Amazon PPC Keyword Research: A 2026 Step-by-Step Guide

Amazon PPC keyword research is the process of finding the searches your product should advertise on, deciding what each one is worth, and turning that list into campaigns. Done well, it decides most of what an account spends in its first six months.

The old version of this post opened with third-party tools. This one starts where we start on the accounts we manage: with the data Amazon already gives every advertiser, much of it free, and how to move it into campaigns.

What Is Amazon PPC Keyword Research?

Amazon PPC keyword research means building a list of shopper searches to target in Sponsored Products and Sponsored Brands, grouped by intent and ranked by what each search is likely to earn. It answers three questions: which searches are relevant, which ones are affordable, and which campaign each one belongs in.

It differs from Google keyword research in one big way. On Amazon almost every search is a shopping search, so volume matters less than whether the shopper buys a product like yours.

Amazon's own keyword targeting guide describes the core loop: start from your listing, run automatic targeting for a couple of weeks, then use the search term report to find the terms that sell and move them into manual campaigns.

If the difference between a keyword you bid on and the search a shopper typed is still fuzzy, our guide to Amazon keywords vs search terms explains it with examples.

What Changed in Amazon Keyword Research in 2026

The method has not changed much. The tools around it have, and several changes affect where the data lives and which searches your keywords now serve on.

Two of these matter most for research. Search terms are getting longer and more conversational, and Amazon is matching more of them for you through broad match, keyword groups and prompts. Both make the search term report more important than any tool's volume estimate.

The Amazon Sources for PPC Keyword Research

We use eight sources, all from Amazon. Each one answers a different question, and none of them gives you everything.

1. Start With Your Listing, Reviews and Product Facts

Write down what the product is, what it's made of, who it's for and what problem it solves. Then read your reviews and your competitors' reviews for the words shoppers use, which are rarely the words in a spec sheet.

Amazon's keyword strategy guide suggests the same start: keywords that define your product's characteristics, materials or benefits, plus your brand name and the terms already on your detail page.

The FAQ adds a practical warning. Keywords have to match what's on your detail page: an ad for bath towels won't get impressions on "beach towels" if nothing on the listing says beach.

2. Type Into the Amazon Search Bar

Amazon's autocomplete is the cheapest volume signal there is. Type your seed term, then the seed plus each letter of the alphabet, and write down every suggestion that fits.

Amazon's own research on its autocomplete says suggestions are usually ranked on customer behavior signals such as clicks and conversion rates. So a suggestion is a query real shoppers type and act on, though Amazon gives no number for it.

Do it logged out or in a private window, on desktop and in the app, and in each marketplace you sell in. Suggestions differ by device and country, and personalization skews them.

3. Use Keyword Suggestions and Suggested Bids in the Console

When you create a manual campaign, the console suggests keywords for the products in the ad group, each with a suggested bid by match type. Amazon's Sponsored Products targeting guide says these recommendations are tailored to your products.

Since the May 2025 switch to theme-based bid suggestions, the bid comes as low, median and high values. The API documentation says they're "calculated from a group of recent winning bids for similar products," with impact metrics showing weekly clicks and orders similar products got at each bid.

Treat both as a starting list and a price check, not as a plan. Suggested bids describe what it costs to win the auction, not what you can afford, and later we show how to check the two against each other.

4. Run an Auto Campaign as a Discovery Engine

An automatic campaign lets Amazon pick the searches and product pages, based on your listing. It uses four targeting groups: close match, loose match, substitutes and complements, each with its own bid.

That makes it the best source of searches you would never have thought of. We launch one per product group with a modest budget and leave it running permanently, because new searches keep appearing.

Bid close match highest, since it converts best for most products, and loose match lowest. Our Amazon PPC match types guide covers each targeting group with Amazon's examples.

5. Mine the Search Term Report

The search term report is where research becomes evidence. Amazon's documentation says it lists every search that got at least one ad click, with spend and orders, over a 65-day lookback.

In 2026 it also includes inferred terms: queries Amazon matched to your product in non-search placements, such as off-Amazon placements on social sites. Read those rows with care, because no shopper typed them into Amazon.

Sort by orders first, then by spend with zero orders. The first list is your harvest, the second your negatives. Our search terms report guide shows how to build it in unified reporting and run n-grams on it.

6. Check Top Search Terms in Brand Analytics

If you're brand registered, Brand Analytics has a report Amazon calls the Amazon Search Terms report, which most sellers know as Top Search Terms. It ranks the most popular searches in the store by search frequency rank and shows the top three products clicked for each.

For each of those three products it shows click share and conversion share. That's how you see who owns a search: if one competitor takes 40% of clicks on your best category term, you know what you are bidding against.

Search by your seed word to find related terms with their rank, or by a competitor's ASIN to find the searches where that product wins clicks. Our Brand Analytics guide covers the other dashboards.

