31:41
Amazon SEO
Andrew, an Amazon brand operator now working with Helium 10, argues that his $100k-to-$7M growth at Touch of Class and NFPA's 600%+ Amazon sales growth came from disciplined Helium 10 keyword research fused with an SEO-driven Amazon Brand Store strategy, and that sellers must now optimize simultaneously for Amazon's A9 search algorithm and its Rufus AI assistant — which he argues share the same backend — using intent- and use-case-based listing writing rather than pure keyword stuffing.
Andrew took Touch of Class's Amazon revenue from $100,000/year to $7 million/year using Helium 10 tools (Cerebro, Magnet, Scribbles) across roughly 4,000 parent ASINs.
He built the Amazon Brand Store into an SEO content hub that ranked in Google, generating $1 million in sales in his final year there from that page alone.
His method for the Brand Store was cross-referencing outside Google search terms against Helium 10-identified transactional Amazon keywords, then building ~10 blog-style pages on that basis.
Core belief: SEO and PPC are 'in symbiotic union' — optimizing the listing before running PPC lifts organic rank, while running PPC before optimization causes a boom-bust cycle once ads are turned off.
He relied on a broad, redundant SKU portfolio (400-500 core products drove ~50% of sales, thousands of long-tail SKUs made up the rest) rather than a few hero products, which is why he leaned heavily on Helium 10's Keyword Tracker.
He currently works with NFPA (National Fire Protection Association), selling code books, where Amazon sales are up over 600% year-over-year.
He argues Amazon's Rufus AI assistant and the A9 search algorithm are intertwined: Rufus reportedly uses A9 to bring in products, builds 'search query plans,' and pulls from the product Search API — so sellers must optimize for both together, not one in isolation.
A 'bat toilet rug' case study (Bradley's own product) is used as evidence: sending one or two purchases against a target keyword acted as a relevancy signal that shifted both A9 search ranking and Rufus's answer to the same query almost simultaneously.
Because Rufus now surfaces price history to shoppers, pricing strategy is framed as increasingly an ethical/trust issue — sellers can't gouge customers around discounts or Prime Day pricing without it becoming visible.
He advocates feature-to-benefit and use-case/intent-based listing writing (not pure keyword stuffing) as what both converts customers and aligns with how Rufus surfaces products.
He built a custom Claude skill called 'intent mapping' that classifies a keyword list by intent type, and recommends structuring PPC campaigns around those intent themes rather than flat keyword lists.
He recommends uploading search term reports and product information directly to ChatGPT or Claude, using reasoning/extended-thinking mode, instead of manually sifting through search term reports.
He treats Helium 10's 'Keyword Sales' metric (distinct from standard search term reports) as a major differentiator for deciding what to bid on in PPC.
FBA (Fulfillment by Amazon) — Amazon's program letting sellers store inventory in Amazon warehouses for Amazon to pick, pack, and ship, which Andrew scaled from 1,000 to 50,000 sq ft at Touch of Class. Apply: Start with a small dedicated FBA footprint and expand warehouse space allocated to it as the Amazon sales channel proves itself.
Freedom Ticket — Helium 10's Amazon training course, taught by Bradley Sutton and Kevin King, where Andrew first learned Amazon selling fundamentals. Apply: Use it to learn keyword research, listing optimization, Seller Central operations, and FBA basics, and to access Andrew's own AI trainings.
Cerebro — Helium 10's reverse-ASIN keyword research tool that Andrew ran repeatedly and 'obsessively' to build keyword lists for Touch of Class products. Apply: Run it against competitor ASINs to surface the keywords driving their traffic, then feed those into category-, product-, and broad-level keyword lists.
Magnet — Helium 10's keyword research tool used alongside Cerebro to build product and category keyword sets. Apply: Use it to expand seed keywords into broader search-term coverage for a given product category.
Scribbles — Helium 10's listing optimization tool used to place researched keywords into titles, bullets, and descriptions. Apply: Feed Cerebro/Magnet keyword output into Scribbles to systematically optimize listing copy without missing target terms.
Listing Builder — Helium 10's newer AI listing-writing tool, co-built with Andrew, designed to write use-case-based, feature-to-benefit copy and to optimize for Rufus as well as traditional keyword search. Apply: Use it to generate listing copy that connects product features to concrete customer benefits and use cases, not just keyword density.
Keyword Tracker — Helium 10's tool for tracking keyword rank position and Amazon's Choice badge status over time; Andrew says he was among its first heavy users. Apply: Track rankings across a broad, redundant SKU portfolio to catch ranking recovery after stockouts and to surface new Amazon's Choice badges as proof points for further investment.
