Amazon Keywords vs Search Terms: Difference With Examples
Nishant Singh
TL;DR
Key takeaways
A keyword is what you bid on in a campaign. A search term is what the shopper actually typed into Amazon.
Match types decide which search terms a keyword can trigger: exact allows a few, broad can allow hundreds.
Backend search terms are a third thing: hidden listing keywords for organic search, capped at under 250 bytes.
Read the search term report weekly: harvest converting searches into exact keywords and negate the ones that waste money.
Keywords and search terms sit in the same row of your Amazon reports, so most sellers treat them as the same thing. They aren't, and mixing them up is how accounts end up paying for searches nobody chose.
This guide explains the difference in plain terms, shows how one keyword turns into dozens of search terms, and walks through the weekly routine we run on the accounts we manage: read the report, harvest the winners, negate the waste.
Amazon Keywords vs Search Terms: The Short Answer
An Amazon keyword is a word or phrase an advertiser adds to a Sponsored Products or Sponsored Brands campaign and bids on. You choose it, you set its bid, and you pick its match type.
An Amazon search term (Amazon's report calls it the customer search term) is the exact query a shopper typed into the search bar. The shopper chooses it, and you only see it after your ad has served against it.
Amazon compares the two in every auction. If a shopper's search term fits one of your keywords under that keyword's match type, your ad is eligible to show.
That is the whole relationship: one keyword, many search terms. A single broad keyword like "dog bed" can match "orthopedic dog bed large", "washable pet bed" and "dog crate mat" in the same week.
Keyword | Search term | |
|---|---|---|
Who chooses it | You, the advertiser | The shopper |
Where it lives | Your campaign or ad group | Amazon's search bar |
Where you see it | Targeting tab, targeting report | Search term report, column "Customer Search Term" |
What you control | Bid, match type, on or off | Nothing directly; you shape it with match types and negatives |
How many | One per row you create | Often dozens per keyword |
Example | "dog bed" (broad) | "orthopedic dog bed for large dogs washable" |
What you do with it | Raise, lower or pause the bid | Harvest it into a keyword or negate it |
A simple way to remember it: keywords are your guesses about how people shop, search terms are the answer.
How Match Types Connect Keywords to Search Terms
Match types are the rule Amazon uses to decide which search terms a keyword is allowed to trigger. They are the bridge between the two, and the reason one keyword can produce one search term or three hundred.
According to Amazon's match type documentation, updated July 29, 2026, all three manual match types account for plurals, misspellings and translations. Beyond that they differ a lot.
Match type | Which search terms it can trigger | Keyword "dog bed" can match | Can't match |
|---|---|---|---|
Exact | The keyword or a close variant | "dog bed", "dog beds", "dog bde" | "large dog bed" |
Phrase | Any search containing the phrase in order | "large dog bed", "dog bed washable" | "bed for dog" |
Broad | Any order, synonyms and related terms | "pet bed", "bed for large dog", "dog crate mat" | Searches with no related meaning |
Broad is looser than most people expect. Amazon's own example says the broad keyword "sneakers" can match "basketball shoes", "cleats" and "foam runners", and that the keyword itself "may not be contained" in the search.
Our Amazon PPC match types guide covers each one in depth. Three rules from the documentation matter most for this topic:
You can't change a keyword's match type after launch. Add the same keyword again with the new match type instead.
The highest bid wins when match types overlap. If "dog bed" is live as exact and broad, Amazon uses whichever has the higher bid for a matching search.
Auto campaigns use targeting groups, not keywords. Close match, loose match, substitutes and complements decide which search terms you get, so every auto campaign is a search term discovery engine.
Sponsored Brands also supports broad match modifiers. Putting "+" before a word ("+kids shoes") forces that word to appear in the search term.
Search Terms Means Two Different Things on Amazon
Here is where most of the confusion comes from. Amazon uses "search terms" for two unrelated things, in two different tools.
