Amazon FBA How to Start in Five Gated Steps

Five gates decide it, in order: market, product, traffic, plan, launch. The market clears $2M per year, the product aims 0.2 stars above the niche average, and Phase 1 buys 200 units. Budget $5,000 to $10,000 for that validation, and write the stop rule before the first order leaves the factory.

The short version

• The gates run in one order. Market, product, traffic, plan, launch. Nothing moves until the gate in front of it clears.
• The market is the first filter, not the product. A niche has to turn over $2M per year, grow year over year, and hold returns under 8%.
• The product bar is a number you can look up. Build for 0.2 stars above the niche average, read from the negative reviews of what already sells.
• Phase 1 is 200 units and $5,000 to $10,000. Up to 4 products can be tested at once, and real capital follows only what validates.
• The stop rule is written before you spend. Four signals across a 60 to 90 day window decide whether you scale, fix, or kill.

The five-step start checklist

This is the sequence we run on every launch, and skipping a gate is what kills most first products. Each item below carries the test for what done properly looks like.

1. Step 1, market. Done properly means one named market turning over at least $2M per year, growing year over year, with a return rate under 8%. Below that floor there is not enough revenue to capture profitably once you pay to acquire the customer.
2. Step 2, product. Done properly means a build brief written from the negative reviews of what already sells, aimed 0.2 stars above the niche average. The rating gap names the complaint to fix, so you innovate where the market asked and nowhere else.
3. Step 3, traffic. Done properly means naming which of the five channels you can win in profitably: organic, paid, promotions, influencer and creator, and off-channel. Write the cost of customer acquisition (CAC) you expect on each one before ordering inventory.
4. Step 4, plan. Done properly means one number for entering and winning this market, staged across two budget phases. Phase 1 validates and Phase 2 scales, so every dollar answers to the market size and the share you can take.
5. Step 5, launch. Done properly means 200 units live and three validation milestones measured on real orders: rating, conversion rate, and cost of customer acquisition. Scale only when all three hold.

What starting costs, in the order you spend it

"I don't know how much money I need to launch." That is the line I hear most from pre-launch sellers, and the answer is that the number is staged rather than single.

Flapen figures as of September 2026. Your landed cost, freight, and category are yours to fill in.

Do the arithmetic with your own inputs. Multiply your landed cost per unit by 200, add the advertising you plan for the validation weeks, and read the total. If it sits far outside the $8,000 to $15,000 we budget for one product, find the wrong input before the purchase order goes out.

Why the majority of brands reach profitability in year one

Our 50 operators run Amazon brands by hand, with sourcing and quality control in Guangzhou and creative studios in Dubai. We launch 100+ brands in a typical year, run about 70 by hand today, and the majority of them reach profitability within their first year. The frameworks behind that were built across 500+ brands.

That outcome is not a talent claim. It is a subtraction claim.

The losers get stopped early, so the survivors carry the year instead of paying for a graveyard. Scale / Fix / Kill reads four signals on every live product: rating trend, return rate, conversion rate, and cost of customer acquisition trajectory. The Kill Criteria set the window at 60 to 90 days, and a product not improving inside it stops.

Hold anyone to that outcome, us included. Ask what share of last year's launches were profitable inside twelve months, then ask how many were killed. A provider who has never killed one has never had a stop rule.

What most agencies will not tell you about starting

Four items belong on your checklist and rarely appear on a proposal. Each one has a version that counts as done properly.

1. Nobody volunteers to write the kill rule. Done properly means the four signals and the 60 to 90 day window sit in the agreement before the first invoice. A monthly fee earns the same whether your product thrives or limps.
2. A service list is not a sequence. Done properly means the provider names which of the five gates you are standing at and what clears it. Anyone quoting a price before sizing your market is selling hours.
3. Starting still costs your own time. Done properly means about 2 hours a month once things are running, and 4 to 6 hours a week during a launch. A partner who needs none of your hours has stopped asking you anything.
4. The Amazon side is yours to confirm. Done properly means every account requirement is checked inside your own Seller Central account rather than taken from a page like this one. Grant access through user permissions you can revoke, never through your login.

Run all four on us first. If the answers do not hold up, start on your own and keep the fee.

Related answers

• Amazon Seller Central account creation
• Amazon FBA shipping calculator
• Top agencies for Amazon DSP and PPC integration
• Which service handles Amazon catalog flat files
• Done-for-you Amazon management: the complete guide

One free thing to do this week, for the pre-launch seller holding $5,000 to $10,000 for a first validation. Take the market you are closest to buying into and write three numbers by hand. Write its size last year, its direction over the two years before, and the average rating on page one.

Then write the two lines that stop you. The return rate you will not go above, and the date you will re-read all of it.

Request the free written audit, and get those numbers and a prioritized fix list back inside 48 hours at no charge from Flapen.