Amazon PPC Keyword Research: A Step-by-Step Guide
- sellerscaleind
- Jul 26
- 13 min read
Amazon PPC keyword research is the process of discovering the shopper searches that generate relevant clicks and profitable sales. A practical strategy uses automatic and broad-match campaigns to collect data, search term reports to identify opportunities, exact and phrase campaigns to control proven terms, and negative targeting to reduce irrelevant spend.
The objective is not to collect the largest possible keyword list. The objective is to identify the search terms that deserve more budget, the terms that need further testing and the irrelevant searches that should be blocked.
This guide explains the complete process in a simple, repeatable way.
Key Takeaways
Amazon PPC keyword research uses advertising data to identify valuable shopper searches.
Automatic and broad-match campaigns are useful for discovering how customers find a product.
Proven search terms can be moved into phrase- or exact-match manual campaigns for better control.
Negative keywords help prevent repeated spending on irrelevant or consistently unproductive searches.
Keyword decisions should be based on relevance, conversions, advertising cost and product profitability—not sales alone.
What Is Amazon PPC Keyword Research?
Amazon PPC keyword research is the process of finding, testing and organising the words or phrases that can trigger an Amazon advertisement.
Unlike traditional keyword research, Amazon PPC research does not stop with estimated search volume. It uses actual advertising data such as impressions, clicks, spend, orders and sales to determine how valuable a shopper query may be.
Amazon allows advertisers to choose keywords manually or use automatic targeting, where Amazon’s systems match the advertised product with relevant shopping queries and placements.
A useful Amazon PPC keyword strategy therefore has two functions:
Discovery: Find new searches and targeting opportunities.
Control: Allocate bids, budgets and negatives based on performance.
The purpose of Amazon PPC keyword research is not to predict every customer search. It is to build a system that continually discovers and evaluates those searches.
Amazon Keyword vs Search Term: What Is the Difference?
An Amazon keyword and an Amazon search term are related, but they are not the same.
Amazon PPC keyword definition
An Amazon PPC keyword is a word or phrase selected by the advertiser to determine which customer searches may trigger an advertisement.
For example, a seller may add stainless steel water bottle as a broad-, phrase- or exact-match keyword.
Amazon search term definition
An Amazon search term is the actual query entered by a shopper or associated with the advertisement when it receives traffic.
A keyword such as stainless steel water bottle might match searches including:
steel water bottle
insulated stainless steel bottle
stainless bottle for office
one litre steel water bottle
leakproof metal water bottle
The keyword is the advertiser’s targeting input. The search term is the customer-behaviour data produced by the campaign.
Element | Keyword | Search term |
Created by | Advertiser | Shopper or Amazon matching system |
Purpose | Controls targeting eligibility | Shows the traffic received |
Example | stainless steel bottle | insulated steel bottle for office |
Main use | Campaign setup and bidding | Analysis, harvesting and negatives |
Keyword optimisation becomes more reliable when decisions are based on real search terms rather than assumptions alone.
Where Are Amazon PPC Keywords Used?
Keyword targeting is primarily associated with Amazon sponsored advertising formats such as Sponsored Products and Sponsored Brands.
Sponsored Products
Sponsored Products ads promote individual product listings. Advertisers can select manual keyword targeting or allow Amazon to target relevant opportunities automatically. A shopper who clicks the advertisement is directed to the product detail page.
Sponsored Products are commonly used for:
Product launches
Direct sales
Keyword testing
Competitor and category targeting
Search-term discovery
Sponsored Brands
Sponsored Brands can help advertisers showcase their brand and products using formats that may include a brand logo, headline, product collection, Store or video creative.
Keyword targeting can also be used within eligible Sponsored Brands campaigns. However, campaign objectives, creative requirements and availability can vary by marketplace and advertiser eligibility.
For a beginner learning keyword harvesting, Sponsored Products usually offer the clearest starting point because automatic and manual targeting can be tested around individual advertised products.
