606: Kevin King On The Future Of Ecommerce: Amazon vs DTC In 2025

Podcast: Download (Duration: 53:21 — 61.3MB)

The Amazon-only gold rush is over, and the winners in 2026 will run a three-legged stool of marketplaces, social commerce, and agentic AI. On this episode of the My Wife Quit Her Job podcast, I sat down with Kevin King, a veteran Amazon seller and the host of the Marketing Misfits and AM/PM podcasts, to get his read on where ecommerce is actually headed. Kevin runs the Billion Dollar Seller Summit and Billion Dollar Sellers newsletter, and he still runs a real Amazon business himself.

His bottom line is blunt. Amazon is now a real business that demands operators, DTC has to be part of any serious brand plan, TikTok Shop is real but tiny next to Amazon, and agentic AI is about to open a gap between sellers who use it well and those who do not. Kevin predicts that some sellers doing $5M a year today will be doing $500K in two years, purely because a competitor started using AI correctly.

Below is the full playbook we walked through: how much cash you actually need to start on Amazon today, whether DTC or Amazon should be first, the categories winning on TikTok Shop, and the specific agentic-AI use cases already producing real returns for advanced sellers.

Get My Free Mini Course On How To Start A Successful Ecommerce Store

If you are interested in starting an ecommerce business, I put together a comprehensive package of resources that will help you launch your own online store from complete scratch. Be sure to grab it before you leave!

Click Here To Download

Key takeaways

• The Amazon gold rush is over. Selling on Amazon in 2026 is a real business that requires logistics, sourcing, inventory, finance, and marketing skills, not a side hustle with a phone dinging on the beach.
• Kevin’s cash rule of thumb is 2.5x your initial landed cost. If you have $10,000 to invest, find a product you can land for $4,000 or less, so you can fund the next inventory order before the first one pays out.
• Amazon should be a launch platform and a shopping-cart-of-choice, not a whole business. Build a real DTC brand alongside it, because you cannot get the customer data or the repeat-purchase experience on Amazon.
• TikTok Shop’s total GMV was around $20-50 billion last year depending on the source, versus Amazon’s roughly $700 billion (1P + 3P combined). It is real, but it is Amazon’s weekend, not a full replacement.
• Tmoo (Temu) is aggressively recruiting US sellers with a “no fees” pitch, but Temu sets the retail price and Amazon sellers will lose their buy box if they list there. Very few Western sellers are winning on Temu today.
• Agentic AI (autonomous AI agents that talk to each other via protocols like MCP) is the single biggest shift ahead. Advanced sellers like Trypoll are already stringing PPC agents, sourcing agents, and creative agents together for tasks that used to take 20 to 30 people a month.
• Copy Coders, a $10,000 custom LLM built on top of Claude by Genesis, has doubled and tripled Kevin’s email conversion rates. The prompt behind it is 19,000 words long, which is the honest reason “just use ChatGPT” is not the same thing.
• Product discovery is moving from Google and Amazon search to LLM chats. Andrew Jassy has already told Amazon internally that generative AI will change how people discover products.

Is Amazon still worth it for new sellers in 2026?

Amazon is still worth it in 2026, but only as one leg of a real branded business, and only for founders who treat it like a real business from day one. The days of slapping a made-up brand on an Alibaba spatula, coasting on organic reviews, and listening to your phone ding at the beach are done.

Kevin has been selling on Amazon since it launched a marketplace, and he is emphatic that the mental model has to change. New sellers need to be comfortable wearing every hat: sourcing, logistics, inventory finance, PPC, listing optimization, customer service, and brand building.

The people who cashed out through aggregator acquisitions in 2020-2021 were right place, right time. Most of the aggregators failed. Most of the founders who tried to launch a second brand after their exit have struggled, because the skill set they had was riding a rising tide rather than operating a mature marketplace.

How much money do you actually need to start selling on Amazon?

The rule Kevin still teaches inside the Freedom Ticket course is that you need at least 2.5 times your initial landed cost in working capital. If your entire budget is $10,000, you should be sourcing a product that lands (landed cost is unit cost plus freight, duties, and inspection) for $4,000 or less on the first order.

The reason for the multiple is cash flow, not conservatism. You need money in the bank to place the next purchase order before Amazon pays you for the first one, or you run out of stock and lose organic rank. Running out is worse than starting slow.

The number you actually need also depends on where you live and what “worth it” means to you. A $1,000-a-month side income can change a family’s life in Pakistan, the Philippines, or parts of Latin America. In the US, it is barely a side hustle, so US sellers need to plan for larger inventory buys, larger PPC budgets, and thinner margins as Amazon fee hikes keep compressing the middle.

Should you start on Amazon or DTC first?

Amazon is still the fastest way to prove that a product sells, because the traffic is already there and buyers show up with credit cards out. Kevin’s metaphor is that Amazon is a freeway of buyers driving by at 60 miles an hour with the window down, and your only job is to stand in the right spot and swipe as they pass.

DTC (a Shopify store or similar) is the opposite. You have to drive every visitor yourself through meta ads, TikTok, cold email, an existing list, or SEO. That is a completely different skill set, and it is why so many Amazon sellers fail when they open a Shopify site and think putting up the store is the work.

