Most Amazon sellers do keyword research the same way: pull thousands of keywords from a tool, dump them into campaigns, and end up wasting a ton of ad spend. The fix is not more keywords. It is separating your keywords by buyer intent before they ever touch a campaign, so each campaign targets one specific shopper motivation.
This process is built for a new product launch, or a relaunch of a product that is dead or has been unranked for a long time. (Mining an existing, live product’s own search-term reports is a separate process.)
What You Are Building: A Buyer-Intent Master List
The end result is a keyword research master list for your product. It contains:
- Your single-keyword (alpha) campaigns
- Keywords segregated into each buyer-intent group
- Competitor keywords and your own branded keywords
- Negative keywords you know do not apply
- Keywords that are better for driving traffic to a storefront than to the product
Once you have this, you have the raw material for your Sponsored Products campaigns. The tool used throughout is Helium 10 (its Cerebro feature), through the Chrome extension. You can sign up free at helium10.com, and any similar third-party tool will work the same way.
This is how you do keyword research where you can segregate your keywords by buyer intent, so you can create campaigns centered around a specific intent your shoppers have.
The Keyword Research Process, Step by Step
Step 1: Pull the Organic Page-One Competitors
Type your main keyword into Amazon the way a shopper would (in the example, “welcome mat”). Open the Helium 10 Chrome extension and click X-ray. Then click filters, turn on “hide sponsored products from results,” and apply, so you see only the products ranking organically on page one, not the ads.
Step 2: Select Nine Competitors Like Your Product
Go through the organic results and select nine competitors that are the same type of product and the same offer as yours, ranking as high as possible. This matters: a utility mat for businesses is not the same as a residential welcome mat, even if both rank for the keyword. Only pick true matches to your product.
Step 3: Run Cerebro and Filter Down to ~150 Keywords
Click Run Cerebro. This pulls every keyword those nine products rank for. Set match type to Organic (you want organic rankings, not sponsored). Before filtering, that is around 8,000 keywords, which is too many.
Do not dump 8,000 keywords into your campaigns. That assumes you are smarter than Amazon and gives the platform no leeway to find long-tail keywords itself, which in today’s Amazon PPC you want it to do. Instead, aim for the 50 to 150 most important, most relevant, highest-traffic keywords. Apply these filters:
- Minimum search volume: 500
- Minimum competitors: 3 (at least three of your nine rank for it)
- Competitor rank max: 50
- Match type: organic
In the example, that leaves 144 terms – a manageable number. Export the data as a CSV, import it into Google Sheets, move the search-volume column next to the keyword, and sort by search volume Z to A.
Step 4: Pick Your 5 Alpha Keywords
Add a column to the left called “alpha keywords.” Choose five keywords with a strong mix of high search volume and high relevance. Skip terms that are too broad (in the example, “doormat” covers too many unrelated products) and skip near-duplicates of ones you already picked. Mark each chosen keyword with an X so you can filter them out easily.
These five become your single-keyword exact-match campaigns for ranking. If you have a lower budget, you may need to niche down to lower-volume, higher-relevance terms, and in some cases tailor your images, primary image, and listing copy to that more specific offer.
Step 5: Build Your Phrase Negations
Scan the list for terms that clearly do not apply to your product. In the example, the mat is outdoor, so “indoor” and “inside” are negated, along with “folding” (it does not fold), “foam” (not the material), “kitten,” “kitty,” “puppy,” and “yoga.” You should find maybe a dozen or two. These go in a phrase negations list so Amazon does not waste your ad spend showing you for uses that do not fit.
Step 6: Segregate Keywords by Buyer Intent (The Key Step)
This is the part most keyword research skips. Group the remaining keywords by shopper intent. In the example, patterns emerge around size (large), style (funny, cute), durability (heavy duty, waterproof), and material (coir, which is coconut husk).
Rather than sort every keyword by hand, use AI:
- Remove your alpha keywords, then copy the keyword and search-volume columns into a new sheet and download it as a CSV.
- In ChatGPT, upload the CSV and ask it to determine the top keyword families based on the different buyer intents in the list. (In the example it returned seven groups, including outdoor placement, indoor placement, durability and size, decorative aesthetic, natural material, and funny/personalized.)