Amazon PPC Keyword Match Types Explained
Amazon keyword match types determine how closely a shopper’s query must relate to the keyword before an advertisement becomes eligible to appear.
Amazon currently describes three main keyword match types: broad, phrase and exact.
Broad match
Broad match provides the widest discovery potential.
A broad-match keyword can match variations, related terms, synonyms and searches in which the keyword words appear in a different order. Amazon may also match based on the meaning of the keyword and the context of the advertised product.
Example keyword:
yoga mat
Possible matched searches may include:
exercise mat
yoga mats for women
thick fitness mat
non-slip workout mat
Broad match is useful when:
The product is new
Search behaviour is not fully understood
The advertiser needs more keyword ideas
The campaign has enough budget for exploration
Broad match requires regular search-term reviews because wider reach can also create irrelevant clicks.
Phrase match
Phrase match balances reach and control.
The shopper’s query generally needs to contain the keyword phrase in the same order, although other words may appear before or after it. Plural forms may also be included.
Example keyword:
yoga mat
Possible matched searches may include:
thick yoga mat
yoga mat for beginners
non-slip yoga mats
Phrase match is useful when:
A keyword theme has already shown relevance
The advertiser wants long-tail variations
Broad match is producing excessive irrelevant traffic
Exact match does not provide enough reach
Exact match
Exact match provides the highest level of keyword precision.
The shopper’s search generally needs to match the targeted keyword word for word and in the same order, although close variations such as plural forms may be included.
Example keyword:
yoga mat
Likely matches include:
yoga mat
yoga mats
Exact match is useful when:
A search term has already generated valuable sales
The advertiser wants more precise bid control
The query is strategically important
Performance needs to be measured separately
Match type | Reach | Control | Best use |
Broad | High | Lower | Discovery and expansion |
Phrase | Medium | Medium | Relevant long-tail searches |
Exact | Lower | High | Proven and priority terms |
Broad match helps discover demand, phrase match refines demand and exact match gives greater control over known demand.
Amazon advises advertisers to test multiple match types and gather adequate data before making major keyword decisions.
The Four-Step Amazon PPC Keyword Harvesting Process
The source material for this guide recommends a four-step workflow: launch discovery campaigns, analyse search terms, graduate successful searches into manual campaigns and use negatives to reduce waste.
Step 1: Launch discovery campaigns
The first stage is to collect data rather than attempt to predict every winning keyword in advance.
Two campaign types are particularly useful.
Automatic-targeting campaign
In an automatic Sponsored Products campaign, Amazon determines relevant shopping queries and product-page opportunities based on the listing and available signals.
Automatic campaigns can reveal:
Search phrases customers use
Close and loose keyword variations
Substitute product opportunities
Complementary product placements
Unexpected long-tail searches
Amazon’s own guidance recognises automatic targeting as a way to discover useful targeting opportunities before expanding into manual campaigns.
Broad-match manual campaign
Create a manual Sponsored Products campaign using a small group of highly relevant seed keywords.
Seed keywords should describe:
The product
Its category
Its main use
Its primary material or feature
Its intended audience
Its size, format or compatibility, where relevant
For example, seed terms for a lunch-box brand might include:
steel lunch box
office lunch box
leakproof tiffin box
insulated lunch container
lunch box for adults
The purpose is to learn how shoppers expand, modify or combine these ideas.
How much budget should a discovery campaign receive?
There is no universal starter budget that works across all Amazon categories.
Set the budget using:
Product selling price
Gross margin
Expected conversion rate
Average cost per click
Competitive intensity
The amount of data required
The attached source suggested a small exploratory daily budget, but the correct amount should be calculated from the product’s economics rather than copied as a fixed rule.
Step 2: Analyse the search term report
Once the campaign has received meaningful traffic, examine the search term data.
The report should help answer four questions:
Which searches generated orders?
Which searches generated profitable or strategically acceptable sales?
Which searches received clicks but no sales?
Which searches are irrelevant to the product?
