Amazon Keyword Research: How to Find Keywords That Drive Sales

Amazon keyword research is not the same as Google keyword research. Amazon searchers are buyers, not browsers. They use specific, often longer phrases because they know what they want. Your job is to find the exact phrases your ideal buyer uses and place them in the right parts of your listing.

Why Amazon Keywords Work Differently

Google ranks pages based on authority, backlinks, and content quality. Amazon ranks products based on sales velocity and keyword relevance. If Amazon's algorithm sees that your product sells consistently when people search for a particular phrase, it ranks your product higher for that phrase. Keywords in your title, bullets, and backend search terms tell Amazon which searches to show you for.

The A9 Algorithm and Keywords

Amazon's A9 algorithm indexes keywords from three main places in your listing:

Your title (highest weight)
Your bullet points and description (medium weight)
Your backend search terms (indexable but not visible to customers)

Keywords in your title that also have sales history for that phrase carry the most ranking power. Backend search terms help you get indexed for additional phrases without cluttering your visible copy.

Where to Find High-Converting Keywords

*Amazon's own search bar:* Type your main product keyword and note every autocomplete suggestion. These are real search queries with real volume. Amazon only suggests phrases that get searched frequently.

*Competitor listings:* Look at the top 5 to 10 competitors for your main keyword. What phrases appear in their titles? What words repeat across multiple top-ranking listings? Shared vocabulary signals what Amazon's algorithm associates with your product category.

*Amazon's "Customers Also Searched For":* On any product page, scroll down to see related searches. These show you the semantic neighborhood your product needs to live in.

*Reverse ASIN tools:* Tools like Helium 10 Cerebro, Jungle Scout Keyword Scout, or DataDive allow you to enter a competitor's ASIN and see every keyword they rank for. This is the most efficient way to identify keyword opportunities your competitors have already validated.

Keyword Research Tools Compared

Helium 10 Cerebro is the most comprehensive reverse ASIN tool. It shows estimated search volume, keyword difficulty (based on competitor ranking positions), and whether a keyword is "organic rank eligible" for your listing.

Jungle Scout Keyword Scout shows search volume trends over time, useful for spotting seasonal keywords. It also shows the number of products competing for each keyword.

Merchant Words is the oldest Amazon keyword database. Its search volume estimates are often cited as less accurate than newer tools, but it covers long-tail phrases that other tools miss.

DataDive aggregates data from multiple sources and is particularly useful for finding gaps: keywords that multiple top competitors rank for but you do not yet.

Keyword Prioritization

Not all keywords deserve equal space in your listing. Prioritize by:

Search volume (high means more potential traffic)
Conversion relevance (does this search lead to your exact product?)
Competition (can you realistically rank for this phrase?)

A keyword with 5,000 monthly searches but perfect relevance to your product is more valuable than one with 50,000 searches but only tangential relevance. Irrelevant keywords that get clicks but not purchases signal bad relevance to Amazon's algorithm.

Where to Place Your Keywords

Title: include your primary keyword (highest-volume, most relevant phrase) in the first 80 characters. Amazon truncates titles in search results at around 80 characters.

Bullet points: use secondary keywords naturally within descriptive bullets. Do not stuff keywords; Amazon's algorithm reads for context and relevance, not just presence.

Backend search terms: add variations, misspellings, and related phrases that did not fit naturally in the visible copy. The backend keywords byte limit is 249 bytes in the US, UK and EU, not 250. Do not repeat keywords already in your title, Amazon already indexes those.

A+ Content and descriptions: these are not primary indexing signals but do contribute to conversion rate, which indirectly boosts keyword rankings.

Common Mistakes

Repeating the same keyword multiple times in the title. Amazon does not give extra weight for repetition; you waste space that could hold additional keywords.

Using competitor brand names in backend terms. Amazon's policy prohibits this and can result in listing suppression.

Ignoring Spanish-language keywords in the US marketplace. Many product categories have significant Spanish-speaking buyer populations. Adding relevant Spanish phrases to backend terms can open a new traffic stream.

Not updating keywords after seasonal events. After Q4, search behavior shifts. Keywords that performed well in November may underperform in February. Search terms and keywords are not the same thing

Sellers use the two words interchangeably and Amazon does not. A keyword is any word Amazon has indexed your listing for, wherever it came from. Search Terms is the name of one specific backend field in the listing editor, under the Keywords tab, and it holds 250 bytes of text that buyers never see.

Getting the distinction right matters because it changes where a phrase belongs. Anything you want a buyer to read goes in the title and bullets. Anything you only want Amazon to index goes in Search Terms. Putting a phrase in both wastes the backend space, because a term already in your visible copy is already indexed.

