[ BDSN ] Do this before spending a single dollar on inventory or ads

Every Amazon seller hits that moment.

You’re doing everything right: the ads are running, listings are optimized, revenue’s in the seven figures, and yet, you’re still chained to the grind.

80-hour weeks. Constant firefighting. Margins that shrink while your effort multiplies.

That’s what sparked a private ESS webinar, a closed-door session for top sellers ready to break that pattern.

We called it The 6R Revolution. It’s a system built to create self-running empires that grow even when you step away.

The room went silent when Liz revealed how each “R” connects to form an unstoppable Amazon growth flywheel.

Now, for a short time, that replay is open to the public.

✅ Why most 7-figure sellers stay trapped in “busy” mode
✅ How top sellers scale to 8-figures with 1/3 the effort
✅ The complete 6R framework — revealed step by step

If you’re ready to free yourself from the constant hustle, this replay could change how you see your business forever.

🎥 Watch the 6R Revolution Replay Now —> Limited Access

👀 HOW AI CAN THINK LIKE YOUR CUSTOMERS

Imagine you're selling lemonade and want to know if kids will buy it. Usually, you'd need to ask hundreds of real kids what they think. But what if you could ask a really smart computer to pretend to be those kids instead, and get answers that are 90% as accurate as asking real people?

That's exactly what this groundbreaking research just figured out how to do.

Scientists taught AI (specifically something called a Large Language Model, or LLM) to act like different types of shoppers.

Here's how it works:

The Old Way:

• Survey 300 real people about a product

• Wait days or weeks

• Pay thousands of dollars

• Get answers that might be biased

The New Way:

1. Tell the AI: "Pretend you're a 35-year-old mom from Texas who makes $50,000 a year"

2. Show it a picture of your product

3. Ask: "Would you buy this?"

4. The AI writes what that person would think

5. Compare the AI's answer to pre-written "reference statements" to score it

The crazy part? The AI got it right 90% as accurately as asking real people.

Why This Is Mind-Blowing

No Training Needed: They didn't have to teach the AI anything special. They just used regular ChatGPT-style AI right out of the box.

Beat Traditional Methods: When they tried using old-school computer programs (machine learning that looked at customer data), those only got 65% accuracy. The AI approach got 88-90%.

It Actually Understands Context: The AI didn't just pick random numbers. It wrote thoughtful responses like:

• "I'm somewhat interested. If it works well and isn't too expensive, I might give it a try."

• "Seems kinda bougie for body wash. I'll stick with what I know."

Real personality. Real reasoning. Not just cold statistics.

How E-Commerce Sellers Can Use This RIGHT NOW

Here's what becomes possible:

1. Test Products Before You Make Them

• Create 5 different product concepts

• Have AI "customers" from different demographics review each one

• See which concept scores highest

• Only manufacture the winner

Example: You're launching a new backpack. Instead of guessing, create AI personas of:

• College students

• Business professionals

• Outdoor enthusiasts

• Parents

Test which design resonates with each group before spending $10,000 on inventory.

2. Write Better Product Descriptions

• Show your product page to AI customers

• See what concerns they raise

• What features do they care about?

• What makes them say "definitely yes" vs "maybe"?

Use those insights to rewrite your listings to address actual objections.

3. Price Testing Without Risk

• Create AI panels at different income levels

• Test price points: $29.99 vs $39.99 vs $49.99

• See where purchase intent drops off

• Find your optimal price point

4. Demographic Targeting

The research showed AI accurately mimicked how different groups respond:

• Age patterns: Middle-aged people showed higher interest than young/old

• Income sensitivity: Lower-income personas were more price-conscious

• Category preferences: Different demographic groups preferred different product types

This means you can predict which customer segments will buy before spending a dollar on ads.

5. A/B Test Everything (Without Real Traffic)

• Different product images

• Different packaging designs

• Different brand messaging

• Different bundle options

Run all your tests in AI first, then only invest in real A/B testing for the winners.

