Amazon Business Model Selection

Part of the AMZ Operator Series (companion to IDS) | Built July 2026 | Distilling: Ryan Daniel Moran · Travis Marziani · Dan Meadors & Dylan Frost (The Wholesale Formula / Fast Track Formula) · Chris Grant & Chris Racic (Clear the Shelf) · Greg Mercer (Jungle Scout) · Stephen Smotherman (Full-Time FBA) · Steven Pope (My Amazon Guy) · Kevin King · Mike Jackness (EcomCrew) · Juozas Kaziukėnas (Marketplace Pulse) · Flips4Miles · Reezy Resells

---

The one-paragraph truth: There is no "best" Amazon model. There is a best model for your capital, your hours, and your risk tolerance — and picking wrong burns 6–18 months. Arbitrage is a job that teaches the platform. Wholesale is a logistics business with thin, repeatable margins. Private label is a capital-intensive brand bet with the only real exit value. Dropshipping on Amazon, as most people practice it, is a policy violation with a countdown timer. This module makes the selection decision once, correctly.

---

#1. THE SEVEN MODELS — MASTER MATRIX

#1.1 Reading the matrix like an operator

• Only Private Label builds a sellable asset. Wholesale accounts and arbitrage skills die with your login. FBA brands transact at ~2.5–3.5x SDE in 2025–26 (Empire Flippers avg ~3.1x; clean brands 3.5–4x per Quiet Light); the 4–6x aggregator era died with Thrasio's Chapter 11 (Feb 2024).
• Net margin ≠ ROI. OA nets "only" 10–20% of revenue but turns capital 6–10x/year — 26%+ monthly ROI on buys is common. PL nets more per dollar of revenue but locks cash in 90–120 day inventory cycles.
• Jungle Scout 2025 baseline: 58% of sellers turn profitable within year one; 61% run margins above 10%; about a third exceed 20%. The median killer isn't margin — it's running out of cash first.

#1.2 Dropshipping FBM — the policy most people are violating

Amazon's dropshipping policy is short and lethal. You MAY dropship only if:

1. You are the seller of record on every document;
2. Your name appears on all packing slips, invoices, and external packaging — no third-party branding;
3. You handle all returns yourself;
4. You are NOT "purchasing products from another online retailer and having that retailer ship directly to customers."

That fourth line kills the "Walmart-to-Amazon" playbook taught on YouTube: a Walmart box with a Walmart receipt lands on the porch → one report → Section 3 deactivation. Amazon Sellers Lawyer reports dropshipping as a top cause of suspensions through 2025. The only compliant version is a genuine wholesale/3PL relationship — a distributor or brand blind-ships under YOUR name with YOUR packing slips. That's not "dropshipping" as sold in courses; it's FBM wholesale with delegated fulfillment. "No-inventory Amazon automation" agencies are account arson for a fee.

#1.3 Handmade & Merch/KDP — the honorable side quests (brief)

• Handmade: application-gated, artisan-only, flat 15% referral fee, professional account fee waived, no listing fees (vs Etsy's $0.20 + 6.5% + payment fees + forced Offsite Ads over $10k/yr); 250M+ buyers vs Etsy's ~90M. Best net margins on Amazon (30–50%) because labor is the hidden cost; ceiling = your hands. Run both platforms; Amazon is the volume channel.
• Merch on Demand: now a three-tier performance royalty system (2026): a $19.99 tee pays ~$2.44 (Creator), ~$4.88 (Plus, 15%+ external traffic), $5.27 (Premium, 35%+). Amazon cut base pay and now pays you to bring traffic. KDP: 60% print royalty minus print cost ($6.49 on a $14.99 paperback); 70%/35% ebooks. 20–50 real titles commonly produce $500–3,000/mo. Neither is sellable; both are cash-flow hobbies that teach keywords.

