Platform Fees Comparison Chart: What You Actually Keep on $1,000

You quote a client $1,000. After commission, payment processing, a withdrawal fee, and one ad placement you never asked for, $743 actually lands in your account.

That 26% gap is the single largest controllable expense in independent work. Not your laptop. Not your software subscriptions. Not your accountant. The platform layer -- the toll you pay for access to buyers -- is the biggest line item most freelancers never put on a spreadsheet.

And it is moving fast. Upwork scrapped its sliding 5/10/20% structure and went flat. Gumroad abandoned its 3.5% + $0.30 model for a flat 10%. Patreon restructured its tiers. Etsy made offsite ads mandatory above a revenue threshold. Substack normalized a 10% cut stacked on top of Stripe. Every one of those changes quietly shifted 2-8 percentage points of your revenue from your column to theirs -- and most people found out by noticing a smaller deposit.

This is the chart nobody publishes: six platform models, the same $1,000 gross, and the actual take-home after every fee is stacked. No wishy-washy analysis. A verdict for each.

Why This Decision Matters More in 2026 Than It Did in 2020

Three shifts turned platform fees from an annoyance into a strategic decision.

First, the fee increases compounded. Between 2022 and 2025, nearly every major platform restructured its pricing. Some reduced headline rates while adding new charges -- contract initiation fees, mandatory ad participation, payout fees, currency conversion spreads. The published fee schedules got longer, not shorter. If you formed your mental model of platform economics before 2023, it is out of date.

Second, AI compressed deliverable prices. When a client can get a first draft, a landing page, or a logo concept generated in ninety seconds, your pricing power on commodity deliverables drops. The Bureau of Labor Statistics continues to show independent and self-employed work as a structurally significant share of the labor force, but the middle of the market -- mid-skill, mid-price, undifferentiated -- is where margin is thinnest. When your gross rate falls 20%, a 20% platform commission hurts far more than it did when rates were climbing.

Third, owned audiences became viable. A newsletter with 3,000 engaged subscribers, a Gumroad storefront, and a Stripe account can now replace a marketplace account entirely for a lot of service businesses. That was not true in 2018. It is barely debatable now.

Stack that on top of the 15.3% self-employment tax you already owe on net earnings, and you start to see why a five-point fee difference is not a rounding error. On $80,000 of gross revenue, five points is $4,000 -- roughly the cost of a good accountant, a decent laptop, and a year of software, combined.

The Comparison Chart: Six Platform Models, One $1,000 Benchmark

Every row below answers the same question: if a customer pays $1,000, what reaches you? Numbers assume US-based sellers, domestic card payments, and published rates as of early 2026. Your mileage varies with country, payout method, and volume tier.

Two things jump out. First, the spread between the best and worst channel is $630 on the same $1,000 -- a 63% difference in take-home purely from where you chose to sell. Second, the cheapest channels require something the expensive channels give you: audience. Etsy and Udemy are not charging you a fee for nothing. They are charging you a customer acquisition fee, dressed up as a percentage.

That reframing matters. A 20% Fiverr commission on a stranger who found you in search is cheaper than a 3% Stripe fee on a customer you do not have.

Deep Dive 1: Freelance Marketplaces (Upwork, Fiverr, Freelancer)

The model: You list services or bid on jobs. The platform handles discovery, escrow, dispute resolution, and payment collection. In exchange, it takes a percentage of every transaction and often charges you to compete.

The real numbers. Upwork moved to a flat 10% freelancer service fee in 2023, but the fee surface did not shrink -- it moved. Connects cost roughly $0.15 each and most proposals cost 4-16 of them. A contract initiation fee applies on first-time client relationships below a spending threshold. Withdrawal fees apply to instant payouts. Fiverr kept its 20% seller commission and charges buyers a 5.5% service fee on top, which means your $1,000 order nets you $800 before you pay $1-3 to move the money. Freelancer.com takes 10% with a $5 minimum, which quietly becomes a 25% cut on a $20 gig. Guru tiers between 9% and 5% depending on membership. Fiverr's seller fee documentation is worth reading line by line before you price anything.