If you already know how to drive traffic (or you have a partner who does), Kevin actually recommends starting DTC first in 2026, then letting Amazon be the overflow for buyers who need the Prime trust badge. If you do not know how to drive traffic yet, start on Amazon, maximize it (add Canada, then Europe), and use the profits to fund the DTC build.

How Amazon and DTC compare for ecommerce sellers today

The two channels are genuinely different animals, and picking one because you already know the other is the most common mistake Kevin sees. Here is how the tradeoffs stack up in 2026.

The strategic answer Kevin keeps coming back to is that serious brands run both. Use Amazon to catch the buyers who will not trust a new domain. Use DTC to build a customer list you actually own, and to sell upsells, subscriptions, and repeat orders that Amazon will not let you offer.

Which strategies are working best on TikTok Shop right now?

TikTok Shop is the shiny object of 2025-2026, and it is real, but the ceilings are lower than most sellers assume. TikTok Shop’s total GMV last year ran between roughly $20 and $50 billion depending on the source you trust, versus Amazon’s approximately $700 billion combined 1P and 3P. That is meaningful, but Kevin’s framing is that TikTok Shop is Amazon’s weekend, not a full replacement.

The categories that actually work on TikTok Shop are impulse-buy, sub-$100, entertainment-adjacent products. Beauty, supplements, small electronics gadgets, pet products, apparel, and art all show up in the success stories. Kevin cited a founder doing $21 million a year selling women’s pants on TikTok Shop, and another doing about $6 million a year selling artwork.

The reason TikTok Shop converts is the seamless in-feed checkout, where the video keeps playing while you check out and there is almost no interruption. That is a purchasing environment purpose-built for impulse, so if your product needs more than a few seconds of consideration or costs more than $100, it will struggle there.

TikTok Shop is also going through real turbulence. They cut back the free-shipping and promo-discount subsidies that fueled early growth, executed a large US layoff, and replaced US executives with China-based leadership. Add the ongoing US ban uncertainty, and this is a channel to test, not a channel to bet your company on.

Is Temu worth selling on for US ecommerce sellers?

Temu (Tmoo) is aggressively recruiting US sellers right now, but the terms are hostile enough that almost no Western sellers are winning there. Temu’s pitch is “no fees,” which is technically true, and misleading in practice.

The catch is that Temu sets your retail price. You send in a product with a wholesale cost you can accept, they mark it up and list it however they want, and if that price is lower than your Amazon price, Amazon will strip your buy box under its fair-pricing policy. Kevin heard the pitch in person during Amazon Accelerate in Seattle and walked out convinced no serious Amazon seller can play.

Temu has also lost roughly half its US traffic and 60-70% of its US sales since the de minimis rule change, and Shein is fighting a similar battle. Unless you are a Chinese-based seller or a first-time seller with nothing to lose on Amazon, Temu is a hard pass in 2026.

What is agentic AI and how are top ecommerce sellers using it?

Agentic AI is a step up from a chatbot. It is a system of specialized AI agents that each own one job (PPC, sourcing, listing creation, image generation, keyword research), and that talk to each other through a common protocol like MCP (Model Context Protocol) to actually execute tasks with minimal human supervision.

Kevin’s mental model is that instead of hiring the top PPC person from a top agency like Clear Ads, you clone that person’s knowledge into an agent. Then you clone your best sourcing person into another agent, your best listing writer into a third, and so on. The human stays in the loop for judgment calls, but a stack of agents can now execute 24/7/365 what used to take 20 to 30 people a month.

This is not science fiction. Trypoll (an ecommerce platform) is already running an agentic stack that handles roughly 70% of the coordination Kevin described. The prediction Kevin is willing to put a number on: some Amazon sellers doing $5 million a year today will be doing $500,000 a year in two years, because a competitor in their category started using agentic AI correctly and pulled away.

The AI use cases already working for average sellers

Most sellers are still using AI at the “rewrite this email” level, which is kindergarten. The next step up is the one that pays back this quarter: video ad creation with tools like Veo 3, positioning-angle testing, avatar generation for UGC ads, customer-service chatbots, and negative-review remediation workflows.

Copy Coders is the specific tool Kevin uses that has doubled and tripled his email conversion rates on the Billion Dollar Sellers newsletter. It is a custom LLM built on Claude by Genesis, priced around $10,000 a year (sometimes sold at $5,000). What makes it different from “just use ChatGPT” is the base prompt, which Kevin says is 19,000 words long and trained on Ogilvy and the classic copywriters going back 100 years.

Custom LLMs on your own content

The other pattern to watch is founders building custom LLMs on top of their own body of work: podcast episodes, courses, newsletters, webinars, and event content. Someone can then query “what does Kevin King think about the Amazon buy box?” and get the actual answer from the actual person, without the internet hallucinations that a generic LLM adds.

Kevin has one of these launching in two weeks, in both a free scaled-back version and a full-content version. I tried building one on my own content roughly a year ago and it was not ready. The models have improved enough that this is worth revisiting in late 2026.

