Posting this because I see so many people struggling with messy keyword structuring, random expansion, and fragmented data in the first month. This is the white-hat, non-lowball, non-fluff strategy I’ve used on multiple new launches to get Amazon to recognize the product, validate conversion, and scale without guessing.
My Product & Context
-
Private label, 100% white hat, no price-war launches
-
Clear direct competitors (similar design, function, material, price within ±$2)
-
Goal: Build a scalable, repeatable ad system — not luck-based orders
What I Observed From Top Competitors
I dug into the top 8 organic ranking listings for my main terms:
-
Most of their driving keywords are 100–300 search volume, not huge high-volume terms
-
Many top sellers hit thousands of monthly orders off clusters of these mid-tail & long-tail keywords
-
Lesson: Sales don’t depend on big keywords — they depend on high-relevance small-to-mid keyword groups
My Actual Launch Ad Strategy (Week by Week)
Week 1 Goal: Teach Amazon What I Am + Prove I Can Convert
This is not about ACoS. It’s about giving Amazon clean data so it stops showing me random traffic.
Setup I Use
-
3 Manual Exact campaigns
-
Grouped strictly by intent / use case / audience — NOT search volume
-
Example groups: core function, usage scenario, target demographic
-
3–5 highly similar keywords per group, total ~9–15 keywords
-
Why? Search volume grouping does almost nothing for product segmentation. Intent grouping tells Amazon exactly who to show me to.
-
1 Auto campaign for learning and discovery
Week 1 Settings (Non-Negotiable)
-
Bidding: Fixed bids across the group (no individual tweaks yet)
-
Set to ~80% of top suggested bid
-
Placement adjustments: NONE
-
New listings don’t need top-of-search pressure yet — we just need valid clicks
-
Budget split: Enough to get meaningful clicks, not waste
Week 2: Expand & Capture Search Terms
-
Keep existing Exact campaigns running
-
Launch 1 Broad campaign using top performing roots from Week 1
-
Add 1 new small Exact group for confirmed high-relevance terms
-
Auto campaign still running for search term mining
Weeks 3–4: Promote Winners & Kill Junk
This is where most sellers mess up: they promote every 1-click, 1-order keyword.
My Hard Rules for Keyword Promotion (Exact Match Lift Criteria)
I only move a search term from Auto/Broad to Manual Exact if:
-
≥15 clicks in Auto/Broad
-
Conversion rate ≥15%
-
Clearly matches my product’s use case (no accidental matches)
Anything below that is luck, not signal. I log it as potential but don’t scale it.
How I Handle The Two Most Annoying Keyword Types
1. Core Big Keywords
-
Convert poorly in Exact
-
Get random orders in Auto/Broad
-
What I do:
-
NO Exact campaigns for these
-
Run them in Phrase or low-bid Broad
-
Only for leftover traffic, not main conversion
2. High-Relevance Low-Search-Volume Keywords
-
Low search, low individual clicks
-
But convert well when grouped
-
What I do:
-
Group 3–8 similar ones together in one Exact campaign
-
Do NOT run single-keyword campaigns for these — they won’t spend budget
-
This becomes my stable, low-ACoS order base
My Actual Bidding & Placement Rules When Scaling
-
When moving a term to Exact:
-
Bid = Auto CPC + 10–20%
-
Still fixed bid initially
-
Only add Top-of-Search adjustments (+20–40% max) once the group hits 5+ orders/week
-
No placement boosts in Week 1–2 — waste of budget
How I Clean Up Fragmented Data
So many terms with 1–3 clicks and 1 order. Here’s what I actually do:
-
Leave them in Auto/Broad for more data
-
No promotion until click threshold is hit
-
Batch negative keywords once per week
-
Instant negatives only for clearly irrelevant terms
My Real Goal With Keyword Expansion
Not to collect as many keywords as possible.
It’s to:
-
Lock in 1–2 high-converting intent groups
-
Rank those organically
-
Let Amazon’s algorithm feed more related keywords automatically
-
Repeat until I have a self-sustaining keyword matrix
Big organic head terms almost always come after you own the long-tail.
If You’re Struggling Right Now
Check these:
-
You’re grouping by search volume, not intent
-
You’re promoting low-data keywords too early
-
You’re fighting big core keywords too soon
-
You’re using dynamic bids when Amazon has no data on your listing
-
You’re over-adjusting bids every day
This framework is boring, consistent, and works across home goods, gadgets, and apparel. No hacks, no secrets — just structured decision making.
Curious what others use for their first 30 days. Let’s discuss.
Answers (6)
Ads are just a tool. You can’t fix a low-conversion listing with ad structure.
But your framework is perfect for:
My only addition:
If a campaign is only getting 1–3 clicks a day, it’s not testing — it’s just sitting there.
Either increase bid/budget or replace the keywords.
People overcomplicate Amazon ads so much.
Expansion isn’t about finding more keywords.
It’s about dominating 2–3 high-converting intent groups, ranking them organically, and letting Amazon feed you more related traffic.
I build “long-tail clusters” and scale those. Once those rank, the big organic terms come by themselves.
Your strategy is exactly how you build sustainable sales without fighting the top dogs.
Your confusion points are exactly where 90% of new sellers mess up.
Week 1 placement boosts = waste.
You don’t need top of search when you have 0 reviews.
Bidding:
Big keywords converting poorly in Exact? Normal. They’re dominated by top brands. Auto/Broad orders are often matching long-tail variations, not the head term itself.
Really wish I saw this when I started.
Grouping by theme/scenario/use case changes everything.
Two quick practical tips:
Also, you don’t need huge search terms to win. I have a product doing 2k+/mo off a big pool of 100–300 search volume terms.
Great post. I tested almost exactly this structure with a similar product.
One thing I noticed:
New listings need volume first, not hyper-precision.
If you only target tiny low-search keywords in week 1, you won’t get enough sessions, and Amazon won’t push you to recommended placements.
I run one core root keyword in Broad early, plus some ASIN targets, to pile up clicks and data. Then we tighten over time.
Amazon rewards listings that generate orders more than ones that are “super targeted but dead.”