- The critical second prompt: ask it to output a CSV of the keywords with their search volume, segregated by shopper intent, with each keyword appearing only once. Give it a priority order (most specific to least specific) so keywords that fit two groups land in the right one.
Import that list back into your sheet and add a filter so you can view one intent group at a time.
Step 7: Build the Campaigns
For each keyword group, create a Sponsored Products campaign: select the product, click manual targeting, then keyword targeting, then enter list. Unselect phrase and broad, keep exact selected, paste your list, and click add keywords. Set your bid (use Amazon’s suggested bid only if you have to; calculating your own from sales data is better), and select down-only bidding.
Placement adjusters are optional; if you already know a placement works or you want to rank for top of search, add a small adjustment, no more than 10 to 20 percent, 30 percent at most. Set a daily budget (the example uses $20) and launch.
You can also run broad campaigns with these keywords. If you do, add your phrase negations under negative keyword targeting as negative phrase. That keeps Amazon free to find new high-performing terms while blocking the ones you already know do not fit.
Why Buyer-Intent Grouping Matters
When each campaign is built around one intent group, you can read that campaign’s click-through rate, conversion rate, and impressions and use them to diagnose your listing. If your “decorative aesthetic” campaign underperforms your others, it signals you are not communicating well to that buyer, which points you to specific assets to fix: your thumbnail, primary image, and title first, then your secondary images, A+ Content, Brand Story, top-voted reviews, and listing copy. That feedback loop is what most sellers never take the time to build.
You can also pull out keywords that signal browsing intent (for example, “home decor products” or “front door products”). Those are better sent to a storefront through a Sponsored Brand campaign than to a single product through a Sponsored Product campaign.
Frequently Asked Questions
Amazon PPC keyword research is the process of finding and organizing the keywords you will target in your Sponsored Products campaigns. Done well, it produces a master list of the 50 to 150 most relevant, highest-traffic keywords, grouped by buyer intent, plus alpha keywords, negatives, and competitor and branded terms.
Pull competitor keywords with Helium 10 Cerebro, filter down to the most relevant terms, then group them by the motivation behind each search (size, style, durability, material, and so on). You can use ChatGPT to segregate the list by intent, keeping each keyword in a single group using a most-specific-to-least-specific priority order.
Aim for roughly 50 to 150 of the most important, most relevant, and highest-traffic keywords. Dumping thousands of keywords into campaigns is inefficient and leaves Amazon no room to discover long-tail terms on its own.
The video uses Helium 10 and its Cerebro feature through the Chrome extension (free signup at helium10.com), plus ChatGPT to help group keywords by intent. Any similar third-party keyword tool follows the same process.
Alpha keywords are the five highest-value terms you pick from your list, with a strong mix of high search volume and high relevance. Each one gets its own single-keyword exact-match campaign built for ranking.
Key Takeaways
- Start keyword research from the nine organic page-one competitors that match your product and offer
- Filter Cerebro down to 50 to 150 keywords rather than dumping thousands into campaigns
- Pick five alpha keywords for single-keyword exact-match ranking campaigns
- Grouping keywords by buyer intent lets campaign performance tell you which listing assets to fix
- Add phrase negations to broad campaigns so Amazon can still find new terms without wasting spend
Related Guides
- Amazon PPC Targeting Guide: Keywords, Products and Audiences
- Amazon PPC Match Types: Broad, Phrase, Exact and the 4th Type
- 12 Amazon PPC Mistakes We Find in Almost Every Account Audit
- The Amazon PPC Tools We Actually Use
Ready to Run Profitable Amazon PPC?
At Sophie Society, my team and I run a more advanced version of this exact process across 800+ brands and $5M/Month in ad spend. If you want us running your campaigns, apply to work with us.
Free Resources
- The Profitable PPC Challenge: My live Amazon PPC training event
- Sophie Society case studies: Real partner results across launches and PPC
- My full keyword research walkthrough: The free step-by-step video this guide is based on, including the keyword master-list template
Based on my video: “Amazon PPC KEYWORD RESEARCH – Ultimate 2026 Step by Step Guide”.
