The one-line difference
In dropshipping you are the seller of record: the money hits your account, the customer emails you when the package is late, and the refund comes out of your balance. In affiliate you are the referrer — someone else takes the payment, ships the box, and eats the return, and you get a percentage that can be reversed weeks later. Everything else in this comparison follows from that single split. Dropshipping buys you margin control and hands you the operational load. Affiliate hands you zero operational load and somebody else’s commission table, which they can rewrite without asking you: Amazon cut rates across several categories in April 2020 with about a week of notice, and publishers who had built their whole site around one of those categories woke up to a different business.
Neither one is passive income, whatever the thumbnail promised. One version has you at 11pm explaining to a stranger where a parcel you have never touched currently is. The other has you watching a dashboard where last month’s earnings quietly get smaller while you sleep. Choose the failure mode you can live beside for a year, because that is the actual decision in front of you.
Comparing on cost
Start on the affiliate side, because it has a published ceiling and no room to negotiate at small volume. On Amazon’s own fixed standard commission income statement as of 2026, Grocery and Health & Personal Care pay 1.00%, Home, Toys, Furniture and Pet Products pay 3.00%, Physical Books, Kitchen and Automotive pay 4.50%, Luxury Beauty pays 10.00%, and the catch-all “All Other Categories” bucket pays 4.00%. Amazon’s Associates Program Operating Agreement was last updated 15 October 2025. Run the arithmetic before you build anything: a $40 home product at 3% is $1.20 per converted click. To clear $1,200 a month you need a thousand purchases, not a thousand visitors — a thousand people who reached the checkout and finished. Plan on roughly 5% of your outbound clicks converting until your own numbers say otherwise, and that thousand purchases is 20,000 clicks a month, repeated every month, forever.
The dropshipping side has no ceiling but a long column of subtractions instead. Here is where the money actually goes on each model.
| Line item | Dropshipping | Affiliate |
|---|---|---|
| Storefront platform | Monthly subscription plus the app stack — reviews, upsells, shipping rules — which usually costs more than the subscription itself; check shopify.com/pricing for your region’s current tier prices | Domain and hosting only, or nothing at all on a social-first setup |
| Payment processing | Card rate on every order, plus Shopify’s published third-party gateway surcharge of 2% on Basic, 1% on Grow and 0.6% on Advanced if you do not use Shopify Payments | None — you never touch the money |
| Cost of goods | Paid to the supplier at order time, days before the customer’s funds settle into your account | None |
| Returns and refunds | Yours, including return shipping on non-defect returns you choose to accept | Commission reversal only |
| Chargebacks | Yours, plus a per-dispute fee — $15 in the US on Stripe and Shopify Payments, refunded only if you win, and you seldom win an “item not received” dispute on a 20-day delivery | None |
| Support labour | Every “where is my order” email, and they arrive in clusters on Monday mornings | None |
The row that hurts is not a fee row. It is the cash-flow row. You pay the supplier the moment the order lands, and your processor pays you on a rolling payout that sits two to five business days behind the sale. At 30 orders a day and $18 of goods per order, roughly $1,600 of your own money is permanently in transit. That figure scales with success, so it grows fastest in the exact week a product finally starts working — which is the worst possible week to learn you are short.
Comparing on requirements
Affiliate has a low setup bar and a high content bar. You need somewhere to publish, an approved application, and disclosure that meets the FTC’s Endorsement Guides at 16 CFR Part 255 — the 2023 revision codified “clear and conspicuous” to mean difficult to miss and, in interactive media, unavoidable. Amazon additionally requires the specific statement “As an Amazon Associate I earn from qualifying purchases” stated clearly and prominently. Below the fold does not count. Neither does 10px grey on a white background. And approval is provisional rather than final: Amazon expects three qualifying sales within 180 days of signup and closes accounts that miss it, so your first deadline after being accepted is a sales target, not a publishing one.