7. Read Search Query Performance

The Search Query Performance dashboard shows your brand's top queries, up to 1,000 in the brand view, with query volume, impressions, clicks, cart adds and purchases. It covers organic and Sponsored Products results on the search results page.

It is the only Amazon source that shows total search volume next to your share of each funnel step. A query where you get 8% of impressions but 1% of clicks is a listing problem; one where you convert well but get few impressions is a keyword to bid on.

Amazon says its numbers aren't meant to match your ad reports, so use it for direction, not reconciliation. Our Search Query Performance guide walks through each column.

8. Test Keyword Groups (Beta)

Keyword groups are a Sponsored Products targeting option that builds sets of related keywords from your ASINs, such as "Keywords related to your brand", and "continuously refreshes them over time." Amazon describes them as simplified broad-match keyword management.

They're in beta and US only, announced in the Amazon Ads API release notes. You can mix them with normal keywords in an ad group, but not with product targets.

We treat them like an auto campaign: a discovery source with its own capped budget, harvested weekly into exact. One more Amazon tool sits near this list: Product Opportunity Explorer groups search terms into niches with volume trends, which helps before a launch.

How to Build a Keyword Map by Intent

A keyword list becomes useful once every term is sorted by intent. Shoppers who type your brand name behave nothing like shoppers who type the category, so they need different bids, match types and goals.

Amazon's guide sorts keywords into generic, long-tail, branded and seasonal. We use four groups, because competitor terms behave differently enough to need their own.

Add a seasonal tag to any term that peaks at a time of year, such as "back to school water bottle". Seasonal terms sit in the group they belong to, with their own budget for the weeks that matter.

Branded Keywords

Branded terms include your brand name or product line. They convert best and cost least, and Amazon's targeting guide recommends bidding on them as brand defense, because competitors may be bidding on them too.

Keep them in their own campaign. If branded sales sit inside a category campaign, they make its ACOS look far better than the category work really is.

Category Keywords

Category terms describe the product without a brand: "insulated water bottle", "stainless steel tumbler". They carry the volume and most of your new customers, and they're the most expensive clicks in the account.

This is where Search Query Performance and Top Search Terms earn their place. They show which category terms are big enough to fight for and who currently wins them.

Competitor Keywords

Competitor terms are another brand's name, with or without a product. They convert poorly unless you beat that product on price, rating or a feature the shopper can see in the ad.

Run them in exact match in a separate campaign, judge them on new-to-brand orders rather than ACOS, and drop rivals you can't beat on the search results page.

Long-Tail Keywords

Long-tail terms are specific searches of three or more words: size, material, use case, audience. Amazon's guide says they "generally have lower search volume but higher relevance."

In 2026 there are more of them. Since shoppers can type questions straight into the search bar, phrases like "water bottle that keeps ice all day for hiking" show up in search term reports, and phrase match picks them up well.

How to Prioritize Amazon PPC Keywords

A good map has hundreds of terms. You can't fund them all on day one, so rank them on four tests, in this order.

1. Relevance. Would a shopper typing this be happy to find your product? If not, it isn't a keyword, it's a negative. No volume fixes a relevance problem.

2. Demand. Search frequency rank in Top Search Terms or query volume in Search Query Performance. Without Brand Registry, autocomplete and impressions in your auto campaign are the proxy.

3. Proof. Conversions in your own search term report, or purchase share in Search Query Performance. One proven order beats any estimate.

4. Affordability. Can the term pay for its clicks at your target ACOS? This is the test most sellers skip.

The Affordability Math

The highest cost per click (CPC) a keyword can carry is your price times its conversion rate times your target ACOS (advertising cost of sales). At a $30 price, a 10% conversion rate and a 30% target, that's $30 x 0.10 x 0.30 = $0.90 per click.

Compare that with Amazon's suggested bid range. If your maximum sits inside or above the range, bid. If it sits well below, the term can't pay at your target yet, so test it in phrase at your maximum or wait until reviews lift your conversion rate.

Use the conversion rate from your own search term report where you have one. For a new product, start from the ad group average and replace it as data comes in.

Target ACOS itself comes from your margin. Our guide to break-even ACOS shows how to set it per product.

Sort the List Into Tiers

The four tests give each term one of four tiers. These decide where the term goes next.

How to Move Keyword Research Into Campaigns

Research only pays when it reaches a campaign with the right structure. This is the order we follow on a new product.

1. Build Campaigns Around Intent Groups

One campaign per intent group per product group: branded, category, competitor, plus one discovery campaign with auto and broad. Separate campaigns give each group its own budget, so a hungry category term can't starve your brand defense.

Our guide to structuring Amazon PPC campaigns covers naming and how many products to put in an ad group.