How AI is changing how people discover ecommerce products

Product discovery is quietly moving from Google and Amazon search into LLM chats. Kevin cited a story from the CEO of Pattern, a $1.8 billion Amazon seller with 16 data scientists on staff, whose youngest daughter did not accept the family’s usual first-car recommendations. Instead she typed into ChatGPT that she was a 16-year-old girl who liked X, Y, and Z, got a specific Jeep recommendation, and trusted the AI over her parents.

That is the pattern that is scaling. Buyers are doing their research in ChatGPT, Gemini, Perplexity, and Claude, then going directly to Amazon or the brand’s site to complete the purchase. Traffic from open Google search to product review sites has been declining steadily for exactly this reason.

Amazon CEO Andy Jassy already sent an internal memo warning that generative AI will change how customers discover products, and told his teams to embrace it or get left behind. For sellers, the takeaway is that being cited by AI (through solid product content, third-party reviews, and structured data) is becoming as important as ranking in Amazon search or Google.

What Kevin King recommends ecommerce sellers do next

Kevin’s practical to-do list for any seller who wants to still be in business in 2028 is short and specific. Start dabbling in AI now, even if you are not technical, because the logic of how to prompt is a learnable skill. Photographers and creative pros who have learned to prompt image models are thriving; the ones who did not are being priced out.

Build a real brand, not just a product SKU. Get into a category with a repeat-purchase mechanic (consumables, apparel, personalized, subscriptions) instead of one-time buys like a ladder someone buys every 20 years. If you already sell on Amazon, expand to Canada and then Europe before you add a second US channel.

Layer marketplaces, social commerce, and AI. Use TikTok Shop for top-of-funnel discovery on impulse-friendly products, use Amazon as the trusted shopping cart, and use DTC as the brand home where you own the customer. Then wire agentic AI into the parts of your workflow that repeat every week.

Frequently asked questions

Is Amazon worth selling on in 2026?

Yes, but only as one channel in a real branded business, and only if you treat it like a real business. New sellers need working capital of at least 2.5x their landed cost, operator-grade skills across sourcing, logistics, and PPC, and a plan to expand into DTC and other marketplaces. The gold rush is over, so treating Amazon as a passive side hustle no longer works.

Should new ecommerce sellers start on Amazon or Shopify first?

It depends on whether you already know how to drive traffic. If you do, start on Shopify (DTC) so you own the customer list and the brand from day one, and add Amazon later as an overflow channel. If you do not, start on Amazon to prove the product on borrowed traffic, then build DTC with the profits.

How much money do I need to start selling on Amazon?

Kevin King’s rule is 2.5 times your first landed inventory cost as total working capital. On a $10,000 budget that means a product that lands for under $4,000 on the first order. US sellers should expect to need more than sellers in lower-cost countries, because $1,000-per-month profit is life-changing income in some markets and barely a side hustle in others.

Is TikTok Shop a good channel for ecommerce brands right now?

TikTok Shop is a real channel for impulse-friendly, sub-$100 categories like beauty, supplements, small gadgets, apparel, and pet products. Its total GMV is still a fraction of Amazon’s (roughly $20-50 billion vs Amazon’s ~$700 billion), and the platform is going through executive turnover and subsidy cuts. Test it for top-of-funnel discovery, do not bet the company on it.

Is Temu worth selling on for US sellers?

Not for most US sellers. Temu’s “no fees” pitch is undercut by the fact that Temu sets the retail price, which will strip your Amazon buy box if the Temu price is lower. Temu has also lost roughly half its US traffic and 60-70% of its US sales since the de minimis rule change, so the platform itself is shrinking.

What is agentic AI for ecommerce in plain English?

Agentic AI is a stack of specialized AI agents that each own one job (PPC, sourcing, listings, images) and coordinate with each other through a common protocol like MCP to execute tasks with limited human supervision. For ecommerce, it means a small team can now do the coordination work of 20 to 30 employees, running 24/7/365 across data volumes no human can process.

How can an average ecommerce seller use AI today without being technical?

The highest-return uses are video ad creation with tools like Veo 3, positioning-angle testing, avatar generation for UGC content, chatbot-driven customer service, and email copy optimization. For serious email revenue, purpose-built tools like Copy Coders (a Claude-based custom LLM by Genesis) can double or triple conversion rates on newsletter and promotional sends.

Where is ecommerce product discovery actually happening in 2026?

Increasingly inside LLM chats like ChatGPT, Gemini, Perplexity, and Claude, then routed to Amazon or the brand’s site for checkout. Google organic traffic to review and comparison sites has been declining for two years, and Amazon’s own CEO has internally warned that generative AI will reshape how buyers discover products. Being cited by AI is becoming as important as ranking in classic search.

Ready To Get Serious About Starting An Online Business?

If you are really considering starting your own online business, then you have to check out my free mini course on How To Create A Niche Online Store In 5 Easy Steps.

In this 6 day mini course, I reveal the steps that my wife and I took to earn 100 thousand dollars in the span of just a year. Best of all, it's absolutely free!