing power nor attractive profit. ## The carrier: route density can matter more than the computer's price The carrier's assumed revenue is **$15**, independent of the computer's $1,299 retail price. Assume $10.50 of variable transport cost, $3 of allocated fixed operating cost and depreciation, and $0.30 of non-operating financing cost. | Carrier: one shipment | Amount | | -------------------------------------- | ----------------- | | Assumed shipment revenue | $15.00 | | Variable transport cost | ($10.50) | | **Contribution** | **$4.50 / 30.0%** | | Fixed operations and depreciation | ($3.00) | | **Operating profit contribution** | **$1.50 / 10.0%** | | Allocated non-operating financing cost | ($0.30) | | Illustrative tax | ($0.30) | | **Net profit contribution** | **$0.90 / 6.0%** | The assumed 10% unit operating margin is a normalized scenario, not UPS's reported 4.1% quarterly GAAP result. We do not transplant a period containing business-wide costs and restructuring into an individual delivery; nor do we erase those costs when assessing the company as a whole. This is the participant most visibly exposed to physical capacity. A nearby stop on an already busy route can be attractive. A failed delivery, a second attempt, a long detour or a costly expedited service can consume the entire margin. Add $3 of variable cost and our $1.50 operating profit becomes a $1.50 loss. A more expensive computer does not automatically pay the carrier more. Package dimensions, weight, service commitment, route and the merchant's contract can matter much more. In our completed-order case the October 12 delivery earns $15; Apple's October 31 payment collects that already-earned revenue. ## Put the five results beside one another These are **illustrative risk-adjusted outcomes over the completed-order schedule**, with each percentage using the participant's own stated revenue/income denominator. The issuer row includes the $12.99 expected-loss budget; the fully repaid individual-loan result is $9.84 pretax and $7.38 net instead. The other rows retain their stated cost estimates and allocations. | Participant | Revenue/income base | Contribution | Operating or bank pretax contribution | Net contribution | | ------------------ | ------------------------- | --------------- | ------------------------------------- | ---------------- | | Apple hardware | $1,299.00 | $418.52 (32.2%) | $268.52 (20.7%) | $199.89 (15.4%) | | Goldman issuer | $88.38 gross model income | −$0.15 (−0.2%) | −$3.15 (−3.6%) | −$3.15 (−3.6%) | | Acquirer/processor | $6.50 retained fees | $4.90 (75.4%) | $2.00 (30.8%) | $1.31 (20.2%) | | Mastercard | $2.60 net network fees | $2.35 (90.4%) | $1.50 (57.7%) | $1.11 (42.7%) | | Carrier | $15.00 | $4.50 (30.0%) | $1.50 (10.0%) | $0.90 (6.0%) | The useful interpretation is a profile: Apple plausibly earns the largest dollar profit; the network can have the highest percentage margin; the issuer's result is acutely sensitive to contract economics and credit; the processor monetizes a small retained service fee; and the carrier's modest unit surplus is sensitive to operational execution. The exact numbers are less robust than that structure. Even the structure is conditional: a deeply discounted product or a materially different fee arrangement can change the ranking. ### A reconciliation that catches hidden double counting Apple's $32.48 payment cost becomes $23.38 issuer income, $2.60 network income and $6.50 processor income. Apple's $15 freight cost becomes carrier revenue. The assumed $65 support is an Apple expense and issuer income. These are transfers within our five-role boundary. Add the modeled operating contributions and those internal transfers cancel. The resulting **$270.37** can also be obtained directly: **$1,299 − $800 product cost − $8 warranty − $110 Apple operating allocation − $33.57 issuer funding − $12.99 credit loss − $38.97 rewards − $6 issuer servicing/overhead − $4.50 processor operating costs − $1.10 network operating costs − $13.50 carrier operating costs = $270.37.** This is a reconciliation of our model boundary, not consolidated financial statements for the economy. The bank funding cost remains an operating input here, while the small non-operating financing allocations for other participants have not yet been deducted. Upstream supplier books remain outside the analysis; their effect is contained in the opening product cost. If that reconciliation fails, someone has usually counted principal as revenue, counted a fee twice, or made a contractual transfer disappear from one side. ## Rewind to the return: which profits survive? Use the same alternative as Part 1: the carrier returns the computer on October 22 and the hardware credit settles October 23, before any installment is paid. The return changes the economics of an order already shipped; it is not the same as never taking the order. Apple refunds the $1,299 price, recovers a computer worth $760 against an original $800 inventory cost, and receives back the $23.38 interchange. The processor retains $6.50 and Mastercard retains $2.60\. The carrier keeps $15 for the outward journey and earns another $12 for the return. The buyer repays the $38.97 Daily Cash adjustment on October 31\. **For the separate financing-support extension, assume the entire $65 is refundable and the issuer pays it back to Apple on October 23.