There is no fixed rule, but below about $1,000 a month you cannot gather enough data to optimize anything. For a single new listing, treat that as the floor for the first 60 days, plan to spend inefficiently while you buy rank, and fund it separately from your inventory budget.
The short version
- $1,000 a month is the practical floor. Not a minimum requirement, just the point below which the data is too thin to act on.
- The first 60 days are tuition. You are paying for keyword truth, not for profit.
- Budget by keyword count, not by revenue percentage. Ten terms tested properly beats a hundred tested badly.
- Never fund advertising out of the inventory line. The two budgets fail at different times and for different reasons.
- Paid traffic is one channel of five. A listing that only ever gets paid clicks is a rented business.
The number that decides it
Everything about a new listing budget comes back to one arithmetic fact: you need enough clicks per keyword to know whether that keyword converts. If a term costs a dollar a click and you need 30 to 50 clicks before the conversion signal means anything, then testing ten terms costs somewhere between $300 and $500 before you have learned one thing.
That is the entire reason $1,000 a month is where I tell people to start. There is no hard minimum on Amazon and there is none at Flapen either, but a $300 monthly budget spread over a hundred keywords generates noise that looks like data, and acting on noise is more expensive than not advertising at all.
So the correct question is not what percentage of revenue to spend. It is how many keywords you are willing to fund to statistical honesty, and for how long.
The pre-spend checklist
Work through these before the first campaign goes live. Each item has a definition of done, because half done here is where the money goes.
- Listing conversion is credible. Done properly means your primary image, title, and price have been compared against the top ten in the category, and you would click yours in that lineup.
- Reviews are not zero. Done properly means you have a compliant path to first reviews in motion before spend starts, not planned for later.
- Inventory covers the ramp plus the reorder. Done properly means running out mid-test is impossible, because a stockout wastes every dollar of ranking you just bought.
- The keyword list is short and deliberate. Done properly means ten to twenty terms you can defend, split into exact and broad, not an auto-generated list of four hundred.
- You know your break-even advertising cost of sale. Done properly means unit economics after Amazon fees, freight, and expected returns, written down before launch.
- You have set a learning budget and a learning window. Done properly means a fixed amount you accept losing, with an end date, typically the first 60 days.
- Negative keywords are running from week one. Done properly means you review search terms weekly and cut what does not convert, rather than discovering it in month three.
- You know which decision each report will inform. Done properly means every weekly number has an action attached to it. Numbers with no attached decision are decoration.
If you cannot tick item one, do not start. Advertising into a listing that does not convert raises the price of finding out that it does not convert.
Where the money should sit
| Phase | Job of the spend | What you are watching | What good looks like |
|---|---|---|---|
| Weeks 1 to 2 | Discovery across your keyword list | Search term report, click-through rate | You can name your ten best terms |
| Weeks 3 to 6 | Concentration on terms that convert | Conversion rate by term, cost per acquisition | Wasted spend falling week over week |
| Weeks 7 to 12 | Buying rank on the few terms that matter | Organic position on those terms | Organic sales rising as a share of total |
| Month 4 onward | Efficiency and defense | Contribution margin, branded term cost | Paid share of revenue falling, not rising |
If paid is still carrying nearly all of your revenue in month six, that is not an advertising problem to solve with more budget. It is a listing, product, or market problem showing up in the advertising account.
Advertising is one channel out of five
There are five ways to get traffic to an Amazon listing: organic search, paid placements, promotions and deals, influencer and creator content, and off-channel traffic you send in from outside Amazon. Most sellers run two, usually organic and paid, then wonder why their acquisition cost only ever goes up.
The relevance to your budget is direct. When paid is the only channel doing work, every competitor's bid increase is your problem, and your cost per click sets your ceiling. Sellers running three or four channels see the paid budget do less work at a better price, because velocity arriving from elsewhere lifts organic position, which lowers the amount of rank you have to rent.
When you interview an agency, ask which of the five they will actually run for you, with the names of the people who run each. The honest ones will tell you which two they are strong at.
What most agencies will not tell you
A large share of what gets billed as advertising management is bid adjustment on autopilot, and the underlying performance is set by the listing, the price, and the review position. That work is less visible and harder to charge hourly for, which is exactly why it gets skipped in favor of dashboards showing bid changes.
The second thing: a low advertising cost of sale on a new listing is often a bad sign, not a good one. It usually means the campaigns are only bidding on your own brand name and the long-tail terms nobody contests, which is a cheap way to look competent while your product never enters the ranked competition for the terms that matter.
Related answers
- Rank the best Amazon PPC courses
- How to rank new Amazon listings from scratch
- Which product analytics to track weekly
- KPIs an Amazon agency should report weekly
- Amazon seller roadmaps and capital: the complete guide
We will review a live advertising account and send the wasted spend back in writing within 48 hours, free, at Flapen.