Customer search terms are what shoppers type. They appear in your advertising reports and in Brand Analytics.
Backend search terms are hidden keywords a seller adds to a product listing so the listing can rank organically. Seller Central calls this the Generic keyword field, and its help page on search terms defines them as "keywords that are used by the search engine for lexical matching between an ASIN and a query."
Advertising keywords and backend search terms never touch each other. A keyword in your campaign does nothing for organic rank by itself, and a word in your backend field does nothing for which ads you buy.
Advertising keyword | Backend search terms | |
|---|---|---|
Purpose | Decides which searches your ads can show on | Helps your listing get indexed for organic search |
Where you set it | Amazon Ads console | Seller Central, listing edit, Generic keyword field |
Visible to shoppers | No | No |
Limit | No fixed length per campaign | Under 250 bytes per ASIN |
Costs money | Yes, per click | No |
Competitor brand names | Can be bid on as keywords | Prohibited |
Backend Search Term Rules in 2026
The limit is less than 250 bytes, set in Amazon's guide to optimizing product discoverability. Spaces and punctuation don't count toward it, and the field stops accepting text once you hit the limit.
The old version of this post said Amazon allowed 2,500 characters across five fields. That is no longer true: Amazon reduced the search terms fields from five to one and combined their contents, so plan for one field under 250 bytes.
Amazon's rules for what goes in it:
Use synonyms, abbreviations and spelling variants, all lowercase, separated by spaces.
Don't repeat words already in your title or bullets, and skip "a", "and", "for", "the".
Never include brand names (yours or anyone else's), ASINs, "best", "cheapest", "new" or "on sale". Amazon says breaking these rules can get the ASIN suppressed.
Our Amazon backend keywords guide covers how to fill the field. The useful link between the two worlds runs one way: search terms that convert in your ads are proven shopper language, so the best of them belong in your title, bullets or backend field too.
How to Read the Amazon Search Term Report
The search term report is where keywords and search terms meet. Each row shows one search term, the keyword or target that matched it, and what it cost and earned.
Amazon's report documentation, last updated May 18, 2026, sets out the facts that shape how you read it:
It only includes search terms with at least one ad click, so its impressions won't match Campaign Manager.
The lookback window is 65 days, with summary or daily views.
Seller sales are attributed over 7 days from the click; vendors also get 14-day columns.
Alphanumeric entries like "b00ipgvvz4" are ASINs: the product page your ad showed on, from auto campaigns and product targeting.
1. Download the Report
In the Amazon Ads console, open Reports, create a report, and pick Sponsored Products with the "Search term" report type. Choose the last 60 days, summary view.
Our search term report guide has the download steps for Sponsored Brands too.
2. Put Targeting and Customer Search Term Side by Side
These two columns are the keyword vs search term question in one view. "Targeting" is what you bid on; "Customer Search Term" is what the shopper typed.
When they're identical, an exact keyword did its job. When they're different, a phrase, broad or auto target reached out to something new, and that row deserves a decision.
3. Sort by Spend, Then Look at Orders
We sort by spend descending and read the top 50 rows first. That's usually where 70% or more of the money sits.
For each row, ask one question: is this search relevant to the product, and is it converting near target ACOS (advertising cost of sales)?
4. Ignore the Last 7 Days
Sales attribute for up to 7 days after a click, so recent clicks still owe you orders. Judge terms on data that is at least a week old.
5. Decide: Harvest, Negate or Wait
Every row with real spend lands in one of three buckets:
Harvest: relevant, converting at or under target ACOS. Make it a keyword.
Negate: irrelevant, or relevant with enough clicks and no orders. Block it.
Wait: too few clicks to judge. Leave it and look again next week.
Two related reports answer different questions. The Search Term Impression Share report shows your share of ad impressions on each search term over 90 days. Brand Analytics shows search terms across all of Amazon, including organic, if you're brand registered; our Brand Analytics guide explains it.
