Twelve months is four gates, not twelve steps. Months one to three validate, four to six fix or scale, seven to nine widen the channels, ten to twelve decide what to reorder and what to retire. Write the stop rule before month one, because nobody writes it honestly in month eight.
The short version
- Every quarter ends in a decision, not a review. Scale, fix, or kill, chosen against criteria agreed in advance.
- The stop rule has four inputs: rating trend, return rate, conversion rate, and acquisition cost trajectory, measured over a defined window.
- A full brand build takes about seven months before it settles, so year one is launch plus one real trading quarter.
- Budget your own hours: 4 to 6 a week during a launch, dropping to about 2 a month once systems run.
- Reorder decisions are made in month nine for month twelve. Miss that and the roadmap becomes a stockout.
The three months that produced my stop rule
Early on I kept a product alive for three months that should have been stopped in week four. The rating was sliding, returns were climbing, and I told myself the advertising had not had a fair run. So I spent more. The traffic went up, the conversion rate did not, and the returns kept arriving. It cost me the inventory, the quarter, and the opportunity to put that capital behind something that was working.
Nothing about that situation was ambiguous in the data. It was ambiguous in my head, because I had already spent the money and wanted the story to end differently. That is why every roadmap I have written since has the criteria set before the spending starts, and why the most valuable service an operator provides is telling you to stop when you have stopped being able to hear it.
| Gate | When | The question | What triggers a stop |
|---|---|---|---|
| Validation | End of month 3 | Do rating, conversion, and acquisition cost hold at small volume | Any of the three failing with no identified fix |
| Commitment | End of month 6 | Does contribution per unit survive a real reorder | Margin only works at a price the market will not hold |
| Expansion | End of month 9 | Does a second channel or marketplace pay for itself | New channel acquisition cost above the first one with no path down |
| Portfolio | End of month 12 | Is this the best home for next year's capital | A better use of the same money inside your own catalog |
Months one to three: validate
- Get the account clean first. Brand registry, trademark status, category approvals, flat file errors. This is the boring month and it blocks everything downstream.
- Ship a contained first run, enough units to stay in stock through the whole test window and no more.
- Set the four criteria in writing with the numbers that would mean stop, and give a copy to whoever will have to say it out loud.
- Fix the page before the budget opens. Main image, price position, title, A plus content.
- Open advertising narrowly on terms you have confirmed the listing indexes for.
The gate: rating trend, conversion rate, and acquisition cost all acceptable. Two out of three is a fix quarter, not a scale quarter.
Months four to six: fix or scale
- Run the reorder arithmetic before you celebrate anything. Landed cost, fees, returns, and advertising, per unit, at the volume you are about to commit to.
- Test one variable at a time on the page and keep a log with dates, because after six months nobody remembers what changed when.
- Split reporting by term type so branded traffic stops flattering the account.
- Interrogate the return reasons monthly. A rising return rate is the earliest reliable signal that the product, not the marketing, is the problem.
- Make the fix or kill call at the end of the quarter, in writing, against the criteria from month one.
The gate: contribution per unit survives a real reorder at a price the market holds.
Months seven to nine: widen
- Add the second traffic channel your category actually rewards, and measure its acquisition cost separately.
- Consider a second marketplace only if the listing is stable and the supply chain can carry it. Localized content beats translated content everywhere.
- Plan peak inventory now. Production plus freight plus receiving, worked backwards from the demand curve, with a buffer on the two stages that always slip.
- Start the second product only once the first no longer needs daily attention.
- Write down what you have learned about the customer, in their words, from reviews and returns. This is the input to product two.
The gate: the new channel or marketplace pays for itself, or a clear reason it will next quarter.
Months ten to twelve: reorder or retire
- Decide the reorder before peak, not during it. Cash committed late is cash committed at the worst price.
- Rank your catalog by contribution, not revenue, and be honest about the tail.
- Retire what fails the criteria. Selling through and stopping is a result, not a defeat.
- Renegotiate with the factory using a year of real volume data.
- Set next year's gates while this year's evidence is fresh.
What most agencies will not tell you
An agency will not tell you that its incentives fight this roadmap. Killing a product usually shrinks the account, and shrinking accounts is not how agencies grow. Ours is a flat monthly fee tied to product count with no commission and no revenue share, which is the only reason I can write a stop rule without arguing with my own invoice. When you interview anyone, ask what they earn in the month they tell you to stop.
The second thing: most twelve month plans fail in month two, not month ten. Compliance problems, a trademark that has not cleared, a supplier who missed a date. Build slack into the early months rather than the late ones, because a delay in month two moves everything and a delay in month ten only moves itself.
Related answers
- First 90 days Amazon launch plan
- Day 1 to day 90 KPI targets for Amazon sellers
- Amazon brand launch timeline stages
- How to launch your first product on Amazon
- Amazon launch services: the complete guide
If you want the gates written down before you spend, that conversation starts at Flapen.