The 250-byte field, measured properly

The limit is 250 bytes, not 250 characters, and the difference bites in non-English marketplaces. A plain ASCII letter is one byte. An accented character such as é is two bytes in UTF-8, and a German ß is also two. On amazon.de or amazon.fr a field that looks like 240 characters can be over the limit, and Amazon does not warn you. It silently ignores everything past the cut.

What to leave out of the field entirely:

Commas. Amazon does not need separators, and every comma costs a byte. Use single spaces.
Any word already in your title, bullets or brand name. It is indexed already.
Repeated words. The field is a bag of words, not a set of phrases, so "yoga mat" and "mat for yoga" are covered by including yoga and mat once each.
Competitor brand names and ASINs. This is a policy violation and a known cause of listing suppression.
Subjective or temporary claims: best, cheapest, new, sale, free shipping.
Plurals where the singular is present. Amazon handles both forms.

What to put in: genuine synonyms, the terms buyers use that you would never print on a package, common misspellings that still get typed, Spanish equivalents for the US marketplace, use cases, and the problem your product solves in the buyer's own words.

A repeatable research session

Write down the one phrase a buyer would type if they already knew your product existed. That is your primary term.
Type it into Amazon's search bar and record every autocomplete suggestion, then type it again followed by each letter of the alphabet and record those too. These are real queries with real volume.
Open the top ten organic results for that phrase. Copy every title into one document and count which words repeat across five or more of them. That shared vocabulary is what Amazon associates with the category.
On three of those listings, scroll to the related searches and the comparison table, and add any phrase you had not thought of.
Read the one-star and three-star reviews of the top two competitors. Buyers describe the problem in words that never appear in marketing copy, and those phrases convert well because they match intent exactly.
Sort everything you collected into three piles: must appear in the title, belongs in a bullet, backend only.

That process takes about an hour per product and costs nothing. Paid tools speed up step two and step three and add volume estimates, but they do not replace step five.

Judging a keyword before you give it space

Question	Keep it if	Drop it if
Would a buyer typing this want exactly your product?	Yes, clearly	It is a category term with mixed intent
Do the top results look like your product?	They match	They are a different product type
Can you rank for it within a quarter?	Some results are weak listings	Every result is an established brand
Does it describe a feature you actually have?	Yes	You would be stretching

The third column matters more than volume. A keyword that brings clicks and no sales tells Amazon your listing is a poor answer for that query, and that signal follows you.

Search Query Performance, if you are brand registered

Brand Analytics in Seller Central carries the Search Query Performance report, and it is the only place a seller sees Amazon's own numbers rather than an estimate. Per query, it shows search volume, how many impressions your ASIN got, click share, cart-add share and purchase share.

Use it two ways. First, find queries where your impression share is high but your purchase share is low: you are being shown and not chosen, which is a listing or price problem, not a keyword problem. Second, find queries with real volume where your impression share is near zero: those are terms you are not indexed for, and they are the ones worth adding.

Without Brand Registry, the closest free equivalent is the search term report from your Sponsored Products campaigns. Running a broad match campaign on a small budget for two weeks buys you a list of the actual queries that triggered your ads, which is genuine Amazon data rather than a third-party estimate.

Frequently Asked Questions

What are Amazon search terms?

Search Terms is a backend field in the Amazon listing editor, under the Keywords tab, that holds up to 250 bytes of text buyers never see. Amazon indexes it for organic search, so it is where synonyms, misspellings and alternative phrasings go when they do not fit naturally into your title or bullet points.

How many characters can Amazon backend search terms be?

The limit is 250 bytes, which equals 250 characters only when every character is plain ASCII. Accented and non-Latin characters take two or more bytes each, so the effective character count is lower on European marketplaces. Anything past the limit is ignored without any warning from Amazon.

Should I repeat keywords from my title in the backend search terms?

No. Amazon already indexes every word in your title, bullet points and description, so repeating them in the backend field wastes space that could hold a term you are not indexed for. Use the field only for words that appear nowhere in your visible copy.

How do I find the keywords my competitors rank for?

Start free: read the top ten titles for your main phrase and note the words that repeat, then check the related searches and comparison tables on those detail pages. Paid reverse ASIN tools automate this and add volume estimates. If you are brand registered, the Search Query Performance report in Brand Analytics gives you Amazon's own query data for your own ASINs.

How long does it take for new keywords to start ranking on Amazon?

Indexing usually happens within 24 to 72 hours of the listing update. Ranking is separate and slower, because position is driven by sales and conversion for that query. Expect a few weeks of sales history on a term before its organic position settles, which is why sellers often run ads on a new term to build that history.