Why This Is Different from Regular Data Analysis

Old approach: Look at what customers did in the past

• "20% of visitors clicked the buy button"

• "Average order value was $47"

• "Conversion rate was 3.2%"

New approach: Simulate what customers will do in the future

• Model specific personas

• Understand why they buy or don't buy

• Get psychological insights, not just numbers

The paper's key phrase: "This shifts from analyzing past data to simulating future customer behavior."

The Secret Sauce: Semantic Similarity Rating (SSR)

Here's the clever trick that makes this work:

Instead of asking AI "Rate this 1-5," which gives weird results, they:

1. Ask AI to explain its feelings in plain English

2. Compare that answer to five reference statements (one for each rating level)

3. Use similarity matching to figure out the score

Why this matters: AI is much better at explaining things than picking numbers. This method plays to its strengths.

Real Results from the Study

They tested this on 57 actual product surveys with 9,300 real people who rated personal care products.

The AI method:

• Matched the distribution of ratings (how many 1s, 2s, 3s, 4s, 5s)

• Ranked products in the same order as humans

• Showed the same demographic patterns (age, income effects)

• Even wrote more detailed feedback than most humans gave

Direct quote from a synthetic customer: "The ease of use and safety are appealing, but I'd want to know more about its effectiveness and any potential side effects."

That's not robotic. That's human-level reasoning.

The Bigger Picture: We're Entering a New Era

This isn't just about surveys. This represents something profound:

We can now model human behavior without humans.

Think about what becomes possible:

• Instant focus groups for any product idea

• Personalized marketing tested on AI twins of your customers

• Risk-free innovation (test 100 ideas, only build the best one)

• Hyper-targeted products designed for specific customer segments

As one insightful observer noted: "Some brands nail their customer understanding, while others fall short. This reveals exactly who the customer is, down to the topics, interests, and triggers that drive action."

The Bottom Line for Sellers

You no longer need to guess what customers want. You don't need expensive market research firms. You don't need to wait weeks for survey results.

You can:

1. Define your target customer

2. Feed their profile to an AI

3. Show the AI your product

4. Get near-human-level feedback

5. Iterate and improve

6. Repeat until you have a winner

All before spending a single dollar on inventory or ads.

This is the future of e-commerce product development. The brands that figure out how to use this first will have an enormous competitive advantage.

The researchers conclude: "When paired with appropriate methods, LLMs can serve as valid synthetic consumers for concept testing."

The computer can now pretend to be your customer so well that it's almost as good as asking real people.

And that changes everything.

🔭 WALMART and OPEN AI PARTNER to INTEGRATE SHOPPING

Walmart and Sam's Club have announced a strategic partnership with OpenAI that will fundamentally change how customers shop.

This move signals a major shift in retail that could impact traffic patterns across all e-commerce platforms, including Amazon.

• Direct ChatGPT Integration: Customers can now complete Walmart purchases directly within ChatGPT conversations with no need to visit Walmart's website or app. This creates a new shopping channel entirely outside traditional marketplaces.

• Proactive vs. Reactive Shopping: The AI learns customer preferences and anticipates needs before customers even search. This means competing retailers are now intercepting purchase intent earlier in the customer journey, potentially before shoppers ever reach Amazon.

• Walmart's Broader AI Strategy: This partnership extends Walmart's existing AI capabilities, including enhanced product listings, faster customer service, and improved associate training—all areas that directly compete with Amazon's operational advantages.

🤖  HOW to OPTIMIZE for RUFUS on AMAZON

This is part 2 of Andrew Bell’s deep dive on how Rufus is changing the way your products are being discovered on Amazon. Read part one first in Monday’s newsletter.

