06-reference/research

mac platform comparison stripe vs ls vs gumroad

2026-05-18·research-brief·source: deep-research
researchmacplatform-comparisonstripelemon-squeezygumroadmerchant-of-recordeu-vatrefund-ratelaunch-prep

MAC Platform Comparison — Stripe Direct vs Lemon Squeezy vs Gumroad

Context

MAC ships at $350 one-time, framed as a tool (not a course), targeting working data/dev engineers. Time-to-first-value target: <5 minutes from purchase to running /dq-plan against a real project. Founder wants to ship soon and is choosing the checkout/billing platform.

Refund-rate environment is hostile: platform-wide info-product refunds are ~21% in Q1 2026, premium-course band is 28%. MAC's tool-not-course framing is the primary lever to beat this, but platform choice is a real second-order lever — especially around MoR (who eats EU VAT compliance), buyer-initiated refund UX, and chargeback liability.

Comparison table

Axis Stripe Direct Lemon Squeezy Gumroad
Effective fee on $350 one-time ~2.9% + $0.30 = $10.45 (3.0%) domestic; ~4.4% + $0.30 = $15.70 (4.5%) international 5% + $0.50 = $18.00 (5.1%) domestic; 6.5% + $0.50 = $23.25 (6.6%) international 10% + 2.9% + $0.30 + $0.50 = $45.85 (13.1%) all-in (platform + processor, before currency spread)
EU VAT / sales tax / MoR You are MoR. Must register for VAT in EU (or use OSS threshold), file remittances, integrate Stripe Tax (+0.5%). Real burden + audit risk. Paddle is MoR (LS rolled under Paddle's MoR infra post-acquisition). Handles VAT/GST in 100+ countries. Zero filing burden. Gumroad is MoR (since Jan 2025). Same zero-burden model.
Refund UX (matters at 21% baseline) You build it. Manual via dashboard or custom flow. Stripe keeps processing fee on refunds ($0.30 + the % — non-recoverable). Self-serve buyer-initiated refund flow; LS keeps the 5% + $0.50 platform fee on refunds. Chargeback fee: $15/case. Self-serve refund UI; Gumroad keeps 10% + $0.50 platform fee AND processor keeps 2.9% + $0.30. On a $350 refund you lose ~$13.70 in non-recoverable fees per event.
Payout cadence / reserves T+2 standard once verified. New high-ticket digital-goods accounts often get 7-14 day initial hold; rolling reserves possible (10-25% for 90-180 days) if flagged as high-risk. Weekly payouts (Mon). Approval can take days-to-week post-acquisition. No public rolling-reserve pattern for low volume. Weekly payouts. Currency converted to USD then back to your local — 1-2% hidden spread.
Embedded vs hosted checkout Full control. Stripe Elements / Payment Element embedded in your domain. Best for <5min TTFV — buyer never leaves the MAC sales page. Hosted checkout (overlay or hosted page). Decent embed but always LS-branded. Buyer leaves your domain. Hosted only. Gumroad-branded overlay or full redirect. Most "amateur" feeling of the three.
Integration time to ship MAC 8-16 hours: webhook handler, license-key issuance, tax registration setup, receipt template, refund flow. 2-4 hours: API + webhook for license issuance, hosted checkout link, that's it. 30-60 minutes: paste embed snippet, configure product, done. License-key delivery via Gumroad's built-in custom field.
Anti-fraud / chargeback Stripe Radar (good) but you eat every chargeback ($15 fee + lost revenue + reserve impact). High-ticket digital is a chargeback magnet. LS/Paddle absorbs chargeback liability as MoR — you're insulated from the financial hit but they may pull funds back if dispute volume spikes. $15 chargeback fee passed through. Gumroad as MoR absorbs chargeback liability. Less aggressive fraud screening than LS/Paddle historically.
Affiliate / referral Build it yourself (Rewardful $99/mo or Tolt is the common stack). Built-in affiliate program, configurable per-product. Standard for indie launches. Built-in affiliate program. Discover marketplace (30% fee) doubles as organic discovery.
Sales data export / analytics Best-in-class. Full API, raw data, BigQuery sync available. Real attribution possible. Solid dashboard + CSV export + API. Adequate for quarterly P&L. Basic dashboard, CSV export. Weakest of the three for proper attribution.
Crypto / international payment Apple Pay, Google Pay, SEPA, BACS, iDEAL, 130+ currencies. No native crypto. Major cards + Apple/Google Pay. No crypto. Major cards + PayPal + Apple/Google Pay. No crypto.
Dev-tool / indie-hacker reputation Signals "serious." Used by every YC startup. Custom checkout = professional. Signals "indie hacker / dev tool." Built by devs for devs, well-loved in the segment. Some post-acquisition stability concerns (checkout errors, slow support reported) but still the segment-native choice. Signals "creator / amateur." Strong with newsletter writers, ebook authors, course sellers. Reads as "side project" to a dev-tool buyer.

Net economics on a $350 sale ($350 sticker, no discount)

Platform Take-home (domestic) Take-home (international) On refund, you lose
Stripe Direct $339.55 $334.30 $10.45 (fees) + tax-reg overhead
Lemon Squeezy $332.00 $326.75 $18.00 + $15 if chargeback
Gumroad $304.15 ~$298 $13.70 minimum

On a 100-sale launch at 12% refund rate: Stripe ≈ $30,000 net; LS ≈ $29,000 net; Gumroad ≈ $26,800 net. Spread between Stripe and LS is ~$1,000 on the launch — material but not decisive.

RAY'S RECOMMENDATION

Pick: Lemon Squeezy.

The $1k fee delta vs Stripe Direct is bought back many times over by (a) Paddle absorbing EU VAT registration and filing — a real liability for a US founder selling globally, and (b) absorbing chargeback financial liability in a 21%-refund-rate environment where high-ticket digital is a known chargeback magnet. Stripe Direct only wins if MAC is heading to $300k+ ARR fast enough to justify hiring/outsourcing compliance, which this launch does not need to assume. Gumroad is disqualified on take-home (13% all-in vs 5%) AND on dev-tool credibility — Gumroad-branded checkout signals "amateur" to the exact buyer MAC targets.

Two gotchas the founder should flag before committing

  1. LS post-acquisition stability risk. Multiple 2026 reports (TheSoftwareScout, DEV community, daveswift) flag checkout errors, double-charges, and slow support since the Paddle/Stripe-adjacent acquisition. Mitigation: (a) test the full purchase → license-delivery flow end-to-end before launch day with a real card, (b) keep a Stripe account warmed up as a 24-hour fallback if LS checkout breaks during launch window, (c) monitor LS status page during launch hours. The $15 chargeback fee is also worth pricing into the refund-rate model — at 12% refund rate with 2% of those going to chargeback, that's another $30-50 in fees per 100 sales.

  2. <5min TTFV path is harder on LS than on Stripe. LS uses a hosted checkout overlay — the buyer leaves the MAC sales-page domain for the payment step, which adds perceived friction and a brand-handoff moment. To hit <5min from "buy" to "first /dq-plan run," the post-purchase license-delivery email + onboarding link needs to be instant (webhook-driven, not the LS default email) and the install instructions need to be one terminal command. Confirm: does LS send the license key via webhook fast enough that the welcome email with install instructions arrives within 30 seconds of purchase? If not, the TTFV story breaks and the LS hosted-checkout friction compounds it.

Sources