He sold 3,000. They sold 100,000. It wasn't PPC - 80/20 of Ecom Sep 4, 2026

One member of our AI Ecom Builders Cohort sells a learning tool. He had a strong August: around 3,000 units.

Then he looked at the category leader. About 100,000 units.

He went to ChatGPT convinced the number had to be wrong.

It wasn't.

The leader isn't winning on PPC. Not on price. Not on listing quality.

They run webinars for teachers. They run discount programs for schools.

So a teacher tells the class which learning tool to buy. The parent shows up on Amazon already decided.

By the time that customer types anything into the search bar, the sale is already over.

We spend all day optimizing the last three feet of the sale and almost nothing on where the decision actually happens.

His next move is small, and I like it: get the product plus a simple manual into after-school math clubs. One state first. Prove it works, then scale.

TAKEAWAY: If you can't explain the gap between you and the category leader using an on-Amazon lever, stop pulling on-Amazon levers. Find the room where your buyer gets told what to buy. It might be a classroom, a forum, a coach, or a trade group. It is rarely the search results page.

The same cohort member found another problem.

He didn't go viral. He didn't get a press hit. He just ranked better than last year.

That was enough to break his forecast.

His forecast was built on last year's sales. But last year's sales were low because he was out of stock. He forecasted against constrained demand and called it demand.

So the safety stock at his 3PL disappeared faster than planned. His reorder is landing about 10 days late. He had to turn PPC down in his peak month while it was still profitable.

Think about that. He throttled ads that were making money because he couldn't afford the sales.

Most sellers plan for the downside. Almost nobody plans for the upside.

TAKEAWAY: Check whether your forecast is built on what you sold or what you could have sold. If you stocked out last year, your history is understating real demand.

These are exactly the problems we will work through at the 7 Figure Seller Japan Mastermind:

Bring your laptop and a real ecommerce problem. You will leave with an Ecom AI Second Brain action plan and one working AI skill for your business.

Oct 13, 2026 and Oct 14, 2026 · Tokyo

Pre-summit ticket: US$599

See the 7 Figure Seller Japan Mastermind details

Amazon's base monthly storage rate for standard-size inventory rises from $0.78 to $2.40 per cubic foot for October through December. Oversize rises from $0.56 to $1.40.

Here is a simple example. Assume 3,000 standard-size units at 0.05 cubic feet each. That is 150 cubic feet of inventory.

That is base storage only. Your real bill can also include storage-utilization, aged-inventory, dangerous-goods, and other surcharges.

TAKEAWAY: Work out what should sell between now and Black Friday. Send that. Hold the rest at a 3PL if the economics are better.

Check Amazon's current FBA storage rates

Canada is applying counter-tariffs of 15%, 25%, and 50% to selected U.S.-origin products covering $27.6 billion in imports. Affected sectors include steel, dairy, appliances, agricultural equipment, pulp and paper, plastics, and electronics.

Goods already in transit when the measures begin are excluded, but you need to verify origin and the exact tariff item.

TAKEAWAY: Update the landed-cost sheet for every affected lane. Do not leave duty assumptions buried in an old template.

See Canada's official counter-tariff list

Trans-Pacific spot rates moved higher again as Q4 cargo, Asian port congestion, and carrier capacity controls collided.

Freightos data cited by FreightWaves put Asia–U.S. West Coast rates at $7,621 per 40-foot container and East Coast rates at $9,791, both up 2% in the latest week. Drewry's global index held at $4,465 per 40-foot container on Sep 3, 2026, with Trans-Pacific gains offset by Asia–Europe declines.

The next pressure point is China's Golden Week. Blanked sailings can tighten effective capacity even if your factory says the goods are ready.

TAKEAWAY: Get a fresh quote now for every Q4 lane. Compare West Coast, East Coast, and inland costs together. Do not budget from the rate you paid on the last shipment.

Read the FreightWaves Trans-Pacific update

Check Drewry's World Container Index

Grokbot is kind of like what OpenClaw was promising to be: an easy-to-use AI agent with almost zero setup. Some people call it the Apple of AI agents.

That's why I say it's like Apple, because it's easy setup. You can connect Gmail with the click of a button. You can spawn multiple agents and group chats very easily. But you're locked in. You have to use Grok, and you can't even choose the model.

In my limited test, it was good for persistent roles like chief of staff, help desk and social content. It plugged into the Gary Brain very quickly. But it wasn't as good for deep analysis, the $20 plan lasted me one day, and Stripe authorization failed several times.

For me, it's not a replacement for Codex or Claude as my day-to-day driver. But it is worth testing for one narrow role, like a social media agent or competitor analysis agent. Start with the $20 plan. Build one bot. Build the chief of staff first.

Read the official Grok Bot introduction

If your AI agent is still using your personal Gmail, there is a cleaner option. This gives an agent its own inbox, with sending, replies, attachments, and no Google OAuth tied to your personal account.

Open this week's surprise link

Talk soon,

Gary

P.S. If your sales history includes stockouts, your Q4 forecast may be too low. Bring that problem to Tokyo and we will work through the data and the workflow together.

See the 7 Figure Seller Japan Mastermind details