---

#2. DEEP DIVE: PRIVATE LABEL — THE BRAND BET

#2.1 Ryan Daniel Moran's "12 Months to $1 Million" framework

Moran's math is the cleanest mental model in the space:

THE MILLION-DOLLAR FORMULA
3–5 products × 25 sales/day × $30 average price ≈ $1M/year

Check:  25 units × $30 = $750/day/product
        4 products × $750 = $3,000/day = ~$90k/month = ~$1.08M/year
        ($1M/year = $83,333/month — you need ~100 total sales/day at $30)

The sequence matters more than the math — person → product → 25/day:

1. Person first. Pick a specific human (the new dad, the home barista, the kettlebell lifter) — not a product niche. Products for a person compound; random products compete.
2. The Grind (months 1–4): launch product #1, do anything non-scalable to reach a sustainable 25 sales/day.
3. The Growth (months 5–8): audience + ads to hold 25–30/day profitably.
4. The Gold (months 9–12): launch products 2–5 to the SAME buyer until you average 100 sales/day.

Moran's operator warnings: price at $30+ (sub-$20 products can't pay for PPC in 2026), plan inventory so success doesn't stock you out, and treat reviews/audience as the compounding asset — 300+ of his students have exited seven figures on this template.

#2.2 Travis Marziani's passion-product variant

Marziani ($10M+ via Performance Nut Butter and courses) attacks the same goal from differentiation: build a product YOU are the customer for, and make it visibly different — not the 47th silicone spatula with a new logo. His launch stack: pre-launch audience (YouTube/TikTok/email) → crowdfund or waitlist validation → launch with external traffic Amazon's algorithm rewards. His published budget reality: ~$5,500–$13,000 all-in (≈$3,000 first inventory, ~$200 samples, ~$250 GS1 barcode, $39.99/mo account, tools, $300–1,000 PPC, trademark $350–600). The passion angle isn't romance — it's content fuel: you can make 100 videos about a problem you actually have.

#2.3 Capital reality in 2026 — the $5k floor vs the $15–30k truth

The "$3–5k start" still exists on paper (one SKU, 300–500 units, sea freight, zero buffer). Operators budget differently because 2025–26 added structural costs:

Rule: if you have less than ~$10k liquid for PL, you don't have a PL budget — you have an arbitrage budget and a savings plan. Sourcing shifts (Vietnam, India, Mexico light-assembly) mitigate tariffs but add MOQ and QC friction. Undercapitalization doesn't usually kill PL sellers at launch — it kills them at the first reorder, when revenue is proving out and there's no cash to buy stock.

---

#3. DEEP DIVE: WHOLESALE — THE WHOLESALE FORMULA METHOD

#3.1 Why wholesale is misunderstood

Wholesale = buying branded products already selling on Amazon at wholesale prices and joining the existing listing as an FBA offer. No listing creation, no launch, no reviews to earn, demand already proven by Keepa history. Dan Meadors and Dylan Frost started with $600 of arbitrage credit-card float, built to $36M+ in Amazon sales, and codified the system as The Wholesale Formula (relaunched 2025 as Fast Track Formula, ~$2,997; ~7,000 students claiming $1B+ cumulative revenue).

#3.2 Reverse sourcing, step-by-step (the TWF engine)

Traditional wholesale: get a distributor catalog, pray something's profitable. TWF inverts it — find the winning ASIN first, then go get it:

REVERSE SOURCING PROTOCOL
1. SCOUT    — Mine Amazon for replenishable, brand-owned ASINs:
              • BSR consistently strong in category (Keepa 90/180-day avg)
              • Price $15–70, not seasonal, not fad
              • Amazon NOT on the listing (or <30% Buy Box share)
              • 2–15 FBA sellers (proof multiple resellers are tolerated)
              • Brand is NOT enrolled in exclusive-distribution lockdown
2. FILTER   — Unit economics: target ≥15% net margin & ≥33% ROI
              after referral + FBA fees at CURRENT Buy Box price.
              Monthly volume ÷ (sellers + you + 1) = your unit share.
3. CONTACT  — Go around distributors. Email/call the BRAND OWNER.
              Ask for a wholesale account as an authorized reseller.
4. PITCH    — Don't beg for an account; offer VALUE ("brand advocate"):
              • Fix their garbage listing (title/images/A+/keywords)
              • Manage/report their Amazon channel, police MAP
              • Consolidate the seller mess they hate
5. NEGOTIATE— Volume price breaks, exclusivity or capped seller counts,
              payment terms (net 30 after order #2–3).
6. REORDER  — This is a replenishment machine: same POs monthly,
              margins defended, relationships deepened.