2. Match Type by Tier

Scale-tier terms go in exact. Prove-tier terms go in phrase, which lets modifiers through while keeping the core meaning. Discovery belongs to auto, broad and keyword groups, which the match types guide explains in depth.

Amazon's guidance is to bid highest on exact, lower on phrase and lowest on broad. That matches how the three convert in nearly every account we see.

3. Set Starting Bids From Your Maximum, Not Amazon's

Start each keyword at the lower of your maximum CPC and Amazon's median suggestion. Amazon's keyword strategy guide suggests at least 25 keywords and about two weeks of data before you start optimizing.

Two weeks is the floor, not the target. Most category terms need more clicks than that before a zero-order result means anything.

4. Add Negatives So Terms Don't Compete

When the same search can trigger your exact keyword and your auto campaign, you lose control of what you pay for it. Negate each exact keyword as negative exact in the discovery campaigns.

Add the obvious irrelevant terms on day one as well: "used", "replacement lid", sizes you don't sell. Our negative keywords guide covers click thresholds and campaign vs ad group levels.

5. Harvest Every Week

Each week, take converting search terms from discovery campaigns, add them as exact keywords, and negate them at the source. That loop is how research keeps running after launch.

We use one or two orders over 14 to 30 days as the bar for a new exact keyword, adjusted for price. Our guide to automating a keyword harvesting funnel shows the full routine.

Organic Keyword Research: Amazon SEO vs Google

Ads and organic search use the same words, so one research process should feed both. This section replaces our older post on Amazon SEO keyword research.

The main difference from Google is intent. Google serves a mix of informational and buying searches; Amazon is almost entirely transactional, so you research what people buy, not what they want to learn.

Where Keywords Go on a Listing in 2026

Space is tighter than it was. Since July 27, 2026, US titles outside media categories are capped at 75 characters, and Amazon may rewrite longer ones; the new Item Highlights field adds 125 searchable characters.

Put the single most important category term in the title. Use Item Highlights for the materials and use cases cut from it, bullets for features in shopper language, and backend search terms for synonyms that fit nowhere visible.

The backend field is limited to less than 250 bytes and bans brand names and ASINs. Our Amazon backend keywords guide and product titles guide cover both rules.

Let Ads Feed Organic

Your search term report is the best organic research you own. A search that converts in your ads is proven shopper language, and if it isn't in your listing, Amazon has less reason to show you there organically.

Every month we take the top converting search terms not already on the listing and place them, then watch Search Query Performance for organic impressions on those queries over the next four to eight weeks. Amazon's dashboard help says this is exactly what it's for.

Mistakes We See in Real Accounts

• Bidding on volume, not relevance. A big category term that doesn't describe your product spends fast and converts slowly.

• One campaign for everything. Branded sales hide category losses, and nobody can tell what's working.

• Taking the suggested bid as the bid. It describes the auction, not your margin. Check it against price x conversion rate x target ACOS.

• Researching once. Searches change with seasons and with how shoppers use Alexa for Shopping. Discovery campaigns should run permanently.

• Harvesting without negating. If a term is in exact but not negated in auto, both campaigns bid on it.

• Ignoring Brand Analytics. Many brand-registered sellers never open Search Query Performance, the only place Amazon shows your share of each query.

• Treating inferred terms as searches. Rows from off-Amazon placements are Amazon's inference, not typed queries. Don't move them into listings unchecked.

• Competitor terms judged on ACOS. They rarely look efficient; judge them on new-to-brand customers and share.

How We Run Keyword Research in Adbrew

Every step above can be done by hand in Amazon Ads Agent and Seller Central. On a catalog of hundreds of ASINs, the weekly harvest alone takes most of a day.

Adbrew is an agentic retail media management platform for Amazon and Walmart, and the research loop runs inside it:

• Automated keyword harvesting. Converting search terms move from auto and broad into exact and are negated at the source, on rules you set for harvesting, negation, bids and budgets. Learn more about Sponsored Ads management.

• Bulk actions. Add hundreds of researched keywords across campaigns, with match types and bids, in one action.

• Adbrew Intelligence. Ask "which converting search terms from the last 60 days aren't targeted in exact yet?" and the agent builds the list from your data.

• Rank and share of voice. Digital Shelf Analytics tracks organic and sponsored rank on the keywords in your map, hourly or daily.

This is the Search Rank and SOV view. It shows total and sponsored share of voice on the keywords you track, and which brands hold the rest.

Once a keyword is in your map, this is how you see whether bidding on it moved your organic position. Our guide to Amazon share of voice explains how to read it. Brands on Adbrew cut ACOS by 46% on average.

Start With the Data You Already Have

Most accounts already hold the answer in a search term report nobody has sorted. Begin there, build the map, and let the weekly harvest do the rest.

Want to see harvesting rules, keyword rank tracking and Adbrew Intelligence on your own account? Book a demo with our team and we'll walk through your keyword map with you.