Dropshipping inverts that — high setup bar, low content bar. You need a business entity, or at minimum a sole-trader registration your processor will underwrite. A sales-tax registration wherever you have nexus. A supplier who will ship blind, with no third-party invoices or branding anywhere in the carton. A returns address that is not your apartment. A refund policy you can honour on your worst week, not your best one. Budget for underwriting friction too: new stores quoting long fulfilment times often get placed under a rolling reserve, meaning a slice of your revenue is withheld during precisely the months you are shortest of cash. Marketplace channels then layer on rules of their own — Amazon publishes a drop shipping policy in Seller Central requiring you to be the seller of record on all packing slips and invoices, so read that page directly before listing anything you do not physically hold.
What only shows up once you commit
Three things nobody puts in the tutorials.
Your supplier’s stock feed lies, and it lies in one direction. The CSV or API feed updates on a schedule — often nightly — so a SKU that sold out at 9am is still “in stock” on your storefront until the next sync. Worse, a feed that has silently stopped updating looks exactly like a feed reporting that everything is fine; the numbers are all there, they are simply frozen. You do not discover any of this from the feed. You discover it from a customer 11 days later asking where the package is, by which point your dispatch metrics are already damaged and your only move is a refund and an apology. The fix is boring and it works: pull the feed more often than the supplier suggests, and manually flag your top ten SKUs to zero the moment supplier stock drops below about 20 units rather than waiting for it to reach one.
Returns concentrate. The NRF and Happy Returns 2025 Retail Returns Landscape put returns at 19.3% of online sales in 2025, and 15.8% of total retail sales — $849.9 billion. That blended rate is not what you experience. In practice a handful of SKUs carry almost all of it: anything sized, anything where the photo oversells the material, anything with an assembly step in the box. The rest of the catalogue barely comes back at all. You cannot see the split until you have roughly 200 orders per SKU, which is further away than it sounds. Once you can see it, cutting the two worst offenders usually does more for margin than whatever pricing experiment you had queued up.
Affiliate earnings are not money until the return window closes. The dashboard credits the order on the day it happens; the reversal lands when the buyer sends the item back, typically four to eight weeks later. Your best month therefore gets debited during some later month, which makes month-over-month comparisons actively misleading for your first quarter — a flat-looking March is often a strong March carrying January’s returns on its back. Track commissions by the month the click happened, not the month the payment cleared, and read the dashboard total as provisional.
Which one to pick
Pick affiliate if you already have traffic and no capital, and if your category pays above roughly 4%. Below that threshold the arithmetic only closes at a content volume most people cannot sustain solo. Pick dropshipping if you have $2,000–5,000 you can leave tied up in float and refunds for six months without needing it back, and if you would rather solve logistics problems than publishing problems. Running both on the same site is a legitimate hedge rather than an inability to decide: affiliate the accessories you do not want to stock, and sell the core product yourself.
Before you decide, make two contacts. First, your prospective supplier’s named account manager — the general support inbox will not answer any of this, and getting routed back to it is itself a result. Ask exactly five things: the 30-day defect rate on the specific SKU you plan to sell; who pays return shipping on a defect, and separately on a change-of-mind return; the daily cut-off time for same-day dispatch and which timezone that clock runs on; whether they ship blind with no branding or invoices in the carton; and the restock lead time when a SKU goes to zero. Keep all of it inside the email thread. A supplier happy to answer on a call but never in writing has just told you how the next dispute will go.
Second, the affiliate manager for the program you are considering, reachable through the contact link on the program’s page inside Impact, CJ, ShareASale or Awin. Ask for the cookie window in hours, the current reversal rate for publishers in your vertical, the payment terms and minimum payout threshold, and whether commissions are clawed back on returns or only on cancellations. A program that will not answer the reversal-rate question is telling you the answer.
None of this guarantees a particular income. It does mean that when a number moves, you will know which of your own decisions moved it.
Related reading: Best Ecommerce Shipping Companies Compared: Which Carrier Fits Your Store
Sources
Figures and deadlines change. Verify at the source before you act.