** That transfer therefore leaves zero net support income or expense in this returned-order scenario. We need a few additional cost assumptions to complete the comparison: $5 of Apple return inspection/handling; $0.40 of extra processor work; $0.05 of extra network work; and $9 variable plus $2 allocated fixed carrier cost for the return leg. The issuer incurs $1 of origination/refund servicing rather than $3 of full-term servicing, and retains its $3 fixed account-cost allocation. Keep Apple's $110 shared operating allocation and the original processor/network fixed allocations. These are selected costs for this scenario, not refund-policy facts. The bank's hardware principal is funded for just **14 days**, October 9–23: **$1,299 × 5% × 14/365 = $2.49**. No installment principal remains after the refund. We assume no default in this resolved return, recover the reward in full, and carry no lifetime-loss budget for an extinguished loan. Actual valuation balances are cleared as Part 1 explains. | Participant | Revenue or income left after reversals | Costs before shared fixed allocations | Result before shared fixed allocations | Operating/pretax result after those allocations | | ---------------------- | ------------------------------------------ | ----------------------------------------------------------------------------- | -------------------------------------- | ----------------------------------------------- | | Apple | $0 | $40 inventory loss + $9.10 retained fees + $27 freight + $5 handling = $81.10 | −$81.10 | **−$191.10**, after $110 shared operating cost | | Goldman issuer | $0 interchange/support; $0 net reward cost | $2.49 funding + $1 servicing = $3.49 | −$3.49 | **−$6.49**, after $3 fixed cost | | Acquirer/processor | $6.50 | $1.60 original variable work + $0.40 refund work = $2.00 | $4.50 | **$1.60**, after $2.90 fixed cost | | Mastercard | $2.60 | $0.25 original variable work + $0.05 refund work = $0.30 | $2.30 | **$1.45**, after $0.85 fixed cost | | Carrier, both journeys | $27.00 | $10.50 outbound + $9 return variable cost = $19.50 | $7.50 | **$2.50**, after $5 combined fixed cost | Apple's **$760 returned inventory remains an asset**, not a fresh expense. We do not charge the original $800 cost again or add its old $40 fixed production allocation again after reversing cost of sales. The $40 reduction in recoverable inventory value already captures the loss. We also assume the unused $8 product-warranty provision is released and the $65 financing support is recovered in full. Inspection and freight, however, have happened and cannot be undone. For net results, retain the original allocated financing costs of $2 for Apple, $0.25 for the processor and $0.02 for Mastercard. Give the carrier $0.50 for both journeys and apply 25% tax to positive pretax results, with no tax benefit on losses. The resulting net contributions are **Apple −$193.10; issuer −$6.49; processor $1.01; Mastercard $1.07; carrier $1.50**. Apple and the issuer have zero net revenue/income in this branch, so a profit-margin percentage using that denominator is **not meaningful**. Dollar losses are the useful measure. The five operating/pretax results sum to **−$192.04**. Independently, the surviving costs outside transfers among these five roles are: $40 inventory loss, $5 Apple handling, $110 Apple shared overhead, $2.49 issuer funding, $4 issuer servicing/fixed cost, $4.90 processor operating cost, $1.15 network operating cost, and $24.50 carrier operating cost. They also total **$192.04**. The retained fees and freight cancel between the participant paying and the participant earning them; the refunded $65 support cancels too. This is the practical difference between revenue reversal and economic reversal. The buyer can be made whole on the hardware price while businesses still consume time, cash, transport capacity, and inventory value. A return might improve one participant's revenue—the carrier earns a second freight fee—while making the combined outcome worse. Our separately modeled sales tax is refunded and recovered from the tax authority, so it creates no revenue or cost here. Part 1 also assumes cancellation and full refund of the annual AppleCare price before any claim. That plan therefore has zero net service revenue in the return branch; any administration and payment costs already consumed still need to be charged to the plan, separately from this hardware table. ## The annual AppleCare plan is a different economic unit The order details separately show **$39.99 per year**. We assume coverage from October 8, 2026 through October 7, 2027, as in Part 1\. At Apple-group level that is a service contract with a coverage period, cancellation risk and uncertain claims, not $39.99 of immediate profit. Its legal provider can be an Apple subsidiary, as discussed in Part 1; intercompany collections must not become duplicate group revenue. For a full earned year, assume $12 of expected covered-service cost, $3 of variable administration and $1.20 of payment/collection costs. That leaves **$23.79 of contribution, or 59.5%**. After an assumed $7 fixed operating allocation, operating contribution is **$16.79, or 42.0%**. With no separately allocated interest and the same 25% illustrative tax rate, net contribution is **$12.59, or 31.5%**. Every cost in that paragraph is hypothetical. The payment amount is a separate all-in service-order assumption, not an extension of the hardware fee split. We do not apply Apple's overall Services margin to AppleCare and call the result an estimate. Services contains businesses with very different delivery and risk profiles. If expected claims cost is $25 instead of $12, the same service plan's operating contribution falls from $16.79 to $3.79\. One covered