.webp)
Introduction
Amazon keyword research is the process of finding the exact words shoppers type into Amazon search before they buy, then placing those words where Amazon reads them: your title, bullets, backend search terms and ad campaigns.
This guide shows the process our team uses on the Amazon accounts we manage, step by step, plus what our own ad data says about which searches actually turn into sales.
Quick answer: how to do Amazon keyword research
In short: collect the words customers use, expand them with Amazon's own suggestions and your competitors' listings, estimate volume, then let a short ad test show which searches actually sell. Put the proven words in your listing, and repeat every month.
What our ad data says about Amazon keywords
We pulled September 2026 Sponsored Products search term reports from 63 Amazon ad accounts we manage across the US, Canada, Mexico, the UK, Australia and six EU stores. Four numbers changed how we do keyword research:
The takeaway: a keyword tool gives you a starting list, but the searches that sell are found in your own search term report. That is why step 6 below is the most important one.
How Amazon search uses keywords in 2026
Amazon's search engine (often called A9 or A10) ranks products in two stages. First it decides whether your product is relevant to a search, mostly from the words in your title, bullets, description, backend search terms and product attributes. Then it decides the order, based on how well the product sells for that search: clicks, conversion rate and sales velocity.
Keyword research covers the first stage. If a word is missing from your listing, you are usually not indexed for it, no matter how good the product is. Paid ads cover the second stage, because sales from a search, paid or organic, help you rank for that search. Amazon's AI shopping assistant Rufus adds a third layer: it reads your listing and reviews to answer questions, so clear, natural language about use cases now matters as much as exact phrases.
9 steps to do Amazon keyword research
1. Build a seed list from your product and your customers
Write down 10 to 20 short phrases that describe what the product is, what it does and who it is for. Then read your reviews and your competitors' reviews and copy the words customers use. Shoppers often say "lemon squeezer" where the brand says "citrus press", and both belong on the list.
2. Expand it with the Amazon search bar
Type each seed phrase into Amazon search and note the suggestions. Then add a letter after the phrase ("citrus press a", "citrus press b") to see more. These suggestions come from real shopper searches, so they are free, current and specific to the marketplace you sell in. Do this in every country you sell in, as phrasing changes between the US and the UK.
3. Check what your competitors rank for
Pick the three to five listings that win the searches you want. Read their titles and bullets for repeated phrases, and run their ASINs through a reverse ASIN lookup in a paid tool to see which searches send them traffic. Our guide on how to find competitor keywords on Amazon covers this step in detail.
4. Use Brand Analytics if you are brand registered
Brand Analytics is Amazon's own search data and it is free with Brand Registry. Two reports matter most:
- Top Search Terms shows how popular a search is and which three products get the most clicks from it.
- Search Query Performance shows, for your own products, the searches that bring impressions, clicks, add to carts and purchases, and your share of each.
A search where you get a high share of impressions but a low share of purchases usually points to a price, image or review problem, not a keyword problem.
5. Estimate search volume with a keyword tool
Put your list into a keyword tool to estimate monthly search volume and competition. Volume numbers are estimates, so use them to rank phrases against each other, not as exact traffic forecasts. You can start with our free Amazon keyword research tool, or use paid tools such as Helium 10, Jungle Scout or Data Dive for reverse ASIN data and larger lists.
6. Test the list with an automatic and broad match campaign
This is the step most sellers skip. Run one automatic campaign and one broad match campaign with your top 20 to 30 phrases for two to four weeks. Then open the search term report and sort by orders. You will find converting searches that no tool suggested, often long phrases, misspellings, or searches in a second language.
In our data, 86% of the searches that converted in the median account were not set up as exact match keywords, and together they brought 65% of keyword sales. Every week, we move the searches with repeat orders into exact match campaigns and the listing. We explain the full routine in our guide to Amazon keyword harvesting.
7. Group and prioritize the keywords
Sort every phrase into one of four groups, so you know where it goes:
| Group | What qualifies | Where it goes |
|---|---|---|
| Main keywords | High volume, clearly relevant, converts in your ads | Title (first 80 characters) and first bullet |
| Supporting keywords | Medium volume, specific features, uses or audiences | Bullets, description and Item Highlights |
| Long tail and variants | Low volume, synonyms, misspellings, other languages | Backend search terms |
| Ads only | Competitor brand names and ASINs, gift and seasonal searches | PPC campaigns only, never in the listing |
8. Place the keywords and check they are indexed
Write the title for a human first, with the main keyword near the start. Use the bullets for supporting phrases in full sentences, and put the remaining single words in the backend field without repeats or commas (see our Amazon backend keywords guide for the 249 byte rules). Our Amazon listing optimization guide covers the full listing.
To check indexing, search Amazon for your ASIN plus the keyword (for example "B0XXXXXXXX citrus press"). If your product appears, it is indexed for that phrase. If it does not, the word is missing, misspelled or in a field Amazon is not reading.
9. Repeat every month
Search behavior changes with seasons, competitors and new product variations. Check the search term report every week or two, refresh Brand Analytics monthly, and redo the full research every quarter or when you change the product.