Strengths. Speed. You can go from zero to a funded contract in 72 hours. Escrow protection means you are not chasing invoices. Dispute systems mean a bad client cannot simply disappear with your work. Search traffic means a well-optimized profile gets inbound leads while you sleep. For anyone in their first 18 months of independent work, that structure is worth more than five percentage points.

Weaknesses. Fee stacking. Connects turn bidding into a pay-to-play lottery -- spend $40 in Connects to win a $600 job and your effective rate drops another 6-7%. Client ownership is restricted: taking the relationship off-platform violates terms and can get you banned. And the flat-rate change means a $3,000 project and a $300 project now cost the same percentage, which penalizes high-ticket specialists.

Ideal profile: You need volume, you need strangers, and you do not yet have a repeatable referral engine. Also good for service categories where clients genuinely comparison-shop.

Cost and effort: 10-20% plus Connects spend of 5-15% of gross for active bidders. Time investment to maintain profile quality and response rates is real -- budget 3-5 hours per week.

Verdict: Use it as a customer acquisition channel, not a home. The moment a client has paid you twice, move the relationship into your own contract and checkout. That single habit is worth roughly $80 on every $1,000 you bill.

Deep Dive 2: Creator Membership Platforms (Patreon, Substack, OnlyFans)

The model: Your audience pays you directly, on a recurring or per-piece basis, and the platform handles billing, paywalls, and churn management. You keep the audience relationship; the platform keeps a slice.

The real numbers. Patreon's current structure runs 8% on its Pro tier and 12% on Premium, plus payment processing of roughly 2.9% + $0.35 per transaction. On $1,000 that is $80 + $29.35 = $890.65 at the low tier, or $850.65 at Premium. Substack takes a flat 10% of subscription revenue and stacks Stripe processing on top: $1,000 becomes roughly $870.70. OnlyFans takes a clean 20%, leaving $800. Kick's 95/5 split is an outlier precisely because it is competing for creators with better known economics.

Patreon's published pricing is transparent, but the important detail is what is not in the headline: currency conversion (roughly 2.5% for non-USD patrons), payout fees to non-US bank accounts, and per-transaction costs on small pledges that disproportionately hurt $3 and $5 tiers. A $3 pledge can effectively lose 20%+ to combined processing -- which is why most successful creators price at $7 minimum.

Strengths. Recurring revenue changes your financial planning entirely. Predictable monthly income means you can decline bad projects. Platforms handle the unpleasant mechanics -- failed card retries, dunning emails, tax forms for international patrons, churn analytics. For a creator with 500+ engaged followers, membership revenue is the highest-quality revenue you can build.

Weaknesses. The fee stack is invisible until you read the statement. Premium tiers on Patreon cost you four extra points for features most solo creators never use. And memberships are fragile: a single algorithm change, a slow month of output, or a public mistake can trigger 10-15% churn in a quarter.

Ideal profile: You have an existing audience somewhere -- social, newsletter, podcast, Discord -- and you publish consistently. Membership is a terrible first product for someone with no audience and an excellent second product for someone with 2,000 engaged followers.

Cost and effort: 8-20% + processing. Effort is ongoing content delivery; expect 20-40% of your working time once you cross a few hundred paying members.

Verdict: If your audience skews general-interest and you want zero infrastructure, Patreon or Substack is the right call and the ~10 point cost is a fair trade for billing infrastructure. If you already have a technical setup, self-hosted Ghost or a Stripe-backed membership plugin nets you roughly $970 on $1,000 -- but you own every failed payment, every refund request, and every tax form.

Deep Dive 3: Direct Checkout and Owned Storefronts (Stripe, Gumroad, Lemon Squeezy, Shopify)

The model: You own the relationship, the list, and the checkout. The only costs are payment processing and software.