How to Optimize for Rufus

1. Map Your Product's Connections Ergonomic office chair connects to:

• Remote work setup → desk → chair → monitor arm

• Back pain solution → assessment → chair → standing desk

• Home office upgrade → furniture → chair → décor

2. Make Your Product Easy to Discover A yoga mat should work for:

• Fitness goals

• Joint health

• Home gym setup

• Meditation

• Physical therapy

• Travel workouts

Each pathway needs different signals in your listing.

3. Help Complete the Mission "Everything needed for X (no additional purchases required)" "Pairs with [items] to complete your setup"

4. Align with Rufus's Thinking

Old Way: Optimize for keywords
New Way: Optimize for storylines

Answer questions Rufus needs answered:

• "Is this for beginners?" → Address experience level

• "Works in small spaces?" → Give dimensions

• "Pairs with [item]?" → Highlight compatibility

5. Add Context Everywhere

Weak: "Yoga mat, 6mm, non-slip"
Strong: "6mm yoga mat perfect for home practice and studios. Non-slip for hot yoga. Eco-friendly materials for mindful living. Physical therapist-recommended thickness for joint support. Pairs with blocks, straps, and bolsters."

6. Memory Builds Loyalty

If a customer likes your kitchen organizer, Rufus remembers. When they need pantry solutions later, your brand has an advantage.

This means:

• First purchase quality is CRITICAL

• Bad experiences stick

• Good experiences lead to more recommendations

• Design product families customers can grow through

7. Outside Reputation Matters

Rufus checks websites, reviews, and expert opinions. A critical YouTube video can hurt your recommendations.

Build authority through:

• Getting featured in buying guides

• Earning expert reviews

• Winning awards

• Creating helpful content others share

8. Balance Old and New SEO

Keep doing (Keywords):

• Research core search terms

• Optimize titles

• Fill backend keywords

• Track rankings

Add this (Context):

• Problem-solution language

• Use case specifics

• Lifestyle alignment

• Mission integration

• Outcome framing

Your Action Plan: Preserve, Evolve, Build

Keep:

• Keyword research

• Title optimization

• Backend terms

• Category accuracy

• Performance tracking

Why: Keywords tell Rufus WHAT your product is. Without them, you won't even be considered.

EVOLVE: Product Content 🔄

Transform this: "Stainless steel chef's knife, 8-inch blade, ergonomic handle"

Into this: "Professional 8-inch chef's knife for home cooks building skills. Handles daily meal prep from weeknight dinners to weekend entertaining. Ergonomic handle reduces fatigue. Used in 5,000+ home kitchens. Pairs with cutting board, honing steel, and knife block."

What to change:

• Descriptions → Mission stories

• Bullet points → Benefits with context

• Images → Show realistic use

• A+ Content → Customer journey maps

• Q&A → Proactive mission support

Start with: Your bestsellers and products with high traffic but low conversion.

1. Map Your Product's Missions

Example: Camping Tent

• Mission 1: Family camping (needs: safety, comfort, easy setup)

• Mission 2: Backpacking (needs: light weight, durability)

• Mission 3: Festival shelter (needs: quick setup, style)

• Mission 4: Emergency prep (needs: reliability, storage)

For each: What do they care about? What else do they need? What worries them?

2. Create Bundle Paths

Design product families for mission completion:

STARTER: Basic coffee maker ($50)
 ↓
CORE: Espresso machine + grinder ($300)
 ↓
ENTHUSIAST: Full setup with frother + scale ($600)

3. Build External Authority

Short-term (0-6 months):

• Submit for reviews

• Contact niche influencers

• Participate in roundups

Long-term (18+ months):

• Become category expert

• Create research others cite

• Build industry relationships

The shift is simple:

Old: Get found → Convert
New: Get found → Fit the mission → Help completion → Build relationship

The sellers who win won't just optimize for algorithms—they'll create products and content that genuinely help customers succeed. Rufus is smart enough to know the difference.

✌🏼 Have a great weekend.

See you again on Monday.

The answer to today’s STUMP BEZOS is
Amazon expects to make $700 million PROFIT from Rufus in 2025