Long tail or short tail: what the numbers show
The common advice is that long tail keywords convert better. Our data shows something more useful: they convert about the same, but they cost less. These are median values across the accounts with enough clicks in each group (September 2026):
| Search type | Conversion rate | ACoS | Share of ad sales |
|---|---|---|---|
| Branded searches (your brand name) | 10.8% | 15% | 33% in accounts with branded traffic |
| Non branded, 1 to 2 words | 7.7% | 42% | 17% |
| Non branded, 3 words | 7.3% | 43% | 18% |
| Non branded, 4 to 5 words | 7.6% | 40% | 19% |
| Non branded, 6+ words | 8.9% | 33% | 5% |
| Competitor ASIN targets | 6.6% | 30% | 16% |
Two lessons follow. Short, broad searches are where the volume is, so they belong in the title, but in ads they are usually the most expensive way to buy a sale. Long phrases are cheap and add up: a listing that is indexed for dozens of them earns steady sales that no single keyword can match.
Free and paid Amazon keyword research tools
| Tool | Cost | Best for |
|---|---|---|
| Amazon search bar | Free | Real shopper phrasing in each marketplace |
| Search term report (Amazon Ads) | Free with ads | Finding the searches that actually sell |
| Brand Analytics | Free with Brand Registry | Search popularity and your share of clicks and purchases |
| Olifant Digital keyword tool | Free | Quick volume estimates and keyword ideas |
| Helium 10, Jungle Scout, Data Dive | Paid | Reverse ASIN lookups, large lists, rank tracking |
For a broader comparison, see our list of Amazon SEO tools.
Common Amazon keyword research mistakes
- Trusting volume over sales. A high volume search that does not convert for your product costs money in ads and adds nothing to rank.
- Stopping at the tool. In our data, one in three converting searches was first found by an automatic campaign.
- Stuffing the title. Repeating phrases hurts click through rate and can get the listing suppressed. See how to optimize Amazon titles without keyword stuffing.
- Putting competitor brand names in the listing. This breaks Amazon policy. Target them with ads instead.
- Never removing words. When a product changes, remove the old features and ingredients from the listing and backend, or you will rank for searches you can no longer satisfy.
- Ignoring wasted spend. In the median account, 47% of spend went to searches with no sale that month. Review them and add negative keywords where they are not relevant.
What to do if your ranking drops after keyword changes
First check that the listing still shows for its main keywords with the ASIN search from step 8. If it does not, an edit may have removed a word or a category change may have dropped attributes. If it is still indexed, the drop usually comes from sales: a price change, a stockout, a new competitor or lower ad spend. Our guides on how to rank #1 on Amazon and ranking drops after a stockout cover the recovery.
Frequently asked questions
How do you do keyword research on Amazon?
Start with a seed list from your product and customer reviews, expand it with Amazon search bar suggestions and competitor listings, check volume with a keyword tool, then run an automatic and broad match campaign for two to four weeks. The search term report from that campaign shows which searches actually lead to sales, and those go into your title, bullets and backend.
What is the best free Amazon keyword research tool?
The best free sources are Amazon itself: the search bar suggestions, the Sponsored Products search term report and, for brand registered sellers, Brand Analytics (Top Search Terms and Search Query Performance). For search volume estimates you can use a free tool such as the Olifant Digital Amazon keyword research tool.
How many keywords should an Amazon listing target?
Most listings have one or two main keywords in the title, five to fifteen supporting phrases in the bullets and description, and the remaining relevant words in the 249 byte backend field. In our data the median account sold through 74 different searches in one month, so the long tail matters more than any single keyword.
Are long tail keywords better on Amazon?
They are usually cheaper rather than more likely to convert. In our September 2026 data, non branded searches of six or more words had a median ACoS of 33%, against 42% for one and two word searches, while conversion rates were within about one point of each other.
How often should you redo Amazon keyword research?
Review the search term report every week or two for new converting searches, and redo the full research every quarter or whenever you launch a variation, change the formula or enter a new marketplace.
Does Amazon use Google keywords?
No. Amazon has its own search engine and its own search data, and shoppers phrase searches differently on Amazon. Google data can give you seed ideas, but volume and priority should come from Amazon sources.
Final thoughts
Good Amazon keyword research is not a one time list. It is a loop: find the words, test them with ads, keep the ones that sell and place them in the listing. Brands that run that loop every week build rankings that hold up when ad spend changes.
Olifant Digital runs this process for established Amazon brands as part of our Amazon SEO and Amazon PPC work. Every account is run by senior account managers with 7+ years of experience, and we back our work with a 60-day guarantee. If you want to see where your listings are missing sales, start with our free marketing plan.
.webp)
Alex is the founder and CEO of Olifant Digital, where his team manages over $100M in annual client revenue across 50+ e-commerce brands. He sets the agency's strategy and methodology, audits client accounts, and reviews the team's work to make sure every brand is on track to hit its growth and profitability goals. He also leads team training, so the performance data and lessons from across all accounts shape how every account manager runs their clients.
.png)
Mike reviews every Amazon article on this blog for strategic and technical accuracy before it publishes. As Director of Amazon Growth at Olifant Digital, he sets marketing strategy across client accounts and personally audits PPC at every stage of growth. He brings 8 years of daily Amazon operations across 7 and 8-figure brands including Beauty by Earth, Ekster, and Bullstrap, the kind of hands-on depth most agency directors delegate away.
