Should you sell your ebook on Gumroad or Amazon KDP?

The per-sale math on a $4.99 ebook

Start with money, because everything else is commentary. On a $4.99 ebook sold direct through Gumroad, the fee is 10% plus 50 cents — about $1.00 off — leaving you roughly $3.99 per sale. No monthly fee, no listing fee; those figures come straight from Gumroad's pricing page, checked September 2026.

The same book on KDP at the 70% royalty option earns 70% of list price minus a small delivery deduction — roughly $3.45 per sale at $4.99. At the 35% option it earns about $1.75. The 70% band runs $2.99–$12.99 on Amazon.com at last check, so $4.99 sits comfortably inside it.

Margin per sale: Gumroad direct wins by about fifty cents. But margin per sale is not profit — profit is margin times copies, and copies are where the two shelves diverge completely.

Discovery: Amazon's shelves versus your links

KDP is a search engine with a checkout button and a subscription service attached. Millions of readers arrive with intent, Kindle Unlimited subscribers borrow without thinking about price, and categories plus keywords put your book in front of strangers while you sleep. For a first ebook with no audience, that built-in discovery is KDP's entire argument.

Gumroad is a checkout link you own: your product page, your customer emails, your bundling and discount codes. Setup takes minutes and the relationship is yours. The tradeoff is total: nobody browses Gumroad the way they browse Amazon, and Discover marketplace sales cost 30% — nearly wiping out the margin advantage.

The honest summary: Amazon rents you shoppers; Gumroad gives you control. No audience plus Gumroad alone means no traffic, and no traffic means no sales regardless of how good the book is.

Control: pricing, readers, and rules

On Gumroad you set any price, change it hourly, bundle freely, discount by code, and email every buyer. Your reader list compounds with each sale — the asset that makes book two launch bigger than book one.

On KDP you play in Amazon's store: price bands gate the 70% royalty, categories and keywords follow their taxonomy, reviews follow their policies, and the customer relationship stays with Amazon. KDP Select adds Kindle Unlimited page-read income and promo tools in exchange for 90-day digital exclusivity — no selling the same ebook anywhere else, including your own site.

Neither is wrong; they tax different things. Amazon taxes control and margin. Gumroad taxes your time spent finding readers. Pick the tax you can afford right now.

• Want shoppers without an audience: KDP (test Select for one term).
• Want margin and reader emails: Gumroad direct with your own traffic.
• Want both: KDP wide (no Select) plus Gumroad — allowed together.

What changes for AI-assisted books

On KDP, one extra step: the AI-content questions. Disclose AI-generated text, images, or translations honestly — even if you edited afterward — and skip disclosure for pure AI-assisted help like brainstorming or proofreading your own words. Answering wrong risks removal; answering right costs nothing. Our disclosure explainer covers the exact clicks.

On Gumroad there is no disclosure gate, but the same honesty applies as marketing: readers who discover undisclosed AI help feel deceived, while a brief process note on the copyright page builds trust. Transparency converts; evasion corrodes.

Either way, verify the manuscript before it touches either shelf: our accuracy workflow exists because one invented statistic refunds harder on Amazon — where reviews are permanent — than anywhere else.

The decision for book number one

No audience yet: publish on KDP for discovery and add Gumroad the week you have any traffic source — a newsletter, a community, a viral post. The Gumroad page costs nothing to maintain and captures full margin from readers you found yourself.

Existing audience, even small: lead with Gumroad direct for margin and emails, keep KDP as the discovery net for strangers. And if the book feeds a course or coaching offer, stay wide of KDP Select — exclusivity forbids selling the ebook on your own funnel pages, exactly where it earns most.

Revisit every 90 days with real numbers: copies per shelf, margin per shelf, and where buyers came from. Our KDP selling walkthrough covers the setup; this page covers the strategy. Data beats dogma — including ours.

The third shelf: where Etsy fits

Etsy belongs in this comparison because for workbooks, planners, and templated guides it often beats both: built-in search traffic from buyers who purchase digital downloads habitually, at $0.20 per listing plus 6.5% transaction and standard payment processing — roughly $4.07 kept on a $4.99 sale by our math, the best margin of the three before volume effects.

Etsy's tax is different from both rivals: listing hygiene (titles, tags, thumbnails, renewals every four months), marketplace rules you don't control, and customers who belong to Etsy first. It is discovery you rent by the listing rather than by the royalty split.

The fit test is product-shaped: printable, visual, or template-driven books thrive on Etsy; straight reading books do better on KDP; audience-driven sales convert best on Gumroad. Match the shelf to the artifact — our Etsy-versus-Gumroad guide covers the direct-side decision in full.