Do not evaluate a search term using only one metric.
Review:
Impressions
Clicks
Click-through rate
Spend
Orders
Sales
Conversion rate
Advertising cost of sales
Return on ad spend
Relevance
Product margin
ROAS measures advertising revenue divided by advertising spend, while ACOS expresses ad spend as a percentage of attributed sales. These metrics describe advertising efficiency, but neither one automatically confirms product-level profitability.
How much data is enough?
The attached workflow recommends reviewing a longer period, such as 60 or 90 days, and allowing recent sales attribution to settle before making decisions.
That can be useful for established campaigns, but the correct review period depends on:
Sales velocity
Daily click volume
Seasonality
Promotion periods
Recent price changes
Listing changes
Stock availability
A high-volume product may produce useful data quickly. A low-volume or expensive product may require a longer testing window.
A keyword should not be labelled successful merely because it generated an order; the sale must also make economic and strategic sense.
Step 3: Move proven search terms into manual campaigns
When a search term repeatedly produces relevant sales, add it as a manual keyword.
Common choices include:
Exact match for terms that deserve precise bids and separate monitoring
Phrase match for proven phrases that may have useful long-tail variations
For example:
Discovery search term:
large steel lunch box for office
Possible harvested targets:
Exact: large steel lunch box for office
Phrase: steel lunch box for office
Exact: office steel lunch box
The attached material proposes three or more sales as a possible graduation threshold. Treat this as a practical starting heuristic, not an Amazon requirement.
A search term may deserve harvesting when it has:
Strong product relevance
More than one conversion
An acceptable ACOS or ROAS
An acceptable contribution margin
Strategic importance
Enough traffic to justify individual bid control
Should the harvested term immediately be blocked in the discovery campaign?
Not necessarily.
Moving a search term into an exact-match campaign does not guarantee that the new campaign will instantly receive the same impressions or conversions.
A cautious process is:
Add the search term to the manual campaign.
Set an appropriate bid and budget.
Observe whether it receives traffic.
Compare performance across campaigns.
Add a negative in the discovery campaign only when isolation is genuinely required.
Amazon states that advertisers do not simply “bid against themselves” by using different match types for the same keyword.
Campaign overlap can still make reporting and budget allocation harder to interpret, so structure should follow the account’s management needs.
Step 4: Add negative keywords
Negative targeting prevents advertisements from appearing for selected searches or targeting opportunities.
Amazon supports negative keyword targeting within several Sponsored Products and Sponsored Brands campaign types. Depending on the campaign, negative targets may be applied at campaign or ad-group level.
Use negative keywords for:
Clearly irrelevant searches
Searches for an incompatible product
Wrong sizes, materials or use cases
Repeated non-converting searches with sufficient data
Searches that create unprofitable sales
Terms being intentionally isolated elsewhere
Negative exact vs negative phrase
Negative exact blocks a specific query while allowing other variations to continue receiving traffic.
Example:
Negative exact:
kids lunch box
This may block that exact search while still allowing more specific, relevant variations, depending on campaign matching behaviour.
Negative phrase blocks searches containing the negative phrase and is therefore more restrictive.
Example:
Negative phrase:
for kids
This could block multiple searches containing that phrase.
Use negative phrase carefully. An overly broad negative can remove profitable traffic that has not yet been discovered.
Negative keywords should remove unwanted traffic without unnecessarily reducing useful discovery.
How to Classify Search Terms
A simple classification system makes search-term analysis easier.
Classification | Typical evidence | Recommended action |
Proven winner | Relevant, multiple sales, acceptable economics | Add to exact or phrase campaign |
Emerging opportunity | Relevant, early sales or promising engagement | Continue testing |
Expensive uncertainty | Relevant but high spend and weak conversion evidence | Reduce bid or test further |
Irrelevant | Does not match product or customer need | Add negative |
Unprofitable converter | Generates sales but exceeds viable cost | Lower bid, isolate or exclude |
Low-data term | Too few clicks or impressions | Wait before deciding |
Use break-even ACOS as a decision reference
A seller should understand the maximum advertising cost the product can support.