The real numbers. Stripe's published rates are 2.9% + $0.30 on domestic cards, 3.9% + $0.30 on international cards, and 4.4% + $0.30 with currency conversion. That is $970.70 net on $1,000 -- the highest take-home of any model on this chart. Gumroad's flat 10% covers processing and hosting, netting $900 with zero infrastructure. Lemon Squeezy charges 5% + $0.50 and handles merchant-of-record tax obligations globally, which for a seller with international customers is worth real money in compliance time. Shopify starts around $39/month plus 2.9% + $0.30 -- and adds up to 2% extra if you use a third-party gateway instead of Shopify Payments.

The critical math here is the break-even against a percentage-based platform. A $39/month Shopify plan equals a 10% commission on $390 of monthly revenue. Below that, the subscription is the more expensive choice. Above $1,500/month, Gumroad's 10% costs you $150 and the subscription model wins decisively.

Strengths. Highest take-home. Full control of customer data, upsells, and pricing. No platform can ban you, change your fee structure, or insert itself between you and your buyer. Email capture means every sale compounds into future free sales.

Weaknesses. Zero discovery. Nobody browses Stripe. You must generate 100% of your own traffic, which means content, SEO, referrals, paid ads, or an audience. You also inherit compliance work: sales tax registration, VAT for EU customers, refund handling, chargeback disputes (Stripe charges $15 per lost dispute), and abandoned-cart recovery.

Ideal profile: You have an existing audience, a referral network, or a repeat client base. Also ideal for anyone selling to businesses, where direct invoicing is the norm anyway.

Cost and effort: 3-10% + $0-39/month software. Effort is front-loaded: building the store, writing the sales page, setting up email flows. Expect 20-40 hours of setup for a proper storefront.

Verdict: If you are doing more than $2,000/month in sales and you have any audience at all, direct checkout is the correct answer, full stop. The 10-17 point take-home advantage over marketplaces is the difference between a business and a side hustle.

I sold digital templates on two marketplaces for three years and thought the 20% was just the cost of doing business. Then I ran the numbers properly. On $6,400 of monthly revenue I was handing over $1,280 in commission and another $180 in ad fees I had not opted into carefully. I moved the same catalog to a Gumroad store plus Stripe invoices for repeat buyers, spent six weeks building an email list from my old customers -- which the marketplace terms technically allowed since I had their email from support conversations -- and my take-home went from about $4,900 to $6,200 on the same volume. The fees were never the problem. Not knowing what they were was the problem.

Priya Raghavan, former e-commerce operations manager, now full-time digital product seller

Deep Dive 4: Product Marketplaces (Etsy, Amazon Handmade, eBay, Poshmark)

The model: You list physical or digital goods in a search-driven marketplace. The platform supplies demand; you supply inventory, fulfillment, and customer service.

The real numbers. Etsy's fee schedule is a masterclass in fee stacking. Listing costs $0.20 per item (renewed every four months, and again on every sale). The transaction fee is 6.5%. Payment processing is roughly 3% + $0.25. Offsite ads take 12% -- or 15% if you are below the $10,000 annual threshold, and participation becomes mandatory once you cross it. Add it up on a $1,000 order and you are at roughly $905 before ads, and roughly $754 after a 15% offsite ad attribution. Amazon Handmade takes a flat 15% referral fee with no listing fees. eBay's final value fees average around 13.6% across categories plus $0.30 per order. Poshmark takes a flat 20% on sales over $15. Depop takes 10%.

The offsite ads detail is the one that catches people. You do not control when a buyer clicks an offsite ad; Etsy attributes the sale and takes 12-15% on top of everything else. Sellers routinely report this turning an 8% margin product into a break-even sale.

Strengths. Genuine demand generation. Etsy has hundreds of millions of annual buyers actively searching. For a maker with no audience and no ad budget, that traffic is otherwise unaffordable. Amazon Handmade brings Prime-adjacent trust and returns infrastructure that would cost you thousands to replicate.

Weaknesses. Fee opacity. Attribution rules that you did not negotiate. Constant policy shifts. And commoditization pressure -- search results push toward the lowest price, which is brutal when a competitor is running the same print-on-demand supplier.

