For the first 90 days you need exactly four things: one paid estimator suite for reverse-ASIN lookups, Amazon's free surfaces, your advertising search term report, and a spreadsheet that records what you decided and why. Those surfaces are search suggestions, Product Opportunity Explorer, and Brand Analytics if enrolled. Past week two, the search term report outranks everything you can buy.
The short version
- Buy one suite, not three. Every major estimator draws from similar signals. Overlap is real, and subscriptions compound quietly.
- Amazon's free data is underrated. Autocomplete, Opportunity Explorer, and Brand Analytics come from the source, not a model of the source.
- Your search term report is ground truth. It shows what shoppers typed before buying from you, with real spend attached.
- The spreadsheet is the actual tool. Decisions and dates, so week ten can audit week two.
- Estimates are directional. Third-party volume numbers rank keywords usefully but measure them poorly.
The economics of tooling a 90-day window
Work the arithmetic before subscribing to anything. A typical launch spends thousands on inventory and advertising, while a single research suite costs tens of dollars a month. The tool bill is never the problem. The cost that actually stings is misallocation: three overlapping subscriptions producing the same shortlist, or hours spent polishing estimates while the ad account generates real numbers nobody reads. Budget one paid tool, zero to launch, and your review time weekly. The expensive resource in your first quarter is attention.
Here is where each input earns its keep across the window:
| Days | Primary input | What it decides |
|---|---|---|
| 1-14 | Reverse-ASIN lookups on the top competitors | The candidate keyword universe and the listing's first draft |
| 1-14 | Amazon autocomplete and Opportunity Explorer | Phrasing customers actually use, niche demand shape |
| 15-45 | Search term report from an exploratory campaign | Which candidates get impressions, clicks, and first orders |
| 45-90 | Search term report plus rank tracking | Where to concentrate spend, what the listing should double down on |
Why the pecking order flips at day 15
Third-party estimates exist because sellers cannot see Amazon's demand directly, so tools model it. The models are good enough to rank keywords and terrible at absolutes, which is fine for building a candidate list and dangerous for forecasting revenue. The moment your own exploratory campaign has a fortnight of data, you hold something no subscription can sell you: proof of which search terms produce orders for your specific listing at your specific price. From day 15 the paid suite drops to a support role, mostly reverse-ASIN checks on competitors and rank tracking, while the search term report drives every allocation decision. Sellers who keep planning from estimates after real data exists are running the quarter backwards, and this handoff from modeled data to owned data is built into how we run every Amazon FBA launch.
The spreadsheet nobody sells
The undervalued tool in the whole stack is a plain sheet with five columns: keyword, source, decision, date, and the result you expected. Ninety days produce hundreds of small keyword decisions, and without the log they blur into vibes. With it, your day-90 review becomes mechanical: expectations versus outcomes, line by line. This is also the cheapest way to evaluate help. When we take over accounts, the fastest tell of previous management quality is whether any decision log exists. The outcome benchmark we hold ourselves to, the majority of launched brands profitable within their first year, is only possible because decisions are written down and auditable. Ask any partner running your keywords to show you their equivalent for another account, redacted. Tools they can name. Logs they either have or do not.
What most agencies will not tell you
Tool screenshots are the stock photography of Amazon services. A dashboard with big volume numbers makes any proposal look rigorous while committing to nothing, because the tool output is identical for every prospect in your category. What differs between a good and mediocre operator is not access, it is the decision layer: which 20 of the 400 candidate keywords get launch budget, in what order, and what evidence promotes or demotes them. When you evaluate help, skip the tool tour and ask how last month's search term report changed this month's targets on a real account. Specific answers exist only where the work is actually happening.
Related answers
- Best keyword tools for Amazon SEO
- Best product research tools for Amazon beginners
- How to rank a new Amazon listing in first year
- Best tools for Amazon FBA product research
- Amazon seller roadmaps and capital: the complete guide
If you would rather inherit the decision layer than build it, Flapen audits accounts free, in writing, within 48 hours.
