Why a Product Can Look Profitable in Research and Still Lose Money Per Unit After Launch

9 min read

I see the same story every few weeks. A seller finds a product, the research looks great, the margin on paper is north of 40 percent, and the first purchase order goes out. A few months later they are selling 40 units a day and cannot work out why the bank balance is not moving.

Nothing was wrong with the research. The research was just answering a different question than the one that decides whether you make money. This post is about that gap, what sits inside it, and how to check it before you place the first order.

What Research Tools Do Well, and Where the Numbers Stop

Tools like AMZScout are good at the job they were built for. They help you see whether a market exists: how much demand there is, how crowded it is, what price range works, and roughly what competitors are selling. That is exactly what you need at the idea stage, and I use that kind of data myself.

The trouble starts when the "margin" shown in research gets treated as a forecast. In most cases that number is the selling price minus the product cost minus Amazon's fees. It is a ceiling. It assumes every unit sells at full price, nothing comes back, nothing sits in a warehouse for months, and nobody pays for traffic.

No research tool can know your real freight bill, your return rate, how much launch ads will cost you, or how long your stock will sit. Those numbers only exist once the product is live, and they are the ones that decide whether a unit makes money.

There is also a human side to this. When you have spent weeks finding a product and negotiating with a supplier, you want the numbers to work. I call it seller bias. You stop questioning the spreadsheet because you have already fallen for the product. The fix is boring: put real costs into the sheet before you get attached.

The Costs That Show Up After Launch

Landed cost is not the factory price. A supplier quotes you $7.00 a unit. Then come freight, duties, inspection, prep, and inbound shipping to Amazon. In the example further down, I use $2.50 as the add-on, so the real cost is $9.50. Plenty of sellers run their numbers on the quote and never update them.

FBA fees follow size and weight, not price. Shipping is a flat cost tied to how big and heavy the packed product is, whether it sells for $15 or $50. A box that ends up an inch larger than planned can push you into a different size tier. That is why cheaper products get hurt most by fee surprises, since the fixed cost takes a bigger bite of a smaller price.

Returns cost more than the lost sale. I have worked with sellers whose return rate was as high as 30 percent and others at 4 to 5 percent. Category matters a lot. For FBA, I like to see returns stay under 8 to 10 percent. Each return can mean a refund, a product that can no longer be sold as new, and handling fees that do not come back to you. A 10 percent return rate does not cost you 10 percent of revenue. It costs more.

Storage adds up quietly. Ordering a big first batch because of a minimum order quantity or a freight discount feels smart. Then the stock sells slower than planned, and storage and aged inventory charges start eating the profit. I have seen a brand pay Amazon thousands of dollars in a single month in long-term storage fees because December stock was still sitting there in January. That money could have been working capital. Instead, it went to a warehouse.

The Launch Ad Spend Problem

A new product has no reviews, no browsing history, and no organic rank. In the early weeks, most of its sales come from ads. On some launches I have managed, 90 percent of sales were ad-driven in the first month, which is normal for a product that has not earned its place yet.

So the real question at launch is not "what is my margin?" It is "what can I afford to pay for a click?" That number is your break-even CPC, and it is simple to work out: selling price multiplied by conversion rate multiplied by the ACoS you can tolerate.

Take a $29.99 product, an 8 percent conversion rate, and a 25 percent break-even ACoS. That gives you about $0.60 per click. If the suggested bid on your main keywords is $1.10, you are underwater on every click, and you only find out after the money is spent. New listings usually convert worse than established competitors, so the real number tends to be lower than the one you hoped for.

Launch spend can still be a good investment. Sometimes it is the right call to lose money for a few weeks to build reviews and rank. But that should be a decision you make with a number in front of you, not something you discover on the statement.

One Product, Two Margins

Here is a worked example. To be clear, these numbers are made up for illustration. They do not come from a real product or account, and your own costs and Amazon's current fees will be different. I chose round figures to show the pattern because the pattern is what repeats. The same product, sold at $29.99, looks very different at three points in time.

Illustrative figures only. Check current Amazon fees and your own costs before using any of these numbers.

Same product, same price, same supplier. The research said $13 a unit. Launch month said a small loss. Once ads settle and the listing builds organic rank, it lands at about $3 a unit, which is a real business, just a much smaller one than the first column promised.

If the seller had planned around $13, they would have bought inventory, hired help, and set targets that this product can never support. How crazy is that? The product was fine. The expectation was wrong.

A Checklist to Run Before the First Order

These are the questions I ask sellers before they commit money to inventory. If you cannot answer one, that is the thing to research next.

1. What is my real landed cost per unit, including freight, duties, inspection, prep, and inbound shipping?

2. Which size tier will the packed product land in, and what is the FBA fee at that tier? Would small packaging changes move it?

3. What return rate do products in this category see, and what does one return actually cost me?

4. How many units can I realistically sell in 60 to 90 days, so the first order does not sit and collect storage fees?

5. What is my break-even CPC, and how does it compare with the suggested bids on my main keywords?

6. How much am I willing to lose per unit during launch, and for how many weeks?

7. What does the product earn per unit once ad spend settles to a normal level, around 10 to 15 percent of sales?

8. Under what conditions do I stop? What would make me pull the plug?

Question 8 matters most, and it is the one nobody writes down. Decide it now, while you are not emotionally attached.

Tracking Real Profit After Launch

Everything above is an estimate until real orders come in. After launch, you want actual numbers, and you want to look at them on a schedule, not only when something feels off.

I like two checkpoints a month. One early in the month, for a rough idea of refunds and returns coming back. One around the 20th, which is the more accurate figure because ad attribution and refunds have settled by then.

You can do the basic version in about ten minutes with Amazon's own reports. Open Payments: go to the Reports Repository, request the report for the month, download the CSV, and filter out the transfer lines, since transfers include money from the previous month's orders. What is left shows what Amazon actually charged you and paid you. Then subtract your product cost, which is your cost per unit multiplied by the quantity on each order.

From there, do it per ASIN. A single account-wide number hides the problem, because one product can carry three others that are quietly losing money. I ended up building Sellerview.ai because I was tired of doing this by hand for every brand I worked with, but a spreadsheet is enough to start. If you also want a quick per-unit check before you place an order, the free Amazon FBA profit calculator on our site does the basic math in a couple of minutes. What matters is the habit, not the tool.

The Takeaway

Research tells you whether a market is worth entering. It cannot tell you whether your version of the product will make money, because that depends on costs and behavior that only appear after launch.

Run your numbers with real landed cost, real fees, a realistic return rate, and a launch ad budget before you order. Then track actual profit once you are live. Sellers who do both are rarely surprised, and the ones who skip it usually find out the hard way.

About the Author

Himanshu Gaba is the founder of Sellerview.ai, a profit analytics platform helping Amazon FBA and FBM sellers see their true margin after fees, ad spend, and cost of goods, not just their revenue. With 7+ years working closely with Amazon sellers on advertising strategy and account management across 300+ brands, he focuses on building tools that give sellers SKU-level profit visibility instead of the blended, top-line numbers most dashboards show. Himanshu writes and speaks regularly about Amazon seller profitability, PPC efficiency, and margin tracking.