A simplified pre-advertising margin calculation may include:
Selling price – Amazon fees – fulfilment costs – product cost – shipping – discounts – taxes or other variable costs
If the remaining contribution before advertising is 25% of selling price, an ACOS materially above 25% may be unprofitable under that simplified model.
However, the acceptable target may differ when the seller is prioritising:
Product launch
Organic ranking
Customer acquisition
Inventory clearance
Repeat purchase
Brand defence
The campaign objective should be documented before keywords are labelled good or bad.
Amazon PPC Campaign Structure Best Practices
Keep keyword groups focused
Large ad groups containing unrelated keywords make it difficult to control bids and interpret performance.
Organise keywords around a clear theme, such as:
Generic category terms
Feature-led terms
Use-case terms
Audience terms
Competitor terms
Brand terms
Exact-match winners
The attached workflow suggests approximately 5–20 closely related keywords per ad group as a manageable starting range.
This is not a compulsory Amazon limit. The right number depends on traffic concentration and how much control the advertiser needs.
Avoid allowing one keyword to consume the whole budget
High-volume keywords may receive most of a campaign’s spend while smaller keywords collect little usable data.
Possible solutions include:
Moving major keywords into separate campaigns
Giving high-priority terms dedicated budgets
Lowering bids on dominant but inefficient keywords
Separating branded and generic targeting
Separating research and performance campaigns
Use a clear naming convention
A useful campaign name should reveal the main setup without opening the campaign.
Example structure:
SP_MAN_PRODUCT_MATCHTYPE_OBJECTIVE
Examples:
SP_MAN_YogaMat_EXACT_Performance
SP_MAN_YogaMat_BROAD_Discovery
SP_AUTO_YogaMat_Research
SP_MAN_YogaMat_PHRASE_LongTail
A consistent naming system improves filtering, reporting and handovers between team members.
Use portfolios to organise product lines
Portfolios can be used to group related campaigns and simplify account-level organisation.
Possible portfolio groupings include:
Product line
Brand
Category
Marketplace
Launch period
Business unit
Campaign naming and portfolios do not improve performance by themselves. Their value comes from making analysis and budget decisions easier.
Common Amazon PPC Keyword Research Mistakes
1. Treating every sale as a successful result
A search term can generate orders and still lose money.
Always compare advertising performance with:
Margin
Discounts
Returns
Fees
Product objective
Lifetime value, when reliably known
2. Adding negatives too early
A term with two clicks and no sale usually does not provide enough information for a confident decision.
Premature negatives can remove valuable long-tail traffic.
3. Ignoring relevance
A low-ACOS query is not automatically strategically useful if it attracts the wrong customer, creates returns or misrepresents the product.
4. Moving every query into exact match
Exact match is useful for control, but not every long-tail query needs a separate campaign.
Prioritise terms with sufficient traffic, sales or strategic importance.
5. Using the same bid for every match type
Broad, phrase and exact keywords can behave differently.
Amazon recommends testing match types with different bids rather than assuming one bid suits all targeting.
6. Forgetting the product listing
Advertising cannot fully compensate for a weak product detail page.
Before scaling keywords, review:
Product title
Main image
Price
Reviews
Availability
Delivery promise
Bullet points
Product images
A+ Content
Offer competitiveness
A keyword may be relevant while the listing itself fails to convert the traffic.
7. Optimising too frequently
Frequent major changes make it difficult to determine what caused a performance shift.
Use a consistent review schedule and document:
Bid changes
Budget changes
Keyword additions
Negative additions
Price changes
Promotions
Listing updates
Stock disruptions
A Practical Keyword Research Checklist
Before launching
Confirm the product listing is retail-ready.
Calculate contribution margin and a working ACOS target.
List the product’s category, features, materials, audience and use cases.
Create automatic and manual discovery campaigns.
