15:16
Amazon PPC
This weekly Amazon/e-commerce roundup argues that purchase-intent and conversion data now cross platforms faster than most sellers adapt: keywords already proven on Amazon can be redeployed on TikTok Shop with outsized results, and AI-referred shoppers already convert far better than normal traffic even though that traffic mostly bypasses Amazon for brand-owned sites — so sellers who build cross-platform and AI-answer-engine strategies now, before it's obvious, stand to capture disproportionate share of the next few years.
Elena of AZ Rank tested dropping three Amazon-proven keywords (from SQP and Product Opportunity Explorer) into TikTok Shop's underused 250-character search keywords field, leaving everything else untouched as a control.
Result: product-level GMV up 79%, impressions up 102%, items sold up 75%; on the product-card channel alone, ~$1,600 GMV, 111 orders, ~93% GMV growth, and a ~7% click-to-order rate from ~82,000 impressions.
The three test keywords (which drove purchases) began outranking the two untouched control keywords (which drove none), showing the same Amazon 'rank flywheel' logic playing out on TikTok Shop.
Kevin King's suggested weekly action: pull high click-share/conversion-share keywords from Brand Analytics SQP, cross-reference with Opportunity Explorer, build a 15-20 keyword list, and paste it into TikTok Seller Center's search keywords field, then wait ~30 days.
Prime Day prep: brands that win started 6-10 weeks in advance (keyword testing, ad tuning, inventory math); Prime Day 2025 drove $24.1B in sales, 307 million items, and 63% of US households ordered two or more times, with 2026 expected to top it.
Helium 10's Carrie Miller and BDSS member Dain Wishon released a free Prime Day checklist (inventory, listing optimization, ad strategy, external traffic, historical trends, post-event momentum) with a worksheet.
Adobe Analytics data (May): AI-referred shoppers generated 53% more revenue per visit, converted at a 54% higher rate, and spent 53% more time on site than non-AI traffic; AI traffic to retail sites is up 138% year-over-year, the highest share since Adobe began tracking in October 2024.
A year ago non-AI visits were worth 128% more than AI visits; that gap has fully inverted so AI-referred visits now command a premium — described as a complete flip in twelve months.
Separate Adobe data: agentic shopping is cutting returns by as much as 69%.
Most AI-referred traffic currently flows to brand-owned websites rather than Amazon, because AI models tend to send shoppers to a brand's own site to complete the purchase.
Claude (Anthropic) was observed serving product cards for the first time this week (via monitoring from Refi by), across electronics, fashion, and beauty — described as rough/early but directionally clear.
Analyst Scott Wingo moved Claude up his agentic-commerce autonomy tracker from 'research' to 'find,' while moving Perplexity down after Perplexity pulled its product cards and buy button entirely.
Anthropic has reportedly been meeting with merchants since at least NRF in January and has said it wants Claude to be able to act on a shopper's behalf to complete purchases; Kevin King says to plan as if Claude reaches 'buy' by holiday 2026.
Kevin King's four-point response to AI shopping: (1) build a clean, fast, machine-readable brand site since Amazon detail pages rarely get the referral; (2) make product data readable via clear specs and structured data; (3) build authority off-site via reviews, comparison content, and mentions on trusted sites; (4) run weekly shopping prompts in Claude/ChatGPT/Gemini to track whether and how your product is surfaced, calling this practice AEO (answer engine optimization).
NRF data: Father's Day 2026 spending hit a record $27.9B (up from $24B in 2025), with average spend of $227 per person (up from $199); top categories by total spend were special outings ($4.8B), clothing ($4.1B), gift cards ($3.3B), electronics ($3.1B), and personal care ($2.3B); by percent purchasing: gift cards 60%, clothing 58%, special outings 55%, gift cards again 52%, personal care 40%.
Adam's PPC framework: top sellers build a five-layer 'stack' rather than hunting for one tool — Research (Helium 10, Jungle Scout, Data Dive), Campaign Management (Amazon Ads Console, Perpetua, Pacvue), Optimization (Teikametrics, Cortile, Ad Badger), Analytics (Amazon Marketing Cloud, Amazon Marketing Stream, Amazon Attribution), and Conversion (PickFu, managed experiments/A-B testing, Canva).
A 'science says' segment citing 19 experiments: framing an identical product as having fewer ingredients raised choice rates by ~22% for a granola bar, ~16% for a juice, ~67% for peanut butter, and pulled up to 44% more clicks on Meta ads.
The effect reverses for products bought for pleasure (e.g., dark chocolate) or nutritional variety (e.g., vitamins/supplements), where more ingredients can signal more flavor or benefit.
Recommended tactic: state the ingredient count prominently in title, bullets, A+ content, and packaging, using a digit ('3') rather than the spelled-out word, citing lesser evil's three-ingredient Himalayan pink salt popcorn as a missed opportunity since it doesn't state the count anywhere on the listing or bag.
Quick hits: Amazon is testing an unpaid discovery module on some brands, quietly rolled out a 5-year purchase dataset in Amazon Marketing Cloud, sellers feel better about Prime Day but are watching margins closely, and Walmart's marketplace is expanding cross-border into Mexico.
Stump Bezos trivia: per the Cloudflare CEO, bots have now passed human traffic for the first time in internet history, with bots at 57.4% of web traffic.
