One Inch, $17 a Unit: Why COGS Is the Only Lever Left

I am writing this from a train station in China with half an hour before my train. No excuses — here is what last week's Prime Deal actually taught me.

Over the last four weeks we nearly doubled revenue. Margin grew even faster. On paper it is a great month, and if I stopped the story there you would probably clap.

But look at the ad spend. Four weeks ago we were spending about $14,000 a week. Now we are at roughly $20,000. That is three thousand dollars a day, every single day, on advertising.

So let me tell you what a Prime Deal really costs.

The Prime Deal math nobody puts on the slide

Prime Deals, Black Friday, Cyber Monday — they are all built the same way, and they are all tricky for the same reason: you pay upfront for the privilege of selling cheaper.

Five, six hundred dollars per ASIN. Cheaper in some marketplaces, more in others, but that is the ballpark. You hand Amazon that money, and in return you get to put 20–30% off on your own listing. That is the whole mechanism.

Then watch what happens downstream. In the US last week our average order value came in around $129. Before the deal, an order was worth about $140 to us. The discount is real and it is coming out of your pocket, on top of the fee you already paid.

And the traffic. A deal listing gets clicked constantly — by people who are browsing deals, not by people who came to buy your product. US ad spend went up roughly 20% during the event purely on click volume. Some of those clicks convert. Plenty of them just bill you.

None of this means don't run deals. It means: know exactly what you are buying, because the invoice arrives in three separate places.

You cannot raise the price anymore

Here is the part that changes everything about how I run this business.

Elasticity on our units is severe. I cannot raise the sale price without damaging ranking. It does not work like it used to. The market is oversaturated — there is too much stuff on Amazon, and honestly there is too much stuff in the economy generally. Somebody will always undercut you, and Amazon will always show them first.

> If you cannot move price, the only Jedi way left to improve efficiency is to lower your cost.

That is it. That is the whole strategy. Everything else is decoration.

Two levers that actually move

There are more than two, technically. But two of them carry the weight:

• Logistics. Negotiate properly with your freight forwarder. Ship more efficiently. Batch your paperwork and customs bureaucracy in bulk instead of drip-feeding it. Every bit of that lands on the per-unit prep cost.
• The purchase order itself. Go into your BOM. Actually look at what your product is made of and which component is eating the biggest share of it. Cheese grater, charging cable, toothbrush, projector screen — it does not matter what you sell. Your product has a bill of materials, that ugly enormous spreadsheet, and almost nobody opens it.

For us, the bottom half of our product is now fully owned. We produce those parts ourselves, on our own machines. Fully controllable, and a significant price reduction the moment it went live.

What your supplier actually wants (it is not what you think)

The beautiful part of digging into your BOM is what happens when you finally go and meet the people making your components.

Raw material suppliers are volcanoes. They have to erupt. They constantly need to push material out to somebody — that is the entire physics of their business. And when the economy is shrinking, that pressure gets worse, not better.

So when you walk in and say: guys, you are erupting my raw material, and I have product feedback, I have a market opportunity, I need a new mold, I want to test a different material — two things happen.

First, they are incredibly happy to work with you. Second, they do not raise your price. Often the opposite: they will ship you material upfront, before payment clears, because in the back of their head they have already done the calculation. This one knows his drill. He is investing in something new. Better to work with him for years than squeeze him today.

That is not a theory. That is what is happening around me right now, which is exactly why I am here and not at a desk.

The size tier almost everyone ignores

Now the part that costs sellers the most money while they are looking somewhere else: size tiers.

Sellers neglect tiers. They obsess over ACOS to the second decimal and never check which fulfillment bucket their box falls into. The jump between tiers is not a rounding error — the fees differ substantially.

In our case, one inch was a $17 difference in fulfillment fee. One inch.

So we went back to the supplier and said we need to shrink the product by an inch. Why? What? Why? Yes — we need to do that. We shrank the product, we shrank the box, we kept every performance characteristic identical, and we immediately saved about $17 per unit. Run that against ten thousand units a year and tell me what other single change in your business pays like that.

We have also lived through the reverse. Pull up fee history on our body unit and you can see it plainly: from around May 2024 through September, Amazon reclassified our unit into a higher tier and our fulfillment cost jumped off its usual ~$16 baseline. It took us three months to redesign the box and get back down. Three months of bleeding, on a metric most people never open.

That is why we built a box optimizer into our platform — it checks which tier your product actually falls into, instead of you finding out from a fee report six weeks later. (I would show it to you here, but I have no internet at this station. You will have to take my word for it.)

The box

While we are on boxes. Everybody in this industry stitches theirs together with glue. Ours folds like Lego. No glue, every dimension locked, thirty seconds to assemble, and it is genuinely sturdy. I am proud of it.

The old box could take six stacked on top. The new one takes fifteen. That is not vanity — that is pallet space, container fill, and storage fees, all in one design change.

Meanwhile Amazon Global Logistics just pulled up to collect. Raw material for roughly 2,500 units heading to the assembly plant. Best freight forwarder we have worked with so far, and believe me, we have worked with a lot of them.

---

If there is one thing to take from this: stop trying to win on price, and go into the weeds of what your product is made of. The margin is down there.

Money saved is money earned.

Written at a train station in China, thirty minutes before departure.