You and 4,304 subscribers are going to enjoy this newsletter or i'll submit a flat file on your account with all your variations set to 'ScentName'.

Welcome back friend!

Know an Amazon Seller who’d love 💕 this newsletter? Share it in one click.
(it really helps us)

Same Fear, Different Century.

Everyone's convinced AI is about to delete half the workforce. It’s not.

Why? Jevons' Paradox. Here’s the basic idea. Most people think when something gets cheaper, we use less of it. The paradox says the opposite is true: When something gets cheaper & more efficient, we use more of it. 

For example, when the steam engine made coal more efficient to use, England didn't use less coal. They used massively more. Efficiency didn't reduce demand. It exploded it.

This pattern repeats everywhere.

• ATMs made basic banking transactions effortless. Banks didn't shrink. They opened more branches and hired more tellers, redeploying them into advisory and sales roles.

• Spreadsheets automated the grunt work of accounting. The profession didn't collapse. It expanded, because suddenly every small business could afford financial analysis.

• The internet was supposed to kill retail. Instead it created an entirely new economy of logistics, fulfillment, content, and digital commerce that dwarfs what came before.

• Cloud computing made server management trivially cheap. Companies didn't need fewer engineers. They built more software, launched more products, and hired more engineers than ever.

Now watch it happen again with AI. Designers aren't getting replaced. They're producing ten times the output, which means companies want more design, not fewer designers. Developers shipping code faster doesn't shrink teams. It expands what teams can build. Every efficiency gain creates new surface area for human work.

That's the paradox. Making work easier doesn't eliminate it. It creates more of it.

So the question was never "Will AI take my job?" It was always "What new work will AI make possible?"

The IVS

Big thanks to my friend Ares Severi for sharing this week’s tip. Ares runs viscacha.ai, an AI agent that is the most advanced way to find winning Amazon products, from niche discovery to manufacturer sourcing.

Here’s what I do before I consider a new launch “done.” I run the listing through Rufus and start asking it questions a real buyer would ask.

• "Is this dishwasher safe?"

• "Will this fit a 15 inch laptop?"

• "Can I use this outdoors?"

Ask 15-20 of them. Rufus suggests its own questions on the product page too, so tap those and screenshot every answer.

Some questions Rufus answers using your listing copy. Some it answers with vague language or pulls from a random review. Some it gets wrong or says the information isn't available. Those last two categories are gaps in your listing that no keyword report or analytics dashboard will surface. A shopper asked a real question, Amazon's AI failed to answer it from your content, and that shopper looked at the next product.

You may have heard of the above before. But here’s where it gets really good:

Paste the screenshots into Claude and ask it to sort every response by confidence level. For each weak or incorrect answer, Claude identifies what to add to your bullets, description, or A+ content so Rufus handles the question next time. You get a prioritized list of specific content gaps based on what shoppers and Rufus are asking about your product right now.

Once you do this and you’re happy with how it turns out, record your screen with Claude Cowork & ask it to do it for you going forward. Then you’ve built a repeatable skill that allows you to see what Amazon thinks about any/all of your products.

Did an extremely good-looking friend forward you this newsletter?

Get the next one, here!

How about some Volcano Surfing?

This week Jackson’s high school class trip went to Nicaragua for 4 days on a bus from Costa Rica. One of the highlights was sand surfing on Cerro Negro volcano, which is still considered an active volcano last erupting in 1999. The only problem is you first have to hike about 2 km up the hill to get to the top, no lifts up there!

Put on the protective “Breaking Bad” gear and away you go! He said it started slow but you really pick up speed further down the slope. Add it to your to do list!

The “Review Moat Strategy” Just Got a Haircut

For years, one of the more effective Amazon strategies was building a review moat through variations. Take your hero ASIN with 2,000+ reviews, attach lower-selling child ASINs under the same parent. Different formulas, different bundles, different specs. Every child inherited the full review pool. Instant social proof on day one.

That's over (At least the way most sellers were doing it).

What Changed

Last month, Amazon started splitting pools on variations where the products are functionally different. Each child ASIN only shows reviews written for that specific product. The rollout is category-by-category through May 31.

Amazon is applying a substitution test: would a customer be somewhat satisfied with receiving variant A instead of variant B? If no, reviews split.

What Still Shares: color, size, pack quantity. Same product, different look. A blue water bottle and a red water bottle still pool reviews. A 3-pack and a 6-pack still pool.

What Doesn't: different flavors, different materials, different specs, bundles vs standalone, different formulas, different fits. If a review for variant A would be misleading on variant B, they split. Amazon explicitly called out chocolate vs vanilla protein powder as an example.

So Should You Break Up Your Variations Entirely?

This is the question everyone's asking, and the answer is: probably not.

Reviews were the loudest benefit of variations, but they weren't the only one. The variation detail page is still one of the best free cross-sell mechanisms on Amazon. A customer searching for your vanilla protein powder lands on the listing and sees chocolate, strawberry, unflavored right there. That's free product discovery you'd have to pay for with ads on a standalone listing.

Traffic consolidation still works too. Eight standalone listings each need to fight for rank independently. Eight variations under one parent means one page captures all that traffic. You rank one child, the rest ride along.

And here’s a “softer” conversion metric: showing a full product line makes you look like a real brand. Customer sees 6 options and thinks "these people know what they're doing." A single standalone listing with 200 reviews and no visible product family looks thinner.

When Splitting IS the Right Move

Break them apart when a child ASIN has bad reviews that were previously diluted across the pool. Those reviews are now exposed, and if the product is genuinely weak, it's dragging down your page experience.

Split when the products are so different that grouping them confuses the customer. If someone looking at a protein powder sees creatine as a "flavor variation," that hurts trust more than it helps.

Split when a child would rank better in a different subcategory on its own, or when you need completely independent keyword strategies per product.

What To Do Right Now

1. Look at your variation families. Which child ASINs are functionally different? Those reviews are getting separated whether you act or not.

2. Check your review distribution. How many reviews does each child actually have on its own? A listing showing 2,400 pooled reviews could drop to 300 overnight. Know your exposure.

3. Enroll vulnerable ASINs in Vine now. Up to 30 reviews per child ASIN. Start building before the split hits (if the listing is eligible).

4. Keep your legitimate variations grouped. The cross-sell, traffic consolidation, and brand credibility are worth it even without shared reviews.

The review moat still exists. It just only works when you've built it around genuinely similar products. And the variation page itself? Still valuable. Just not as a Trojan horse for fake social proof.

WiFi Sucks During Travel? Do This:

Google Maps lets you download full city maps for offline navigation. Do this on wifi before you leave. Open the area you need, tap your profile picture, tap "Offline maps," and select the region. It takes about 30 seconds per area.

Another one to try: Maps.me. It pulls from OpenStreetMap, which is community-maintained and weirdly thorough in places Google has barely touched. Southeast Asia, Central America, rural Costa Rica, parts of Europe. Download both before you leave.

Why both? Because they fail differently. Google is better for driving directions and business hours. Maps.me is better for finding that unmarked trail to the waterfall your Airbnb host mentioned, or the walking path from your rental to the beach that Google thinks doesn't exist.

One more thing: download the maps for a wider area than you think you'll need. Day trips happen. Plans change. You may find yourself in the middle of the Monkey Trail, by mistake (like me).

To your success,

Rob & Max