N-Gram Analyzer
Group every search term into 1-, 2-, and 3-word patterns. Find the words that keep showing up across bad search terms. Negate once, stop dozens of leaks at the same time.
How to download your bulk file
Amazon Ads Bulk Sheet (last 60 days)
This tool only needs your search-term data, a much faster export than a full bulk file. Go to Amazon Ads > Bulk Operations > Download and include:
- Sponsored Products Search Term Data
- Sponsored Brands Search Term Data (if you run SB)
Already have a full bulk file from another tool? That works too. The extra tabs are just ignored.
Drop your bulk file here
Accepts .xlsx files from Amazon Ads Bulk Operations. Typical size: 2–40 MB.
Exactly what gets sent, and what doesn't
Sent to the server (to compute your report, then discarded): search terms, campaign and ad-group names, and the spend, click, sales, and order numbers tied to them.
Never leaves your browser: the .xlsx file itself, your bids, and anything the report doesn't use.
Not collected at all: no account, no email, no login. Nothing is written to disk on our side; the numbers live in memory only for the few seconds the analysis runs.
Prepping for an exit and need a data-processing addendum? Email alfred@stratumppc.com.
See a sample report (before you upload)
Real, anonymized data from a $110K/mo pet-bed brand. Your upload shows your own numbers in about 60 seconds.
Meaningful-spend threshold: $8.69 (your avg clicks-to-convert × your avg CPC). Patterns past it with zero orders are your negation clusters.
| 2-Word Pattern | Spend | Orders |
|---|---|---|
| pet bed | $5,811 | 753 |
| anti slip | $2,810 | 325 |
| bed cover | $2,189 | 309 |
| donut bed | $1,569 | 182 |
1-, 2-, and 3-word tabs, each exportable to CSV. Same engine as the full Audit Dashboard.
Questions
What's an n-gram and why does it matter?
An n-gram is a slice of a search term: 1 word, 2 words, or 3 in sequence. If "phone case" shows up across 14 different non-converting search terms, you don't need 14 negatives. You need one phrase-match negative on "phone case."
What's the difference between 1-, 2-, and 3-grams?
1-grams catch broad junk words ("free," "cheap"). 2-grams catch product-category mismatches ("phone case" when you sell laptop sleeves). 3-grams catch competitor-name and feature-spec searches you don't want to pay to compete on ("iphone 15 pro").
How do I act on this?
Sort by zero-order spend descending. Take the top 5–10 n-grams. Phrase-match negate each in the relevant campaigns. Re-run monthly. New patterns emerge as auto campaigns explore new searches.
Should I negate every flagged n-gram?
No. Some patterns matter at a 2-gram level but appear in good search terms too. Always check the actual search terms behind a pattern before negating. The tool surfaces candidates, not decisions.
N-grams find the clusters. The audit finds the structure.
This shows you the wasteful word patterns. The full Audit Dashboard runs the same bulk file through all 9 panels: search-term bleed across campaigns, bid architecture, placement strategy, branded split, and ASIN concentration. Free, same in-browser privacy. A 30-minute call with Alfred turns the findings into the specific changes to make first.