Shopify vs Amazon vs Walmart

For most brands past a certain size, both — and the useful version of the question is which SKUs go where. Your own store carries the highest margin per order and gives you the customer relationship, the email address, and the second purchase, but you pay for every visit. Marketplaces bring demand you didn't have to buy, at the cost of a referral fee on each sale, a customer who belongs to the platform, and fulfillment economics that some products simply cannot survive. If your product has real brand demand and a repeat-purchase motion, weight toward the own store. If it has strong category demand and little brand demand, marketplaces will usually reach buyers your own store would have to pay full price to acquire.

Both charge a referral fee on each sale, set by category, and both publish their own fee schedules — those schedules are the authority for your specific category, and they change, so quoting a percentage here would age badly and mislead you. What matters more than the headline percentage is everything stacked on top of it: fulfillment and storage fees if you use the platform's fulfillment, return processing, and advertising, which on a competitive listing is closer to mandatory than optional. The number to calculate is contribution margin per unit after all of that, per SKU. Plenty of products look fine at the referral-fee line and go negative by the time returns and ads are counted.

On Amazon and Walmart, multiple sellers can list against the same product. The Buy Box — Amazon calls it the Featured Offer — is the default add-to-cart on that shared listing, and the overwhelming majority of orders go to whoever holds it. Price, fulfillment method and speed, in-stock rate, and seller performance metrics all feed the decision. The reason it matters more than it sounds is the failure mode: when you lose it, nothing visibly breaks. Your listing is still there, your ads still spend, and your sales quietly stop. On your own store there is no Buy Box, because you are the only seller on your product page.

No, and trying is one of the more common ways brands break listings. Your store catalog can be the source of truth, but each marketplace has its own required attributes, identifier rules, and content standards, and each one enforces them differently. A feed mapped for Amazon pushed at Walmart tends to produce rejected or suppressed listings rather than a clear error. The workable pattern is one canonical catalog plus a per-channel mapping layer that stays in sync — which is exactly the work that gets skipped at launch and then blamed on the marketplace six months later.

Sometimes, and whether that's a problem depends on margin rather than on principle. Some marketplace orders are genuinely incremental — buyers who would never have found or trusted your site. Some are customers who would have bought from you directly and just cost you the referral fee and the customer relationship. The way to tell is to look at whether own-store revenue actually declined when the marketplace launched, and to compare contribution margin per unit on each channel rather than comparing channel revenue. Brands that price and merchandise identically everywhere tend to see more cannibalization than brands that differentiate bundles, sizes, or subscription offers between channels.

Amazon first in most categories, simply because the demand pool is larger and the onboarding door is easier — Walmart runs an application and approval process that screens on business history and fulfillment capability. Walmart is the better second move once one marketplace is already running well, and it's occasionally the better first move in categories where its shopper base over-indexes and seller competition is thinner. What rarely works is opening both at once: two catalogs, two attribute schemas, two fulfillment programs, and two sets of performance metrics is more operational load than most teams can absorb while also running the store they already have.

Contribution margin by channel and by SKU, calculated after referral fees, fulfillment, storage, returns, and advertising — not revenue, and not the ROAS figure on the platform dashboard. ROAS can look identical on two products with completely different real profit once fulfillment and return rates are accounted for, which is why a channel can post healthy platform numbers while the bank account doesn't move. Buy Box win rate belongs in the same report for marketplace channels, because a listing that lost the Buy Box will show declining sales with no obvious cause anywhere else in the data.

2026 US market research puts a meaningful store or channel build at roughly $5,000-25,000 for a basic-to-mid-market implementation, with complex commerce reaching $30,000-100,000+, and ongoing management commonly $2,000-10,000 per month. That is market context, not ULEY's price. ULEY's rate card is four options and nothing else: a $2,500 one-time Audit Sprint, a $3,500 fixed-scope Automation Build, a $3,500/month Growth Retainer, and $175/hour for scoped work that fits none of them. Channel builds are scoped as a fixed-price Automation Build agreed before work starts, so the number is settled up front rather than discovered halfway through.