Skip to content
Sellarix

Marketplaces vs Your Own Store: Should You Sell on Amazon, eBay and TikTok Shop?

The first time I watched a product go viral on someone else's platform, I was thrilled for about a week. Then the invoice landed. The marketplace had...

The Sellarix team · 1 Apr 2026 · 6 min read

The first time I watched a product go viral on someone else's platform, I was thrilled for about a week. Then the invoice landed. The marketplace had taken its cut, the ad fees stacked up, and worst of all, I had no idea who any of those buyers were. Hundreds of orders, zero email addresses. That's the moment most operators realize the real question isn't where do I sell more. It's who actually owns the relationship when the dust settles. I've run this experiment from both sides. I've launched products that lived and died on Amazon search, and I've built a direct store that started slow and compounded into something I actually controlled. Neither is "right." They're tools, and they have wildly different price tags once you read the fine print. So let me walk you through what each channel really costs, who keeps the customer, and how I'd run more than one at the same time without losing my mind.

The story the fees tell

Start with the headline number, because it shapes everything else. In 2025 Amazon is on track to capture about 40.4% of US retail ecommerce sales, roughly $491.65 billion, per eMarketer.[1] That's not a channel you ignore. When Amazon and Shopify-powered stores are added together, they account for close to half of US ecommerce, according to Marketplace Pulse.[2] So the market has basically split into "rent reach from a giant" or "build your own and keep the margin." Here's where it gets concrete. Amazon's referral fees run from 5% to 45% depending on category, and they froze those percentages for both 2025 and 2026.[3] Home and Kitchen sits at 15%, Clothing at 17%, Electronics at 8%. eBay's standard final value fee is around 13.25% plus a per-order fee of $0.30 to $0.40, though categories range from roughly 2.5% to 15%, and most categories nudged up by as much as 0.35% in February 2025.[4] TikTok Shop is the cheap newcomer on paper: a 6% referral fee per order in the US (5% for some jewelry), with a 3% promo rate for new sellers in their first 30 days.[5] That 6% looks like a bargain until you remember TikTok Shop is a content-commerce machine. You'll likely pay creators affiliate commissions on top, and ads there need a $50/day minimum budget. The platform fee is low; the cost of actually getting seen is not.

Chart: Typical fee each channel takes per sale
Chart by author. Sources: Amazon referral fee schedule (Seller Central), eBay 2025 final value fees, TikTok Shop US referral fee. DTC figure reflects a typical ~2.9% payment-processing rate, not total cost. One caveat I want to be honest about: that DTC bar showing roughly 2.9% is only the payment processor. It does not include your platform subscription, apps, or the cost of driving traffic, which is the part marketplaces hide inside their fee. On your own store, customer acquisition is your line item to own. On a marketplace, it's baked in but you don't control it.

What you're really buying with each channel

Channel Reach Typical fee Who owns the customer Data you get Who it fits
Amazon Huge. ~40% of US ecommerce demand in one place 5%–45% referral; 8%–17% common Amazon. You rent the buyer Aggregate only; no buyer email/contact High-intent search products, commodities, replenishables
eBay Broad, deal/collectible-skewed ~13.25% + \$0.30–\$0.40/order eBay, but buyer info is less locked down More order/buyer visibility than Amazon Used, refurbished, niche, rare, parts
TikTok Shop Discovery-driven; can spike fast 6% referral (3% new-seller promo) + creator/ad costs TikTok. Algorithm owns demand Limited; content/engagement signals Impulse, trend-led, visual, sub-\$50 products
Your DTC store Whatever you build. Starts at zero ~2.9% payment + your acquisition cost You. Full relationship Everything: email, behavior, LTV Brands, repeat purchase, subscriptions, higher AOV

Warehouse with workers, pallets and boxes
Fulfillment is the unglamorous engine behind every channel. Photo: Tiger Lily via Pexels (Pexels License).

How I'd actually evaluate it

Forget loyalty to any platform. Ask three questions. Does my product get searched for or discovered? If people already type your product into a search bar, Amazon and eBay win because the intent is sitting there. If your product needs to be shown to be wanted, TikTok Shop's discovery engine is a cheat code. A $24 gadget that demos well in 15 seconds belongs on TikTok before it belongs anywhere. What's my margin after the worst-case fee? Run the math on a real SKU. A $30 Home and Kitchen item on Amazon loses $4.50 to the referral fee before fulfillment. If your gross margin can't absorb 15% to 20% plus ad costs and still leave profit, the marketplace isn't a growth channel, it's a slow leak. How badly do I need the customer back? This is the one people undervalue. If you sell something bought once a decade, owning the customer matters less. If you sell coffee, supplements, skincare, anything with repeat purchase, every marketplace sale is a relationship you rented and handed back. On your own store that same buyer becomes an email subscriber, a winback campaign, a lifetime value number you can actually grow. The honest answer for most operators is not either-or. It's marketplaces for reach and validation, your own store for margin and retention. Use Amazon or TikTok to prove demand and acquire first-time buyers cheaply, then work relentlessly to migrate repeat purchases to a channel you own. Insert cards, offer a better price or bundle on your site, run retargeting against the traffic you can track. Running both does add operational weight: separate listings, inventory sync, channel-specific content, more places for things to break. This is exactly where an all-in-one backend earns its keep, keeping catalog, stock and orders in one place instead of five tabs. If that's the wall you're hitting, Sellarix is built for running storefront and channels together. Use it or don't, the principle stands: your tooling should make multi-channel boring, not heroic.

The takeaway

Marketplaces sell you reach and rent you the customer. Your own store sells you margin and keeps the customer, but you pay for every visitor. The smart play is to treat marketplaces as the top of your funnel and your store as the place relationships go to compound. Validate on Amazon, discover on TikTok, defend your margin on your own turf. So here's my question for you: of your last hundred orders, how many of those customers could you email tomorrow? If the answer is "none," you don't have a sales problem. You have an ownership problem.

Sources

Replace six tools with one

Join the waitlist to be first on the platform, or book a demo.

Join the waitlistBook a demo