repair can exceed a year's premium; profitability is evaluated across a pool and time, not by declaring a single claim-free month pure profit. That is why collecting the annual price and recognizing the year's service revenue are different events. The cash arrives before the uncertainty is fully resolved. ## A reasonable range is more useful than a spurious exact answer The ranges below change **one input at a time** around the base case. They are stress tests, not probabilities, market quotations, or combined best/worst cases. | Participant | Illustrative change | Operating/pretax result range | | ------------------ | ------------------------------------------------------------------------ | ----------------------------- | | Apple | Opening inventory cost $700–$900, with its fixed $40 component unchanged | $368.52 to $168.52 | | Issuer | Merchant financing support $0–$100 | −$68.15 to $31.85 | | Mastercard | Net network fee $1.30–$3.90, holding modeled costs fixed | $0.20 to $2.80 | | Acquirer/processor | Retained fee $3–$10, holding modeled costs fixed | −$1.50 to $5.50 | | Carrier | Variable shipment cost $9–$13.50 | $3.00 to −$1.50 | The network and processor rows are isolated participant sensitivities. If their retained fees change in a complete system scenario, update the merchant fee or another participant's share as well; the fixed $32.48 pool cannot support a larger share for everyone simultaneously. For a real commercial decision, replace the assumed numbers with the acquiring statement, issuer and reward terms, product cost and warranty experience, freight invoice, and relevant cost allocations. The model then becomes a way to test the decision rather than merely explain it. ## Profit margin still does not tell us whether capital is well used Apple's inventory can be on its books before I order. The bank's financial asset can remain after Apple has its cash. The processor may bridge a settlement gap. The carrier and network rely on shared infrastructure built before this purchase. Those differences make the timing of capital essential. For illustration, assign equity equal to 10% of the loan's $688.42 time-weighted average principal: **$68.84**. At a 12% annual required return over 356 days, that capital carries an **$8.06 hurdle**: $68.84 × 12% × 356/365\. This is an investor's required return, not an interest bill or an additional operating expense, and not a statement of Goldman's regulatory requirement. The main risk-adjusted case already loses $3.15 before that hurdle. In the $100-support scenario, pretax profit is $31.85 and modeled after-tax profit is $23.89\. After the $8.06 capital charge, **$15.83** remains. The fully repaid $65-support case earns $7.38 after tax, which is positive but still falls short of this illustrative hurdle. Risk, price, and capital can change the conclusion even when the customer pays every installment. Similarly, a carrier's positive contribution may justify one extra package on an existing route but not justify buying a new fleet. A network's attractive transaction margin does not tell us the return on a new market-entry investment. Unit economics is an input to capital allocation; it does not replace it. ## What this changes for the person making the decision **For the buyer:** compare the full purchase obligation and total charges, not just the monthly payment. The product's $38.97 Daily Cash and the timing benefit of 0% can be useful, but neither makes an unnecessary $1,299 purchase inexpensive. Taxes, separate service plans, and other card balances follow their own terms. [Apple's installment terms](https://support.apple.com/en-us/102730?ref=abundance.alloconomy.com) **For a merchant's employee or leader:** evaluate incremental contribution after payment, financing, freight, warranty and returns. Measure how many financed orders are genuinely additional. Test whether improved conversion pays for the offer across all eligible orders, including purchases that would have happened anyway. **For the businesses earning fees:** protect the economics that belong to your role. A lender should price funding, losses, rewards and capital together. A processor should measure retained fees after onward payments and service cost. A carrier should measure route and service complexity. A network should separate volume growth from net pricing and incentives. **For an investor:** identify the revenue denominator, accounting presentation, capital exposure and contract dependence before comparing margins. A large operating margin can coexist with a small dollar opportunity. A thin unit spread can disappear with a modest change in losses or funding. A company-wide margin may conceal very different products and customers. The computer arrives in one box. The economics do not. One company sells the product, another finances the wait, two enable the payment, and another performs the delivery. Each earns a different price for a different service, carries a different risk, and operates on a different clock. Accounting tells us what each has recorded. Economics asks whether the work, time, uncertainty and capital were adequately rewarded. **Previous:** [One Mac mini, Five Sets of Books](https://abundance.alloconomy.com/alloconomy/one-mac-mini-five-sets-of-books/) *Company-reporting references retain their stated dates. Fulfillment, repayment, and return scenarios revised September 12, 2026\. All negotiated fees, costs, support payments, and allocations are illustrative assumptions. Rounded cents make the model reconcile; they do not imply measurement precision.*