Ideal profile: Physical makers, print-on-demand sellers, vintage dealers, and anyone whose product benefits from being discovered by strangers.

Cost and effort: 9.5-20% plus ads. Effort is heavily front-loaded into listing photos, SEO, and review generation -- 40-60 hours to get a shop to steady-state.

Verdict: Use marketplaces for discovery, then convert buyers into direct customers. Insert a card, a thank-you note, or a follow-up email that invites them to your own storefront. That is not a loophole; it is the only way the 15% economics work long-term.

Deep Dive 5: Course and Content Platforms (Udemy, Skillshare, YouTube, Music Streaming)

The model: You create once, the platform distributes forever, and you take whatever share of revenue the platform decides.

The real numbers. These are the harshest economics on the chart and they are almost never discussed honestly. Udemy pays instructors 37% of list price when a course sells through Udemy's own search and marketing, 50% on Udemy Ads-driven sales, and 97% when you drive the sale yourself with your own referral link or coupon. On a $1,000 organic sale, you keep $370. Skillshare pays from a royalty pool based on minutes watched, which for most instructors works out to somewhere between $0.05 and $0.15 per minute of engagement -- a figure that swings wildly quarter to quarter with zero transparency. YouTube pays creators 55% of ad revenue on long-form video. Spotify pays roughly $0.003 to $0.005 per stream, meaning a million streams generates somewhere around $3,000-$5,000 before your distributor's cut.

Udemy's instructor revenue share is the important one to internalize, because it reveals the actual structure: the 37% is not a fee, it is a customer acquisition cost. Udemy spends heavily on paid search and its cut reflects that. The 97% tier exists for instructors who bring their own students -- and almost nobody uses it because building an audience is hard.

Strengths. Zero marketing effort. Passive income with no ongoing client management. A course that earns $400/month on Udemy requires roughly zero maintenance after launch. For someone who wants to teach but hates selling, this is the trade.

Weaknesses. You do not own the student email list on most platforms (Udemy restricts direct contact in many contexts). Pricing is set or heavily influenced by the platform. Algorithm changes can cut your revenue 70% overnight with no recourse. And the implicit ceiling is low: most Udemy instructors never cross $500/month.

Ideal profile: You have genuine expertise, you want to teach, and you are willing to treat revenue as a bonus rather than a plan.

Cost and effort: 45-63% revenue share. 80-200 hours to produce a course that meets platform quality standards.

Verdict: Treat Udemy and Skillshare as a lead-generation channel, not a business. Publish the course cheaply, capture what audience you can, and migrate serious students to a self-hosted platform like Teachable or Kajabi where you keep 90%+ and own the customer relationship. The launch is the free sample; the real business is the thing you sell afterward.

The Hybrid Layer Nobody Charts: Gig and Rental Apps

If you drive, deliver, or rent, the same logic applies with different numbers. Uber's service fee averages around 25% of the fare and varies by market and trip. DoorDash commissions run 15-30% for restaurants plus delivery and marketing fees. Airbnb charges hosts either a 3% host-only fee or up to 14-16% under simplified pricing, with guests paying roughly 14% on top. None of these are negotiable, and all of them changed materially between 2019 and 2025. The pattern is identical: a low headline number, a complex stack, and a platform that owns the customer relationship.

Best For: Match Your Situation to the Right Channel

Here is the verdict section. Find your situation and commit.

You have zero clients and zero audience. Freelance marketplace. Pay the 10-20%. It is tuition. Spend six months building reviews and repeat business, then start migrating your best clients to direct checkout.

You have a specialist skill and five or more years of experience. Talent network or direct. Apply to Toptal-style networks where the client markup is invisible and you keep 100% of your quoted rate -- or go fully direct with Stripe and outbound sales. Do not compete on a marketplace with junior sellers who will undercut you.

You have 2,000+ engaged followers. Creator membership or direct checkout. Membership gives you recurring revenue; direct checkout gives you the best margin. Most successful operators run both: a membership for baseline income and a storefront for one-off products.

