Here is a fact that should unsettle every brand selling on Amazon in the US: of the ten largest marketplaces in the world, six are Asian. Amazon sits at the top with $845 billion in global GMV but Pinduoduo is growing at 27.6% annually, Douyin at 56.7%, Kuaishou at 31.4%. Amazon's own growth? 8.8%.
From inside the US, this is invisible. Amazon is e-commerce. It holds the wallet, the fulfillment, the reviews, the Prime habit. The default for any seller remains what it has been for fifteen years: build on Amazon, optimize for Amazon, live and die by the Buy Box.
That default is starting to cost sellers something real, and most of them don't see it.
Amazon is the finest intent-capture machine ever built. A customer who types "wireless earbuds under $50" is already 90% of the way to a purchase. Everything about Amazon from the algorithm, to the A9 black box is engineered for that last mile. What Amazon has never figured out is the other mile: the one where the customer doesn't yet know they want earbuds. That mile happens on TikTok, in Discord servers, in group chats and increasingly inside an AI chat window. Douyin's 56.7% CAGR is a discovery story. The social layer isn't a missing feature for Amazon but a missing metabolism. You can't bolt discovery onto a system optimized for conversion.
Look at the US band below Amazon. Temu: $22B. TikTok Shop: $15B. Walmart 3P: $15B. Three wildly different business models all landing within $7B of each other. That's the market pricing the "alternative to Amazon" ceiling and it sits around $20B. Amazon at $300B in US third-party GMV is not under existential threat.
But Rufus is. Andy Jassy claimed Amazon's AI shopping assistant drove $12B in incremental sales in 2025. Half of that in the US would make Rufus alone a top-seven US marketplace. The fact is that, for Amazon, the incremental growth is actually coming from discovery surfaces because intent has run out of room.
Here is the reframe: being an "Amazon seller" is becoming a structural liability because Amazon is shrinking, but because Amazon is becoming infrastructure rather than identity. The sellers who win the next five years will treat Amazon the way DTC brands treated Shopify a decade ago – as a checkout layer. Their discovery happens somewhere else; Amazon is simply where the transaction closes. The threat isn't that TikTok Shop displaces Amazon. It's that brands who own their demand on other surfaces capture the margin that used to flow to sellers who were just good at Amazon SEO.
So the question for every brand in the US right now is not "how do I rank higher?" It is: where does my demand come from, and who owns it?
If the answer is "Amazon's algorithm," you are not a brand. You are a listing. And listings, historically, do not survive the transition from platform to protocol.
Lumian can help you get discovered on Rufus. Book a Free Consultation.
In this week’s issue:
Marketplace: CC Payments, Fuel Surcharges, Drone Delivery
Seller Central Update of the Week: Product Opportunity Explorer
Tweet Spotlight: Limiting Beliefs
Marketplace Madness
Quick recap: Amazon had quietly told a group of sellers it would stop letting them pay for ads with a credit card starting April 15, and instead deduct ad costs straight from their retail proceeds, killing the 30 to 60-day float and card rewards sellers had been banking for years. After heavy pushback, Amazon has now pushed that change to August 1, confirmed it only affects a small group of directly contacted advertisers and added Pay by Invoice (Net 30) as an alternative that roughly mirrors the old float. They're also offering $2,500 a month in click credits for five months to cushion the transition.
Why it matters:
If you got the email, Pay by Invoice is probably the move as it preserves most of your working capital window, even if you lose the cash back. Just know you have to actively opt in through the Billing section before August 1 or you'll get defaulted into proceeds deduction. Also worth flagging: operators are warning the free credits could temporarily inflate CPCs in August and September as advertisers bid more aggressively with house money, so don't anchor your Q4 bid strategy to auction data from that window.
Starting April 17, Amazon is tacking a 3.5% fuel and logistics surcharge onto fulfillment fees for FBA sellers in the U.S. and Canada. The company blames rising oil and logistics costs tied to the Iran war. Amazon says the hit averages about 17 cents per unit depending on size and weight and it's calculated on your fulfillment fees rather than your sale price, so it scales with how much space and handling your product takes up.
Why it matters:
17 cents sounds small until you multiply it across thousands of units a month, and on low-ASP or bulky items it can quietly eat a real chunk of your contribution margin. This is the moment to pull your FBA fee preview for your top SKUs, rerun your unit economics with the surcharge baked in, and decide where you can afford to hold price versus where you need to nudge up a dollar to protect margin.
Andy Jassy's shareholder letter basically declared war on waiting. Prime Air drones will cover 30 million customers by year-end and Amazon Now is testing sub-20-minute delivery. Same-day has already moved 500 million packages in 2026. Skeptics argue 90% of shoppers don't actually need it and the margins are brutal but Amazon isn't quite trying to be profitable at speed but to make everyone else unprofitable at it.
Why it matters:
If your SKU lives in Amazon’s top 90,000 SKU catalog that qualifies, you just got a free conversion lever as sub-hour delivery badges move the needle on impulse buys in a way price rarely does. If it doesn't, this is your cue to fight for it, because the gap between "Prime fast" and "everything else" is about to become the new buy-box.
Lumian offers AI-powered PPC + DSP management starting at just $1,500/month
Seller Central Update of the Week
Amazon rolled out a new beta feature inside Product Opportunity Explorer called "Discover Unmet Demand," and it's aimed squarely at the whitespace hunt every seller does manually. The tool surfaces keyword clusters where shopper demand is high but conversion rates are low, meaning lots of people are searching and clicking but not actually buying — a signal that existing listings aren't doing the job. Benchmarks are calculated dynamically by category and price band, and sellers get the underlying data on search volume, impressions, clicks, and purchases at the keyword level.
The signal here is worth paying attention to. Amazon is continuing to pull product research workflows that used to live in Helium 10, Jungle Scout, and Brand Analytics directly into Seller Central, and "unmet demand" is the kind of insight most sellers were paying third-party tools to approximate. If the beta holds up, it becomes a faster way to spot product gaps, validate variations, or sharpen listings on keywords where the category is underperforming. Worth poking around in Opportunity Explorer this week to see if your catalog is surfaced.
Tweet Spotlight
Meme Therapy
How did you like this week's newsletter?
Best,