Toqeer Khalil reminds us on Linkedin that the number of sellers in the Amazon Marketplace is shrinking.

Active Amazon sellers dropped from 2.4 million in 2021 to 1.65 million by the end of 2025.

New seller registrations just hit a decade low. Down 73% from the 2021 peak.

Sounds like bad news. It's not.

Look at what's happening underneath:

⮞ Third-party sellers now move 62% of all units sold. An all-time high.
⮞ Over 100,000 sellers clear $1M+ a year. Nearly double 2021.
⮞ Traffic per active seller is up 31%.

Fewer players. More revenue each. Higher stakes.

This isn't a shrinking opportunity. It's a consolidating one.

The brands winning are pouring money into operations, content, and ad infrastructure. The ones losing are still running their 2020 playbook.

Stop burning your time on manual tasks when AI can run your marketing and backend.

In this episode of Marketing Misfits, Norm Farrar and Kevin King chat with Andrew Erickson, a former Qualcomm engineer who built and sold two physical product brands (including a wooden map company) for seven figures.

From hand-writing postcards for early Etsy reviews to launching highly advanced AI-driven e-commerce software, Andrew breaks down exactly how he uses Model Context Protocol (MCP), autonomous coding agents, and "Logic Lead Magnets" to scale businesses.

Whether you are trying to capture more emails with simple custom apps or want to learn how to securely host virtual AI workers, this episode is a masterclass in modern e-commerce automation.

BTW, Andrew will be doing an AI workshop at BDSS Market Masters 4 next month in Austin.

🔥 Join the Marketing Misfits Newsletter

Two days in and the numbers are nuts.

Day 1 of the four-day event did $8.3 billion in U.S. online sales across all retailers. That's up 5.3% over last year and the biggest e-commerce day of 2026 so far.

It blew past Adobe's $7.9 billion projection. Adobe is holding its full-event forecast at $26.3 billion, which would be a 9% jump over last year. For context, those numbers come from tracking around 1 trillion site visits across 18 categories and 100 million SKUs. This is a big sample, not a guess.

What shoppers are doing

Average order size is sitting at $48.89. And people aren't doing one-and-done. Nearly half of households (49%) placed two or more separate orders.

Translation: buyers are coming back multiple times. If your listing converted once, you had real shots at the repeat add-to-cart.

Where the money went

Day 1 demand clustered in electronics, appliances, tools, and home and garden. Electronics alone more than doubled versus a normal June day. Up 105%.

Everyday essentials climbed too, and some categories went vertical:

That essentials bump is no accident. Amazon leaned hard into grocery and household this year. Trash bags and dishwasher pods are quietly eating share from the pricier discretionary stuff as shoppers feel the squeeze.

The discount window

Here's a number worth saving. Price cuts on Day 1 ran 10% to 24%, and Adobe expects the rest of the event to stay in that same band.

So if you're benchmarking your own promo depth, that's the zone that moved volume this year. You did not need to torch your margin to compete.

Why June this year

Amazon pulled Prime Day out of July for the first time in five years. It now runs June 23 through 26.

The reason: dodging the FIFA World Cup and July 4th. Prime VP Jamil Ghani said both events drove the shift. Amazon is even positioning World Cup watch parties and holiday gatherings as demand drivers for food and household items.

The Amazon Effect is real

When Amazon moved the date, the whole retail calendar moved with it. Walmart and Target ran parallel events to ride the wave and pull extra shoppers online.

That's the lesson. Amazon sets the tempo and everyone else dances to it.

And the price of admission to all of it: Prime is still $139 a year and nearly 75% of all US adults have it.

Two American companies invented entire product categories from scratch.

Both had huge brands. Both are now getting run over by Chinese rivals anyway. The lesson for sellers is simple. A strong brand is not a moat. If you don't have one, Chinese sellers will eventually put you out of business.

Quick story. Someone moves to Taiwan, takes up scuba diving, and goes shopping for an underwater camera. Simple search. It turned into a lesson on where consumer tech is actually heading.

There was a time the answer was obvious. Buy a GoPro.

GoPro invented the action-camera category back in 2002. As recently as 2022 it still owned an estimated 75% of the global market. Then it flipped. Fast.

In early June, GoPro told regulators in a filing there is "substantial doubt" about its ability to keep operating. Global share down to roughly 18% in three years.

Ask any diver, any travel creator, any outdoor person what to buy now. You get one of two answers. DJI or Insta360. Both Chinese. Together they own over 80% of the action-camera market.

What did GoPro blame in the filing? Rising costs from the global memory chip shortage.

What it left out: years of falling sales and brutal pricing pressure from DJI and Insta360 that left zero room to absorb a shock.

The pattern is familiar. GoPro peaked in 2015, then got squeezed by smartphones and a string of failed product bets. DJI entered in 2019 and kicked off a price war. Insta360 showed up in 2022. Both shipped new models faster and stacked on AI editing features

GoPro could not match.

Yes, GoPro made its own mistakes. Analysts call them self-inflicted wounds. But manufacturing scale plus rapid product cycles made a comeback close to impossible.

Now look at robot vacuums. Same movie.

iRobot invented the home cleaning robot and named it Roomba. Chinese brands Dreame and Roborock took over. iRobot filed for bankruptcy in December and got bought by its own Chinese manufacturer, Picea Robotics.

Here's the seller takeaway

This is the new default. A Western brand creates a category. Chinese competitors out-ship it, out-price it, and bolt on AI faster. Action cameras. Robot vacuums. EVs. As AI gets baked into more hardware, the gap could widen.

There is one crack in the story. Trust and data.

Robot vacuums run on cameras, sensors, microphones, and detailed maps of your house. That is now a security conversation. The same concern already put limits on Chinese drones and EVs in the US. GoPro's own escape plan tells you everything. It is exploring a pivot into defense and aerospace.

So what's the play if you sell physical products?

Here's the hard lesson. If you don't have a moat, Chinese sellers will eventually put you out of business. Not maybe. Eventually. They out-ship you, out-price you, and copy your best features in a single product cycle.

A strong brand alone does not save you. GoPro had one of the most recognizable brands on the planet. Roomba was so dominant the word became a verb. Both still got run over.

A real moat is the stuff scale cannot copy overnight and a factory in Shenzhen cannot just clone. Patents and defensible IP. Data privacy and country-of-origin trust where it actually matters to the buyer.

A community that buys from you because of who you are. Distribution or relationships a competitor cannot buy their way into. Service and an experience that travels with the customer, not the SKU.

If your only edge is "we have a nice logo and good reviews," you are renting your position. The minute a Chinese competitor decides your category is worth taking, the rent comes due.

GoPro is now betting its survival on people who care about security, not cost-conscious AI households. That is them scrambling to build a moat after the fact.

Build yours before you need it. Worth asking, honestly, which side of that line your products sit on right now.

✌🏼 Have a great weekend.

See you again on Monday.

The answer to today’s STUMP BEZOS is
Spending on influencer marketing is expected to hit $12.42 billion this year.