Separate branded, generic and competitor themes where necessary.
Use a clear naming convention.
During analysis
Review actual search terms, not only targeted keywords.
Check relevance before checking profitability.
Identify terms with repeat sales.
Identify irrelevant and repeatedly wasteful terms.
Compare ACOS or ROAS with product economics.
Account for seasonality, promotions and recent listing changes.
During harvesting
Add proven queries as exact- or phrase-match keywords.
Give important targets suitable bids and budget.
Avoid immediate negatives unless traffic isolation is intentional.
Monitor whether the new campaign receives impressions and conversions.
Document every major change.
During optimisation
Add carefully selected negative keywords.
Reduce bids on expensive uncertain terms.
Increase control around proven high-value terms.
Keep discovery campaigns active where they continue to find useful demand.
Repeat the process regularly.
Frequently Asked Questions
What is Amazon PPC keyword research?
Amazon PPC keyword research is the process of identifying, testing and organising the words and shopper queries that can trigger Amazon advertisements. It combines initial seed-keyword planning with real campaign data, including impressions, clicks, spend, orders and sales. The goal is to discover valuable searches, target proven terms more precisely and reduce spending on irrelevant traffic.
How do you find profitable keywords for Amazon PPC?
Start with an automatic Sponsored Products campaign and a manual campaign containing closely related broad-match seed keywords. After the campaigns collect sufficient data, review the search term report for relevant queries that generate repeated sales at an economically acceptable advertising cost. Move the strongest terms into phrase- or exact-match campaigns and continue monitoring profitability.
What is the difference between broad, phrase and exact match on Amazon?
Broad match provides the widest reach and can include variations, synonyms and related queries. Phrase match generally requires the keyword words to appear in the same order, with additional words permitted around them. Exact match is the most restrictive and is best suited to searches that already justify more precise bidding and performance control.
When should you add a negative keyword on Amazon?
Add a negative keyword when a search is clearly irrelevant, repeatedly consumes spend without sufficient commercial value, attracts the wrong customer or needs to be intentionally excluded from a discovery campaign. Avoid adding negatives based on too little data. Use negative exact for narrow exclusions and negative phrase only when broader blocking is justified.
How many sales should a keyword get before moving to exact match?
There is no universal sales threshold required by Amazon. Three sales can be used as an initial heuristic, but the decision should also consider relevance, click volume, conversion rate, ACOS, margin and strategic importance. A high-value product may need fewer orders but more evaluation time, while a high-volume product may require stronger evidence.
Should automatic Amazon campaigns be paused after finding good keywords?
Automatic campaigns do not need to be paused simply because they discovered successful search terms. They can continue identifying new queries and product-targeting opportunities. Sellers should pause or reduce them only when the campaign no longer supports the account’s discovery objective, repeatedly wastes budget or duplicates traffic in a way that makes control difficult.
How often should Amazon PPC keywords be reviewed?
Keyword reviews should follow the account’s traffic level and business cycle. High-spend accounts may need frequent checks for major waste, while deeper harvesting can be completed weekly, fortnightly or monthly. The supplied workflow recommends making keyword harvesting a recurring monthly discipline, but high-volume or seasonal campaigns may require more frequent attention.
Conclusion
Amazon PPC keyword research works best as a continuous feedback loop:
Discover shopper searches.
Evaluate relevance and economics.
Harvest proven terms.
Add negatives carefully.
Monitor the result.
Repeat the process.
Automatic and broad-match campaigns help reveal demand. Phrase and exact match provide more control over proven opportunities. Negative keywords reduce repeated waste. The search term report connects these activities by showing how real shoppers interact with the campaigns.
The strongest Amazon advertising accounts are not built from one perfect keyword list. They are built through disciplined testing, measurement and refinement.
If you are an Amazon seller spending ₹25,000 or more each month on ads and need help identifying wasted spend, profitable search terms and campaign-structure problems, contact SellerScale for an Amazon PPC account audit.
Comments