TikTok Shop Search Keywords Cross-Pollination — A technique, tested by Elena of AZ Rank, of copying already-proven high-converting Amazon keywords into TikTok Shop's largely unused 250-character search keywords field on the theory that purchase intent doesn't change across platforms. Apply: Pull top-ASIN keywords with high click share and conversion share from Amazon Brand Analytics SQP, cross-reference with Product Opportunity Explorer, build a 15-20 keyword list, paste it comma-separated into the TikTok Seller Center listing's search keywords field, save, and wait about 30 days while watching product-card metrics.
SQP (Search Query Performance) — An Amazon Brand Analytics report showing which search queries drive clicks and conversions for a seller's ASINs, used in the video as the data source for the TikTok Shop keyword test. Apply: Pull SQP data for top ASINs and combine it with Product Opportunity Explorer signals to identify keywords with high click share and conversion share worth reusing on other sales channels.
AEO (Answer Engine Optimization) — The practice, as framed by Kevin King, of optimizing a brand's data and content so AI models like Claude, ChatGPT, and Gemini recommend the product when shoppers ask for suggestions, distinct from optimizing for on-platform Amazon search ranking. Apply: Stand up a fast, machine-readable brand website, structure product data with clear specs, build third-party authority through reviews and comparison content on sites the models trust, and run shopping prompts in Claude, ChatGPT, and Gemini weekly to track whether and how your product appears versus competitors.
Agentic Commerce Autonomy Tracker — Analyst Scott Wingo's framework for ranking AI assistants on a maturity spectrum from 'research' to 'find' to 'buy' based on how autonomously they can complete a purchase for a shopper. Apply: Watch which direction a given AI platform is moving on the tracker (e.g., Claude moving up to 'find,' Perplexity moving down after removing its buy button) to decide which AI surfaces to prioritize monitoring and optimizing for.
Five-Layer PPC Tech Stack — A framework, credited to Adam, arguing that scaling Amazon PPC requires assembling a connected stack across five layers rather than relying on a single tool: Research (Helium 10, Jungle Scout, Data Dive), Campaign Management (Amazon Ads Console, Perpetua, Pacvue), Optimization (Teikametrics, Cortile, Ad Badger), Analytics (Amazon Marketing Cloud, Amazon Marketing Stream, Amazon Attribution), and Conversion (PickFu, managed A/B experiments, Canva). Apply: Map current tools to each of the five layers, close gaps layer by layer, and ensure research, bidding, optimization, analytics, and conversion tools share data instead of treating PPC as isolated ad management.
Fewer-Ingredients Framing Effect — A marketing/psychology finding, cited from 19 experiments, that describing a product as having fewer ingredients increases purchase likelihood because shoppers read 'fewer' as more natural and healthier, with the effect reversing for pleasure-driven or variety-driven categories like chocolate or vitamins. Apply: State the ingredient count prominently in the title, bullets, A+ content, and on-package copy using a digit rather than a spelled-out number — unless the product is bought for pleasure or nutritional variety, in which case emphasize more ingredients or benefits instead.
The AI-vs-non-AI traffic value gap didn't just narrow, it fully inverted within twelve months — non-AI visits were worth 128% more a year ago, and AI-referred visits now command a premium instead, a magnitude and speed of reversal larger than the individual conversion-rate stats alone suggest.
The AI shopping race is uneven rather than uniformly advancing: Claude is moving toward autonomous buying (per Wingo's tracker) at the same moment Perplexity is retreating and pulled its buy button entirely, meaning sellers can't assume 'optimize for AI' means the same thing across every model.
The TikTok Shop test's real proof point wasn't just that Amazon keywords worked, but that the flywheel dynamic itself transferred: the three keywords with purchases began outranking the two purchase-free control keywords, replicating Amazon's own rank-feeds-rank loop on an entirely different platform.
The 'fewer ingredients' framing effect is not a universal rule but flips direction based on purchase motivation — functional/health products benefit from claiming fewer ingredients, while pleasure-driven or variety-driven categories benefit from the opposite framing, so the same tactic requires opposite execution depending on category.
Kevin King frames AEO as structurally different from Amazon SEO not just in tactics but in control: it's a discovery surface where the seller doesn't own the ranking mechanism at all, which is why his prescribed response is building authority and structured data off-Amazon rather than optimizing an on-platform listing.
«her hypothesis was keywords that already convert on Amazon should work on Tik Tok shop too because purchase intent doesn't really change between platforms, right?»
— 01:44
«The three keywords that got purchases outrank the two control keywords that didn't get any purchases.»
— 02:51
«AI referred shoppers generated 53% more revenue per visit than nonAI traffic. They converted their rate 54% higher and they spent 53% more time on site and viewed more pages.»
— 05:22
«The AI sends you a shopper who already did the research, already narrowed the field, and shows up ready to spend.»
— 06:09
«If you're invisible to the crawler, you're invisible in the answer. Period.»
— 08:18
«Sellers who treat this like the early days of Amazon search and move before it's obvious, they're going to own the next few years.»
— 09:13
«framing your product as having just a few ingredients instead of many can make shoppers up to 22% more likely to pick it.»
— 12:23
«The purpose of business is not to make money. The only reason to start a business is to deliver some product or service to humanity that makes their life better.»
— 14:36
«The answer is 57.4%. More than half of all internet traffic is bots.»
— 14:55
Reception
No comments are available to gauge audience reception.
A densely packed weekly news digest that trades depth for breadth — each of its seven-plus segments (TikTok keyword arbitrage, Prime Day prep, AI shopping data, Claude's product cards, PPC stack framework, ingredient framing study) is individually well-sourced and actionable, but none gets more than a few minutes, so it functions better as a trend-scanning briefing than a deep tutorial on any single tactic.
15:16