You make physical goods. Product marketplace for discovery, own storefront for repeat buyers. The Etsy-to-Shopify migration is the single most reliable profit improvement available to a small maker.

You want to teach. Self-hosted course platform first if you have any audience. Udemy only if you have none and are willing to accept a 37% share as a marketing cost.

You want maximum take-home and will do the work. Stripe plus an email list. This is the highest-value option on the entire chart and it is also the hardest, because nobody hands you customers.

The Decision Framework: Five Questions, Asked in Order

Work through these in sequence. Stop at the first one you cannot answer.

One: Can you generate 20 customers per month without a marketplace? If yes, skip marketplaces entirely. If no, marketplace first.

Two: What is your average sale price? Below $50, percentage-based platforms are almost always better -- the fee scales with your small transaction and you avoid monthly subscriptions. Above $500, subscription-based software wins fast.

Three: How much of your revenue is repeat business? Above 40% repeat, move to direct checkout immediately. Repeat customers do not need discovery, and you are paying discovery fees on them for nothing.

Four: What is your monthly volume? Under $400/month, use percentage platforms and keep zero fixed costs. Over $2,000/month, run the break-even: monthly software cost divided by your net margin percentage. If software is under 3% of revenue, switch.

Five: How much compliance work can you absorb? Direct checkout means sales tax registration, VAT handling, and chargeback management. If that is a non-starter, Lemon Squeezy or Gumroad at 5-10% buys it back cheaply.

If you want a structured way to run this comparison against your actual numbers rather than a generic table, use the Income Architect at Workings.me. It is built for exactly this -- mapping income channels against each other so you can see which combination produces the best net result rather than the highest gross.

Six Fee Traps That Cost You More Than the Commission

1. Currency conversion spreads. Platforms rarely charge a visible FX fee. They bury 2-2.5% in an exchange rate that is slightly worse than the interbank rate. On $40,000 of international revenue, that is $1,000 you never see on a statement.

2. Withdrawal and payout fees. Instant payouts on Upwork run about $0.99 each; Fiverr charges varying fees by method; Patreon charges on non-US bank payouts. Ten instant withdrawals a month is $120/year for the privilege of access to your own money.

3. Payment processing on top of commission. Patreon, Substack, Etsy, and Shopify all layer processing on top of their platform fee. Fiverr and Udemy bake it in. Always ask which one you are looking at.

4. Ad attribution you did not opt into. Etsy offsite ads and Amazon sponsored listings both trigger on traffic you assumed was organic. Read the attribution rules before you price a product.

5. Failed payment recovery. On direct checkout, a failed card is your problem. Roughly 5-10% of subscription payments fail at some point, and recovering them requires dunning emails and retries. That is real time, and the platforms charging 8-12% are charging for it.

6. Chargeback fees. Stripe charges $15 per dispute regardless of outcome. Five disputes a year is $75 plus the lost product plus your time assembling evidence.

Insider Tips From People Who Have Done the Migration

Run both channels during the transition, always. Nobody successfully moves from a marketplace to direct checkout overnight. Keep the marketplace active for discovery while you build direct revenue in parallel. Cut ties only when direct revenue consistently exceeds marketplace revenue for three consecutive months.

Track effective fee rate, not headline. Once a quarter, divide total platform charges (commission + processing + connects + ads + software + payout fees) by gross revenue. That single number is your platform tax rate. Most people guess 15% and discover it is 26%.

Price for the stack, not the sticker. If your channel costs 20% all-in and you want $60/hour, charge $75. Too many freelancers set rates against the headline fee and then wonder why their effective hourly rate keeps falling.

Never let a marketplace own your email list. Where terms allow it, move every buyer into a system you control. A 2,000-person list is worth more than a five-year marketplace review profile, because the list cannot be delisted.

Recheck fees every January. Every platform on this chart has changed its structure at least once since 2022. The chart you built eighteen months ago is probably wrong.

Do the math before the emotional move. Leaving a marketplace because the commission feels unfair, before you have 100 direct customers, is the most common financial mistake independent workers make. The fee was never the problem. The dependency was.