The fastest free way to find Amazon keywords: pull Amazon’s search bar autocomplete for your root terms, rank them against Brand Analytics’ Top Search Terms report, then mine your own search term report for the queries that already convert. Those three sources are Amazon’s actual data — the same data every paid tool is estimating from and selling back to you with a dashboard.
That’s the whole method. The rest of this post is how to run each source properly, how to tell a keyword worth targeting from a keyword worth ignoring, and where each type of keyword belongs on the listing — because finding them is half the job, and most sellers fumble the other half. [Rolls up sleeves. This is the fun kind of spreadsheet.]
What Counts as an Amazon Keyword?
An Amazon keyword is any query a shopper types into the search bar — and that definition matters, because Amazon shoppers don’t search like Google shoppers. Nobody on Amazon types “best ways to improve sleep quality.” They type “magnesium glycinate 400mg” with their card already out. Amazon keywords are shorter, more literal, and closer to the wallet.
You’re hunting three species: head terms (the two-or-three-word category queries with big volume and big competition), long-tail terms (specific, lower-volume queries that convert like crazy because the intent is precise), and attribute terms (the sizes, counts, materials, and use cases shoppers bolt onto a head term). A healthy listing targets all three — on purpose, in different places.
How Do You Find Amazon Keywords for Free?
Start where the shoppers are: Amazon’s search bar. Type your root term slowly and record every autocomplete suggestion — then add a letter (“collagen a”, “collagen b”) and record again. Autocomplete is ranked by real search behavior, so you’re reading Amazon’s demand data straight from the tap. Twenty minutes of this produces 50–100 candidate keywords, and it costs exactly nothing.
Then structure the hunt:
Where the Most-Searched Terms on Amazon Actually Live
Every few months an article goes around listing “the top 100 Amazon keywords.” Read the date, then close the tab. The genuinely most-searched terms on Amazon change weekly — they swing with seasons, launches, and whatever TikTok did last Tuesday — and the only current list is inside Amazon Brand Analytics: the Top Search Terms report, which ranks every query on the platform by search frequency, with the top three clicked products for each.
Here’s how to use it without drowning: search the report for your root terms, note the search frequency rank of each variant, and watch which competitor products soak up the clicks. A term where the top three products take 90% of clicks is a fortress. A term where clicks are scattered is a door left open. That click-concentration read is worth more than any volume estimate a third-party tool will sell you.
Paid tools estimate Amazon’s data from the outside. Brand Analytics is Amazon’s data. When the two disagree, believe the one wearing the badge.
Mining Your Search Term Report (Your Unfair Advantage)
Everything above is public information — your competitors can read the same autocomplete. Your search term report is the one source they can’t see: the literal queries shoppers typed before clicking your ad and buying. A query with orders against it isn’t a keyword candidate; it’s a proven converter that deserves exact-match campaigns, a place in your title or bullets, and a bid strategy of its own.
Run the loop weekly: harvest converting queries into exact match, push the best ones into listing copy, and feed the junk into negative keywords so it stops billing you. That single loop is most of what a $500/month agency retainer buys, and it takes about forty minutes.
Borrowing Competitor Keywords (Legally, Calm Down)
This is the one job where a paid tool earns its subscription: reverse-ASIN lookups — paste a competitor’s ASIN, see the keywords it ranks for. But you can approximate it free: read the top five competitors’ titles and bullets and list every attribute term they repeat (those repetitions are not accidents); check which queries they dominate in Brand Analytics; and read their reviews for the vocabulary real customers use. Customers write reviews in search terms — “bought this for my standing desk setup” is a keyword wearing a costume. Ten minutes in a competitor’s one-star reviews also hands you the objection vocabulary your bullets should be answering, which is a conversion play disguised as keyword research.
Where Each Keyword Actually Goes
Finding keywords and placing them are different skills, and placement is where listings quietly fail. The hierarchy: your title carries the primary head term and top attributes; bullets carry secondary terms and use cases woven into benefit copy, not stuffed like a suitcase; backend search terms carry everything that didn’t fit — synonyms, alternate spellings, regional terms — within the 250-byte limit set out in Amazon’s own guidelines and covered in our backend keywords guide; and A+ Content carries almost nothing for Amazon’s index but supports conversion, which is what the ranking algorithm is really watching. Organic rank follows conversion; keywords just buy you the audition. If you want the full placement discipline applied to a live listing, that’s the core of our Amazon SEO service.
How Many Keywords Should One Listing Target?
More than one, fewer than you think. A workable structure: one primary head term the whole listing is built around, three to five secondary terms carried by the bullets, and twenty to forty long-tail and attribute variants spread across backend search terms and your exact-match ad campaigns. Amazon indexes phrases, not densities — once a term appears once in an indexed field, repeating it does nothing except make your copy read worse.
The practical test for whether a keyword makes the roster: would a shopper typing this query be satisfied landing on this exact product? If the honest answer is “sort of,” the click will cost you twice — once in ad spend, once in the conversion-rate damage that drives ACoS up and organic rank down. Relevance isn’t a virtue on Amazon; it’s the billing model.
And when the catalog grows, run this per ASIN — not per brand. Two products sharing one keyword set end up bidding against each other in the auction and splitting rank signals in search. Same discipline as everything else on this site: per-ASIN math or it didn’t happen.
The Mistakes That Waste All This Work
FAQ
Three free sources cover most of what a paid tool would tell you: Amazon’s search bar autocomplete, the Search Query Performance report in Brand Analytics (free with Brand Registry), and your own search term report — the exact queries shoppers used before buying from you.
The most searched terms are visible in Brand Analytics’ Top Search Terms report, which ranks every query on Amazon by search frequency. The top of the list is dominated by branded and seasonal terms — and it changes weekly, which is why the ranked list beats any static article claiming to know the top keywords.
Amazon allows 250 bytes in the backend search terms field — bytes, not characters. Exceeding the limit can cause the entire field to be ignored. Full details in our backend keywords guide.
Not to start. Brand Analytics, autocomplete, and your search term report are free and use Amazon’s own data — which is where paid tools estimate from anyway. A paid tool earns its fee later, for competitor reverse-ASIN lookups and rank tracking at scale.