How to Harvest Search Terms Into Keywords
Harvesting is turning a converting search term into a keyword you bid on directly, usually as exact match. It's how auto and broad campaigns pay for themselves: they discover, and exact campaigns exploit.
Amazon's own Sponsored Products targeting guide suggests running an automatic campaign for about two weeks, then using its search terms to build manual campaigns.
1. Set a Harvest Rule
Pick a rule before you look at the data, or every term starts to look promising. Ours is usually two or more orders in the last 30 days at or under target ACOS.
2. Add the Search Term as an Exact Keyword
Put it in an exact match campaign or ad group with its own bid. Here is the worked example from the image above (example numbers):
The auto campaign search term "orthopedic dog bed large" got 40 clicks at $1.20, so $48 of spend.
It produced 4 orders at $45, so $180 in sales: a 26.7% ACOS against a 30% target.
A sensible starting bid is target ACOS × average order value × conversion rate: 0.30 × $45 × 10% = $1.35.
3. Negate It Where It Came From
Add the same term as a negative exact in the auto or broad campaign that found it. Otherwise both campaigns compete for the same search, and you can't tell which bid is setting the price.
Use negative exact, not negative phrase. Negative phrase on "dog bed large" would also block the next long-tail term the auto campaign might find.
4. Watch It for Two Weeks
A harvested keyword sometimes performs differently in exact than it did in auto, often because the bid or placement changed. Check it after 14 days and adjust the bid.
The keyword harvesting funnel guide covers the rules in depth, and our campaign structure guide shows how auto, broad, phrase and exact campaigns fit together.
How to Negate Search Terms That Waste Money
Negating means adding a search term as a negative keyword so the campaign stops showing ads on it. It's the other half of the routine.
Amazon's negative targeting help page and the match type documentation set two negative match types:
Negative exact blocks the term and close variants. Up to 10 words and 80 characters.
Negative phrase blocks any search containing the phrase. Up to 4 words and 80 characters.
We negate in two ways. Irrelevant terms go on the first click: if you sell dog beds and a search for "cat tree" shows up, block it now.
Relevant terms need a click threshold. A fair one is about three times your clicks per order: at a 10% conversion rate, 30 clicks with no order. Negating at 8 clicks throws away terms that would have converted.
Our negative keywords guide has the full threshold math, plus when to negate at campaign level and when at ad group level.
What Changed in 2025 and 2026
The keyword vs search term distinction still holds. What changed is how much of the search term side Amazon now decides for you.
Some "search terms" are now inferred, not typed. Amazon's search term report page (May 18, 2026) says the report now covers terms "inferred as best match to your product in non search context (such as off Amazon placements on social sites)."
The match type page adds that when no search term exists, Amazon infers one from context. Treat odd-looking terms with care before you negate them.
Prompts run on your keywords. Sponsored Products and Sponsored Brands prompts entered open beta in November 2025 and became generally available in the US on March 25, 2026. They use your existing campaign targeting and are billed per click.
Amazon picks more keywords itself. Sponsored Products keyword groups, in beta in the US, target sets of related keywords chosen from your ASINs and refreshed over time. Full-Funnel Campaigns, in open beta since September 29, 2026, hand format and budget allocation to Amazon's AI across sponsored ads, display, video and streaming TV.
Branded Conversations arrive in January 2027. Amazon announced on September 29, 2026 that Branded Conversations will turn Sponsored Brands prompts into a multi-turn shopping conversation, again using your existing campaign targeting.
Our view: the more targeting Amazon automates, the more the search term report and your negatives become the place you keep control. Read it every week.
Mistakes We See in Real Accounts
Treating the keyword as the search term. A broad keyword with a good ACOS can hide two search terms doing all the work and ten wasting money. Judge search terms, not just keywords.
Copying ad keywords into the backend field one for one. The backend field wants synonyms and variants your listing doesn't already say, not your bid list.
Harvesting without negating at the source. The term then serves from two campaigns, and your exact bid stops controlling the price.