Nothing forbids all three eventually: KDP for readers, Etsy for searchers, Gumroad for your own traffic — with the single constraint that KDP Select exclusivity, if you enroll, bans the other two for the term. Sequence beats simultaneity: win one shelf, then add the next with earnings.

Pricing experiments that decide for you

Opinions about the right price are cheap; thirty-day experiments are data. Run launch price for a month, change exactly one variable, and compare copies, revenue, and review velocity — our pricing walkthrough maps length to starting bands so the first test starts sane.

Test free strategically, not desperately: a free weekend on a wide book (never under Select exclusivity confusion — know your terms) spikes sampling and reviews, while permanent free belongs to lead magnets feeding a funnel. Free is a tactic with a purpose and a date, not a surrender.

Watch the metric that matters per shelf: on KDP, conversion and review velocity move the algorithm; on Gumroad, email capture per visitor compounds beyond the sale; on Etsy, click-through from search decides everything. Different shelves, different scoreboards — optimize the one you're standing in.

Revisit the whole mix every 90 days with the per-sale math from this page refreshed against current fee pages: fees change, your audience grows, and the shelf that won last quarter owes its position nothing. Strategy is maintenance, and maintenance has a calendar.

• One variable per test, thirty days per test, revenue plus reviews scored.
• Free with a purpose and an end date — never as default.
• Re-decide the mix quarterly against current fee pages.

Switching shelves without losing sales

Migrations fail from simultaneity, not strategy: leaving KDP Select, add the direct page first and confirm it converts before touching exclusivity settings, then exit Select at term end with auto-renewal off. Overlapping availability beats gap days — readers buy on impulse, and impulses don't wait for DNS.

Carry the social proof with you: reviews stay on Amazon, so quote the best lines (attributed, verbatim) on the new page, and email existing buyers the new direct link with a reason to care — a bonus chapter, a bundle price, early sequel access. The audience is portable even when the reviews aren't.

Price the new shelf deliberately rather than mirroring: direct pages can bundle, discount by code, and test freely, while marketplace prices anchor expectations. Our pricing walkthrough sets starting bands; the new shelf earns its own thirty-day experiment from day one.

Watch the exclusivity calendar like a contract, because it is one: Select terms, promo schedules, and price-match delays all constrain timing. Plan migrations at term boundaries with a written checklist — the switches done calmly in week one beat the scramble when a promo collides with a move.

Keep every shelf's admin current during transitions: tax forms, payment details, and contact emails verified before they're needed. Nothing stalls a migration like a payout hold discovered mid-launch — boring paperwork, checked early, is invisible; checked late, it's the launch story.

After the move, run the per-sale math again with real numbers: copies per shelf, margin per shelf, traffic source per buyer. The strategy page you're reading made predictions; your dashboard grades them. Re-decide quarterly, and let the shelves compete honestly for your next book.

Bundles, discounts, and the email asset

Direct shelves win on mechanics marketplaces forbid: bundle the ebook with its workbook, discount by code for communities, price-bump the second title at checkout. Average order value is the lever Amazon never hands you — a $9.99 book plus a $7 companion beats a single $14.99 listing with room to spare.

Discounts need fences to protect value: time-bound launch codes, segment-specific offers, bonuses that expire. Permanent discounting trains buyers to wait; fenced discounting trains them to act. The fence is the strategy — without it you're just cheap.

The email list is the actual asset every shelf builds toward: direct buyers join automatically, marketplace buyers join through back-matter incentives (bonus chapters, templates, sequel alerts). A thousand past buyers reachable for free is what makes book two launch bigger than book one — no algorithm required.

Sequence the funnel deliberately: free magnet captures, tripwire converts, core ebook delivers, bundle ascends. Each asset from our free-versus-paid guide has its rung; shelves are just where rungs get stood on. Build the ladder once, sell climbs forever.

Measure the mechanics monthly: attach rate on bundles, redemption on codes, signup rate per buyer. Small numbers, reviewed regularly, compound into the margin advantage that makes direct selling worth its traffic burden — unmeasured mechanics decay into unused features within a quarter.

One caution: complexity taxes the buyer too. Three clear offers beat nine clever ones — bundle simply, discount rarely, and email consistently. The shelf that feels effortless to buy from wins the repeat purchase, which is where side-hustle math turns from linear to compounding.

Sources checked (September 2026)

All fee and program figures verified against official pages this month; confirm before pricing — terms change.

• Gumroad pricing: https://gumroad.com/pricing
• KDP digital pricing (35%/70%): https://kdp.amazon.com/en_US/help/topic/G200634500
• KDP Select terms: https://kdp.amazon.com/en_US/help/topic/G200798990
• KDP AI-content rules: https://kdp.amazon.com/en_US/help/topic/G200672390
• Etsy fees: https://www.etsy.com/legal/fees/