Where to sell ebooks: Amazon, Etsy, or direct?

Four serious options, each a different trade of reach against control. Amazon KDP offers the biggest built-in audience and pays 35–70% depending on price and territory. Etsy offers search-driven buyers who love workbooks, planners, and printables, for roughly 9.5% + $0.25 per sale plus a $0.20 listing fee. Gumroad (or any direct checkout) keeps the most per sale at 10% + $0.50 — but brings zero traffic. Your own book page earns trust and emails while retail handles the transaction.

That last model is exactly how EarnDraft book pages work: your live page is the lead and entry point, with purchase links out to the retailers. There is no on-platform book checkout to compare here — the page is reach, and the fee row belongs to wherever the reader actually pays.

Most authors end up multi-channel within a year. The question is only where to start, which is what the widget answers from your situation rather than from platform hype.

Run on a $9.99 US sale, the published schedules shake out like this. Amazon at 70% keeps about 30% plus delivery (~$0.30 on a small file). Gumroad direct takes 10% + $0.50. Etsy takes 6.5% transaction plus ~3% + $0.25 processing, with the $0.20 listing fee spread across your sales. Direct wins per-sale; Amazon usually wins per-month through sheer volume.

Two asterisks. Gumroad's 30% Discover rate applies only to sales through its marketplace, not your own links — don't blend the two. Etsy's Offsite Ads can add 12–15% on attributed sales, which quietly makes it the priciest channel for advertised orders.

Fifty copies at $9.99 is $499.50 of reader spend. On Amazon at 70% you keep about $335; via Gumroad direct about $425; via Etsy about $427 before listing fees. The per-sale gap between best and worst is under $2 — but over a year at this volume it's roughly a $1,000 decision.

Now flip the lens to volume, because channels don't sell equally. If Amazon's discovery delivers 80 sales where your own links deliver 20, Amazon pays $536 against direct's $170. Reach beats margin until you own traffic — which is why the widget asks about your audience, not just your price.

The synthesis most independents land on: start where the buyers already are, build the list from every sale, and migrate margin to direct as the audience compounds. Start Amazon, end everywhere.

KDP Select buys Kindle Unlimited page-read royalties, Countdown Deals, and free-promo days — priced at 90-day digital exclusivity. That trade wins when your readers live inside KU (genre fiction, serial nonfiction) and you have no audience to sell direct to. Model the KU side with the KENP estimator before enrolling.

Exclusivity loses when you sell elsewhere meaningfully: Etsy planners, direct workshop bundles, library distribution, or a warm list that converts on your own links. A single corporate workshop licensing 100 copies dwarfs a quarter of KU reads — don't lock that door for promo tools you won't use.

Remember it's a 90-day term, not a tattoo. Enroll, measure one full term of reads plus promos against your wide baseline, and re-decide with data. Authors who re-decide quarterly beat authors who decide once forever.

Your own page does three jobs retail can't: it captures emails, it tells the story at full length, and it survives algorithm changes. Every channel should point home — Amazon's Author Central, your Etsy shop announcement, your Gumroad receipts — because followers on rented land are a campaign away from gone.

EarnDraft gives every book that page automatically: cover, contents, preview reader, and purchase links to wherever you list. Treat it as the hub and retailers as spokes — announce on your page, sell wherever the reader prefers, collect the email in between.

One honest limit: a page without traffic is a brochure. The page compounds whatever you feed it — newsletter swaps, social proof, search — but feeds itself to no one. Budget promotion the same week you publish, not someday.

No audience, straight ebook, exclusivity acceptable: start KDP Select and chase KU reads plus reviews for 90 days. Small audience or non-exclusive: go wide from day one — KDP plus a direct link on your own page. Workbooks, planners, printables: start Etsy, add KDP second once reviews prove the concept.

Large audience: sell direct first (highest margin, most emails), keep Amazon as the discovery spoke. Whatever you pick, put every edition's numbers through the royalty calculator so the decision stays arithmetic, not tribal.

Then prove it with a 90-day experiment: one primary channel, one challenger, identical ammunition (same price, cover, blurb). Track units, revenue after fees, and emails weekly per channel — and pre-commit the decision rule: the challenger wins if it beats the incumbent on profit-per-hour for two straight months. Experiments without pre-commitment end with whichever dashboard you looked at last winning.

Revisit quarterly. Channels change fees, your audience grows, and last year's right answer expires. The authors who win distribution treat it as a portfolio review, not an identity.

Retail is only the middle of the distribution map. Libraries (through aggregators) pay per borrow and put your book in front of readers who never shop indie shelves. Bundles with complementary authors trade audiences at zero ad cost — your workbook plus their course beats either alone. And a single corporate workshop can license 100 copies at full price, dwarfing a quarter of retail royalties.

These channels share one prerequisite: a professional artifact. No organizer licenses a sloppy PDF, and no bundle partner risks their reputation on an unproofed draft. The production budget isn't just retail preparation — it's the ticket to every higher-margin channel on this list.

Sequence them by effort: retail first (always on), libraries second (one setup afternoon), bundles third (relationships take quarters), direct licensing last (needs proof and testimonials). Each layer compounds the last because every channel's buyers join the same list — which is why the lead magnet generator belongs in every distribution plan.

When the numbers say so, reviewed quarterly. Leave (or deprioritize) a channel when its hourly return trails your best channel for two consecutive quarters — counting setup, maintenance, and customer support, not just fees. A store paying $40 a month that costs you three hours isn't a channel, it's a hobby with a login.

Common exit triggers: Etsy Offsite Ads turning your margin negative on attributed sales, a direct checkout whose support load exceeds its premium over Amazon, or KDP Select terms whose page-read payouts stopped beating your wide baseline. Exits should be scheduled decisions with dates, not frustrated midnight deletions — delist cleanly, redirect links to your own page, and email the list first.

Never exit to zero channels. Keep at least retail-plus-one (your page counts as the plus one) so no single policy change or fee hike can zero your income. Distribution is a portfolio: rebalance it, don't gamble it.