The pricing page looks clean: 10% + $0.50 for profile and direct-link sales; 30% when a new customer finds and buys through Discover. No monthly fee—Gumroad only takes a cut when you sell. That is the public pricing-page line.

The trap is treating the first number as all-in. The Help Center is blunter: the direct 10% + $0.50 excludes card processing (commonly about 2.9% + $0.30) and PayPal fees; Discover’s 30% folds processing into one cut. Stack platform and card fees, and a US-card direct sale lands near 12.9% + $0.80—the “about 13%” people quote is usually that total, not the prettier 10% on the marketing page.

The spread is not just a rate card. The product question is which capability Gumroad prices as checkout infrastructure, and which it prices as customer acquisition.

One receipt, two businesses

Direct sale: the buyer hits your link, profile, email, or social post. Gumroad hosts checkout and delivery, and since January 1, 2025 acts as merchant of record for worldwide sales-tax collection and remittance. You bring the person to the door; the fee sits closer to “payments + store plumbing.”

Discover sale: the buyer browses or searches on Gumroad and meets your product for the first time. The platform story is “we found you a customer,” so the take jumps to 30% with processing included. Omid Saffari’s August 2026 check against the live site and Help Center puts it sharply: Discover behaves like customer-acquisition cost, not a payment fee. If you already paid for the click via email, ads, or a partner, and the sale is still attributed to the marketplace, you are buying your own traffic at a 30% rate.

That is the counter-intuitive bit. Many creators hear “list on a platform = distribution” and treat 10% as the whole story. The real design is dual-track: buy convenience when you own the audience; buy acquisition when the market finds the buyer. The fee gap is intentional pricing of two different jobs.

A small ledger makes the decision concrete

Take a common $50 digital product (rates vary by card and region; the table follows Help Center card figures for the direct path).

Attribution	How fees stack	Rough take	Rough keep
Direct card	10% + $0.50, plus ~2.9% + $0.30	~$7.25	~$42.75
Discover	Flat 30% (processing included)	$15	$35

Same SKU, one attribution row apart, nearly 2× the fee. Lower prices make the fixed cents hurt more: Omid’s example puts a $5 direct card sale near a ~30% effective rate—low-ticket downloads, single prompts, tiny icon packs sit in the blast zone. Bundle up, or move to a cheaper checkout stack.

Blended rates matter too. At 80% direct / 20% Discover, your platform-side weighted cut already exceeds the “I’m on a 10% platform” feeling; push Discover higher and the books drift toward a quiet 15%–20% platform. The dangerous number is not any single published line—it is how many orders land on the expensive one.

Attribution hygiene is margin work

The product move is concrete: ads, newsletters, and social bios should land on the direct product URL, not dump people into Gumroad search to “happen” to buy you. Omid’s ops line is almost identical—traffic you brought should not take the 30% marketplace path.

Discover is still usable. When it truly delivers buyers you cannot reach, 30% can be list-buying or trial CAC. The scoreboard is repeat purchase, owned-list capture, and upsell into higher tickets. If the relationship ends at one $50 sale with $15 to the market, the economics need no footnote.

Gumroad still fits a clear persona: validate a paid download or membership fast, skip a monthly SaaS bill, and avoid becoming the global tax operator. Merchant-of-record tax transfer is a major reason to accept double-digit variable fees. Once the audience is stable and sales are predictable, you are funding Gumroad’s cut while supplying the traffic yourself—convenience then has to repay enough tax, delivery, and support relief, or lower platform fees / self-hosted checkout win.

Sources

• Gumroad pricing
• Gumroad Help Center: fees
• Omid Saffari: Gumroad Review (verified Aug 2026)