Most sellers approach Amazon keyword research backwards. They find high-volume terms first, then try to force their product to rank for them.
This is expensive. And usually pointless.
A 50,000 search volume keyword means nothing if shoppers clicking that term want something you don't sell. You'll burn ad budget, tank your conversion rate, and wonder why "doing everything right" isn't working.
Real keyword research answers three questions: Is your product the right answer to this search? Do buyers on this term actually convert? Can you realistically win here?
Volume comes last. Not first.
Why Most Amazon Keyword Research Fails
The standard playbook looks like this: Pull competitors into a tool. Export their keywords. Sort by search volume. Target the big ones.
Simple. Logical. Wrong.
Here's what this approach misses:
- Search intent mismatch. "Baby spoon" might have 80,000 monthly searches. But if 70% of those shoppers want silicone spoons and you sell stainless steel, you're fighting for 30% of a market that doesn't know you exist.
- Conversion reality. Some keywords look great on paper but convert terribly for your specific product. The data exists to know this before you spend.
- Competitive position. Ranking for "wireless earbuds" when you're a new brand against Apple and Sony isn't strategy. It's wishful thinking.
High-volume keywords are not automatically high-value keywords. The distinction matters more than most sellers realize.
The OBG Amazon Keyword Research Process
We built our process around a simple principle: Find terms where you're already winning or can realistically win. Then dominate those before expanding.
This feeds directly into our PPC Lifecycle Framework. During the Launch phase, we target branded terms and exact match high-intent keywords only. No broad experimentation. No hoping for the best. Controlled aggression on terms we know convert.
Here's how we find those terms.
Step 1: Competitive Reverse Engineering with DataDive
We start with DataDive to pull the complete keyword universe for your category. Not just your competitors' keywords—their actual ranking positions, estimated traffic, and historical trends.
But we're not looking for their biggest keywords. We're looking for gaps.
Specifically:
- Keywords where top competitors rank poorly but still get sales
- Long-tail terms with lower volume but higher purchase intent
- Emerging keywords that established players haven't targeted yet
This gives us a raw list. Usually 500-2,000 terms depending on category depth. Most of these won't make the cut.
Step 2: Intent Validation Through SQP Data
Search Query Performance reports are the truth serum for Amazon keyword research.
SQP shows you—for each keyword you have impressions on—your actual click share, cart add share, and purchase share versus the market. No estimates. Real performance data from Amazon itself.
We cross-reference our DataDive list against SQP data to answer the critical question: When shoppers search this term and see your product, do they buy?
This is where most keyword lists get cut in half. Terms that looked promising based on volume reveal themselves as poor fits. Others that seemed marginal show surprisingly strong conversion rates.
We're specifically looking for keywords where your conversion rate meets or exceeds market average. These are terms where your product genuinely answers the shopper's question.
Step 3: Competitive Feasibility Analysis
Even if a keyword converts well, you need to evaluate whether you can actually win there.
We assess:
- Current organic position. Are you on page 1, page 3, or nowhere?
- Top competitor review counts. Fighting a 15,000 review listing with 50 reviews is a long game.
- Ad saturation. Some keywords have 8+ sponsored placements before organic results. The economics change.
- Brand dominance. Is this keyword owned by one major player, or is it fragmented?
Keywords where you convert well AND have a realistic path to page 1 become primary targets. Keywords where you convert well but face entrenched competition become secondary—worth defending if you're there, but not worth heavy investment to crack.
Step 4: Tiered Keyword Classification
Every keyword that survives our process gets classified:
Tier 1 — Defend and Dominate: High conversion rate, achievable ranking, meaningful volume. These get exact match campaigns, aggressive bids, and ongoing organic optimization.
Tier 2 — Test and Validate: Promising indicators but unproven. We run controlled tests with capped budgets before committing.
Tier 3 — Monitor Only: Relevant to your product but currently out of reach. We track these for future opportunities as your authority grows.
This classification directly informs our PPC structure. Tier 1 terms get priority budget. Tier 2 terms get test budgets. Tier 3 terms stay off the active campaign list entirely.
Real Example: How This Works in Practice
When we started working with NumNum Baby, the obvious move was targeting high-volume baby feeding keywords. "Baby spoons" alone has massive search volume.
But our research showed something different. NumNum's pre-spoon design—specifically built for self-feeding babies learning to eat—converted exceptionally well on more specific terms. "Baby led weaning spoon." "Self-feeding baby utensils." "Pre-spoon for infants."
Lower volume. Much higher intent match.
We dominated those terms first. Built review velocity and organic rank on keywords where NumNum was genuinely the best answer. Then expanded outward as the listing's authority grew.
Result: 30x revenue growth in 18 months, from $100K to $3M annual run rate. That trajectory led directly to an 8-figure exit for founder Doug Gonterman.
It started with keyword research that prioritized fit over volume.
Common Amazon Keyword Research Mistakes We See
Chasing competitor keywords blindly. Your competitor's top keywords work for their product, their reviews, their price point. They might not work for you.
Ignoring SQP data. If you have Brand Registry, SQP reports are free intelligence most sellers never check. You're making decisions in the dark.
Treating all keywords equally. Flat campaign structures with the same bids across 200 keywords waste budget on losers while underfunding winners.
Researching once and forgetting. Keyword landscapes shift. Competitors enter. Trends change. Quarterly research refreshes aren't optional.
Optimizing for traffic instead of profit. Revenue is vanity. Contribution margin is sanity. A keyword driving $50K in sales at 40% TACoS is worse than one driving $20K at 10% TACoS.
The TACoS Connection
Every keyword decision ultimately flows to one metric: Total Advertising Cost of Sales.
TACoS = total ad spend ÷ total revenue. It's the only metric that tells you whether your advertising is actually building a profitable business or just buying sales you'll lose the moment you stop spending.
Our keyword research process optimizes for TACoS from the start. By targeting terms where you naturally convert well, you need less ad spend to generate the same sales. Your organic rank builds faster because Amazon's algorithm rewards conversion rate. Your TACoS drops as organic sales increase.
This is why we reject the "target everything and let the algorithm sort it out" approach. It works for Amazon—they get paid either way. It doesn't work for brands trying to build profitable channel economics.
Tools We Use and Why
DataDive for initial keyword discovery and competitive intelligence. Gives us the broadest possible view of category keyword opportunities.
Search Query Performance (SQP) for conversion validation. Amazon's first-party data beats any third-party estimate.
Datarova for ongoing digital shelf monitoring. Tracks rank changes and competitive movement over time.
The tools matter less than the process. You could do this manually if you had unlimited time. You don't. So we systematize it.
Work With OBG
If you want to see how this would work for your brand, book a free strategy session. We'll audit your account, identify the fastest wins, and map out exactly how we'll execute. And if we don't increase your profitability in the first 30 days, you don't pay. Zero risk.