Black Box — A Helium 10 product research tool mentioned by Andrew in the context of his broader toolset. Apply: Referenced only in passing as part of the Helium 10 suite; no specific application described in the transcript.
Tiered keyword-list architecture — Andrew's practice of building separate category-level, product-level, and broader reusable keyword lists rather than one flat list. Apply: Segment keyword research by scope (category vs. specific product vs. reusable cross-catalog terms) so lists can be applied efficiently across thousands of SKUs.
Brand Store as SEO content hub — Treating the Amazon Brand Store page like an SEO blog with themed content (e.g., 'top 10 metal wall art for a Victorian-style home') to rank in Google, exploiting Amazon's domain authority. Apply: Publish blog-style, keyword-targeted content and video on your Brand Store page to capture Google search traffic in addition to on-Amazon search traffic.
Cross-referencing Google and Amazon keyword data — Andrew's method of pulling outside Google search terms and matching them against Helium 10-identified transactional Amazon keywords. Apply: Compare external Google search-term data to Helium 10 keyword data to find overlapping terms worth targeting in both Brand Store SEO content and Amazon listings.
Manufacturer video content for Google SEO — Using manufacturer handcrafting/process videos on Brand Store content pages because video aids Google search ranking. Apply: Source and embed manufacturer or product-process video on Brand Store SEO pages to improve their Google ranking potential.
Amazon's Choice badge tracking — Monitoring Amazon's Choice badge attainment via Keyword Tracker as a signal of ranking success. Apply: Track new Amazon's Choice badges to demonstrate ROI on keyword/SEO work and justify further investment internally.
SEO/PPC sequencing theory — Andrew's framework that SEO and PPC are 'symbiotic' and should not be separated — running PPC after listing optimization reinforces organic rank, while running PPC before optimization causes an organic collapse when ads stop. Apply: Fully optimize a listing's keywords and content before launching PPC campaigns, so ad spend compounds with (rather than substitutes for) organic ranking.
Broad SKU portfolio redundancy strategy — Building sales resilience through hundreds of interchangeable, deeply-stocked SKUs rather than concentrating on a few hero products, so any single stockout has minimal impact. Apply: Diversify catalog depth so no single product's stockout meaningfully damages total revenue, and rely on rank tracking to confirm quick recovery after restocks.
Custom GPT building — Andrew's personal practice of building custom ChatGPT-based GPTs, scaling from an initial ~350 to 700. Apply: Build purpose-specific custom GPTs for recurring workflows as a way to operationalize AI use across many discrete tasks.
Rufus optimization — A distinct optimization discipline for Amazon's Rufus AI shopping assistant, treated as separate from but linked to traditional keyword SEO. Apply: Write listings that answer customer questions and use-cases directly (not just stuff keywords) so Rufus can surface and answer with your product.
A9 algorithm — Amazon's core search/ranking algorithm, described as working in 'symbiotic union' with Rufus. Apply: Continue optimizing for traditional keyword relevance and search ranking signals, since Rufus is said to draw on A9 rather than replace it.
Search query plans — The mechanism Andrew describes by which Rufus extracts the most relevant products before surfacing them to a shopper. Apply: No direct seller action described; presented as background mechanism informing why Rufus results overlap with search results.
Product Search API — The backend Andrew claims Rufus draws from to source products, based on his study and a related patent. Apply: Treat Rufus optimization and standard Amazon search optimization as the same underlying task rather than two separate systems to game.
Relevancy-signal seeding — Sending a small amount of deliberate purchase traffic against a specific target keyword to trigger an indexing/ranking shift, as demonstrated in the bat-toilet-rug case study. Apply: When a new or niche listing isn't surfacing for an exact-match keyword despite being indexed, drive one or two purchases specifically through that keyword to signal relevance to Amazon.
Feature-to-benefit writing framework — A copywriting framework connecting specific product features to the concrete benefits they produce for a given use case (e.g., mirror length to how much of the body it shows). Apply: For each product feature, write the specific benefit it delivers for a specific use case, rather than listing features alone or assuming bigger/more is always better.
Use-case-based / intent-based listing copywriting — Writing listing copy around customer use cases, fears, and behavior rather than pure keyword placement, described as core to both Listing Builder and Rufus's underlying design. Apply: Structure listing content around how customers actually use and evaluate the product, since this is what drives conversion and what Rufus is built to answer.