Negative phrase on short words. "Case" as negative phrase on a phone accessories account can block "phone case with card holder", which might be your best search.
Reading the report on day one of a test. Fewer than 20 clicks rarely tells you anything about a relevant term.
Never looking at ASIN rows. Product pages show up in the search term column too. Harvest the converting ones into product targeting.
A full Amazon PPC audit checks for all of these alongside bids, budgets and structure.
How We Run Search Terms in Adbrew
Everything above works by hand in the Amazon Ads console. On one account it's an afternoon a week; across twenty accounts it's the first job that slips.
Adbrew is an agentic retail media management platform for Amazon and Walmart, and the search term routine runs inside it on rules you write and can read:
Harvesting and negation in one rule. Set the lookback, order threshold and target ACOS. Adbrew adds the converting search term as an exact keyword and the negative exact at the source in the same step.
Threshold-based negation. Negate relevant terms only after a click threshold you set, or one calculated from each campaign's own conversion rate. Brand terms can be excluded.
Search terms and keywords in one view. Filter terms across every campaign in the Sponsored Ads management workspace and act on them with bulk operations.
Agents for the reading. Ask Adbrew Intelligence for "search terms with over $40 spend and no orders in the last 60 days, excluding brand terms", review the list and approve. The same works from Claude or ChatGPT through our MCP server; nothing applies until you approve it.
Every change lands in Change History with the rule that made it. Brands on Adbrew cut ACOS by 46% on average, and Seller Presto saves more than 250 hours of manual optimization a month, much of it on exactly this work.
Bid on Keywords, Manage by Search Terms
Keywords are where you start. Search terms are where the money actually goes, so that's where the weekly work belongs.
Want harvesting and negation running on rules you control? Book a demo and we'll walk through your search term report with you.
Frequently Asked Questions
What is the difference between keywords and search terms on Amazon?
On Amazon, a keyword is the word or phrase an advertiser bids on in a Sponsored Products or Sponsored Brands campaign. A search term is the query a shopper actually typed. Match types decide which search terms each keyword can trigger, and the search term report shows which ones it did.
Are Amazon backend search terms the same as advertising keywords?
No. Backend search terms are hidden words in a listing's Generic keyword field that help the product rank in organic search, limited to under 250 bytes. Advertising keywords are bid on in the Amazon Ads console and decide where ads show. Neither affects the other directly.
What is the character limit for Amazon search terms?
Amazon limits the backend search terms (Generic keyword) field to less than 250 bytes per ASIN in the US. Spaces and punctuation are not counted. English letters and numbers are one byte each, while characters such as German umlauts can take two or more bytes.
What is the Amazon search term report?
The Amazon search term report lists the customer search terms that triggered your Sponsored Products or Sponsored Brands ads and got at least one click. It shows the keyword or target that matched each term, with clicks, spend, sales and orders, over a 65-day lookback.
Why does my search term report show ASINs instead of words?
Alphanumeric entries such as "b00ipgvvz4" in the Customer Search Term column are ASINs. They mean your ad showed on that product's detail page, which happens in automatic campaigns and product targeting. You can target converting ASINs directly or block poor ones with negative product targeting.
How do I turn a search term into a keyword?
To turn an Amazon search term into a keyword, add it as an exact match keyword in a manual campaign with its own bid, then add it as a negative exact in the auto or broad campaign where you found it. Harvest only terms that meet a rule, such as two orders in 30 days under target ACOS.
Should I use the same keywords in my listing and my ads?
Use the language that converts in both, but in different places. Search terms that convert in ads belong in your title and bullets when they fit, and synonyms or variants in the backend field. The backend field shouldn't repeat words already in the title or contain brand names.
What are negative search terms on Amazon?
Negative search terms are negative keywords: searches you tell Amazon not to show your ads on. Negative exact blocks one term and its close variants, up to 10 words. Negative phrase blocks every search containing the phrase, up to 4 words. Both are capped at 80 characters.