Intent mapping — A custom Claude skill built by Andrew that takes a keyword list and organizes/matches it by intent type. Apply: Run a keyword list through an intent-classification process to group terms by underlying shopper intent rather than treating them as a flat, undifferentiated list.
PPC campaign structuring by intent-type/theme — Organizing PPC campaigns around keyword intent themes (e.g., splitting 'metal wall art' into sub-themes) instead of one broad campaign. Apply: Break a broad product category into themed sub-campaigns based on intent-mapped keyword groups to remove guesswork on what to bid on.
Keyword Sales metric — A Helium 10 feature showing sales attributable to specific keywords, distinguished from standard Amazon search term reports. Apply: Prioritize bidding and optimization decisions based on which keywords are shown to actually generate sales, not just which appear in search term reports.
LLM-assisted search term report analysis — Uploading search term reports and product information to ChatGPT or Claude in reasoning/extended-thinking mode to analyze keyword relevance instead of manual review. Apply: Upload the product's own information first, then the search term report, to a reasoning-enabled LLM (e.g., ChatGPT with extended thinking, or Claude with adaptive thinking) to get relevance-graded keyword analysis.
Price-history-aware ethical pricing strategy — A pricing approach responsive to Rufus/Amazon surfacing historical pricing (including pre-Prime-Day prices) to shoppers, framed as an emerging ethical constraint rather than a purely commercial lever. Apply: Avoid pricing patterns that look manipulative once price history is visible to customers (e.g., inflating pre-sale prices), balancing profitability against this new transparency.
Andrew claims outside Google search terms substantially overlap with Amazon's own transactional keywords, and exploited that overlap by cross-referencing Google-search data against Helium 10 keyword data to build Brand Store content aimed at ranking in Google itself.
He claims Amazon Brand Store pages carry enough of Amazon's domain authority to rank organically in Google search on their own, effectively functioning as an SEO blog rather than just a storefront.
The relevancy-signal-seeding technique described in the bat-toilet-rug case: deliberately sending a tiny amount of purchase traffic (one or two orders) against a specific target keyword with zero prior search volume flipped both organic search ranking and Rufus's answer for that exact query within under an hour, offered as evidence the two systems share a backend.
Andrew claims (citing his own study and 'a relevant patent') that Rufus's answers are drawn overwhelmingly from Amazon's actual product Search API rather than being mostly personalized per shopper, though some answers are personalized.
Bradley notes that testing thousands of ranking variables shows little additional differentiation over testing merely tens or hundreds of them.
Rufus's price-history transparency is reframed not as a UX feature but as a structural constraint on seller pricing behavior — Andrew told a colleague Rufus has effectively become a pricing-strategy consideration he never expected it to touch.
Rufus is described as capable of supporting conditional purchase logic (e.g., 'buy this if there's a discount of X%'), extending its role from product discovery into purchase-timing behavior.
Andrew's Touch of Class strategy deliberately avoided a hero-SKU concentration model, instead using deep catalog redundancy (thousands of niche SKUs each selling only once or twice a month) so that any single stockout barely dented overall sales.
«how cool is that pretty cool I think.»
— 00:13
«To this day, Bradley, I read one ebook per day.»
— 04:08
«Keyword research SEO is the art of making what is invisible visible through the right, you know, search terms, I believe.»
— 08:14
«Fun fact, we have the largest collection of metal wall art on Amazon. It's not even close.»
— 08:55
«We had a million dollars in sales our last my last year when I was with them on just that brand store page.»
— 10:54
«Guess what? We had the right keywords.»
— 14:53
«SEO and PPC are in symbiotic union with each other, right? You cannot separate those two.»
— 16:57
«I had built around 350 custom GPTs, uh, which by the way, I'm up to 700 now.»
— 18:08
«We just recently were we're up 600%, uh, well over, um, 600% uh, year over year.»
— 18:39
«I intend to prove that definitively here within the next couple months.»
— 21:37
«these things are intertwined. You don't really optimize for one and not the other.»
— 23:44
«You can't just gouge them»
— 24:29
«keep me away from people who just say intent or people who just say keywords without mentioning Rufus a thing.»
— 27:39
«make sure also to upload the information of your product first so it knows that.»
— 30:50
Reception
Viewers found the episode highly valuable and practical, praising its actionable Amazon selling insights.
This is a Helium 10-branded interview blending a genuine practitioner's growth story with vendor tool promotion, and its boldest technical claim — that a single case study proves Rufus and A9 draw from the same backend — rests on one anecdote rather than the broader study Andrew says he is still working to prove.
31:41