#3.3 Why distributors kill the margin

Brand-direct: you pay true wholesale (~50% off retail). An authorized distributor stacks 20–30% on top of that and sells the same catalog to every other seller who asks — no exclusivity, no moat, and after Amazon's ~30–35% total fee load there is often nothing left. Distributors are for ungating invoices and gap-filling; brand-direct relationships are the business. TWF's own economics: aim ≥$3 profit/unit, 15%+ net margins, and win by owning 20 boring accounts nobody on TikTok will ever film.

#3.4 The 2026 brand-gating reality

Gating got worse and better simultaneously. Worse: Amazon gates at brand, category, and ASIN level; ungating invoices must be commercial wholesale invoices (10+ units, dated within 90–180 days, matching your legal name/address), and Amazon cross-references whether your supplier is actually authorized — random distributor invoices increasingly bounce. Transparency/Brand Registry programs let brands block unauthorized sellers outright. Better: if YOU are the authorized seller with the letter, gating is your moat — every gate that blocks a rando protects your Buy Box share. In 2026 wholesale, the sales skill (winning brand accounts) IS the defensibility.

---

#4. DEEP DIVE: ARBITRAGE (OA/RA) — THE CASH-FLOW TREADMILL

#4.1 What it is, honestly

Buy discounted retail (online = OA, in-store = RA), resell on Amazon at the spread. No brand, no moat, 100% labor-fueled. It is ALSO the fastest, cheapest real-world education in fees, prep, Buy Box mechanics, Keepa, and account health — which is why the smartest voices (Chris Grant & Chris Racic of Clear the Shelf, Stephen Smotherman of Full-Time FBA, Flips4Miles, Reezy Resells) all frame it as the on-ramp, not the destination.

#4.2 Keepa graph literacy — the non-negotiable skill

Keepa (~€19/mo) is the arbitrage X-ray. Read it or donate money to those who can:

KEEPA DECODE RULES
• RANK DROPS = SALES. Each downward "icicle" on the green BSR line
  ≈ at least one sale. Count drops/month ≈ minimum monthly velocity.
• Judge the 90/180-DAY AVERAGE, never today's price (spike = trap).
• ORANGE SHADING = Amazon in stock. Amazon holding the Buy Box
  >30% of the time? Walk away — you won't outlast their pockets.
• OFFER COUNT RISING while price falls = race to the bottom forming.
  Model your exit price at the 90-day LOW, not current Buy Box.
• BUY BOX ROTATION: ~80%+ of sales flow through the Buy Box.
  Your share ≈ months_volume ÷ (FBA sellers + you). If 10 sellers
  share 100 sales/month, your "winner" moves 9 units. Pass.
• Sub-1% category BSR ≠ safety if drops are seasonal — check last
  year's Q1, not last month's Q4.

#4.3 The workflow stack

• RA: Amazon Seller App / ScoutIQ scan → SellerAmp SAS ($19.95–34.95/mo) for instant ROI + eligibility ("can I even sell this?") → buy at ≥$3 profit AND ≥30% ROI after all fees → prep → ship.
• OA: Tactical Arbitrage ($59–159/mo; scans 500+ retail sites against Amazon) or lead lists (OA leads services ~$30–100/mo) → SellerAmp/Keepa validation → stack discounts (cashback portals 2–8%, discounted gift cards, store cards) — discount stacking is where half the OA margin actually comes from (Chris Grant's core play).
• Guardrails: check gating BEFORE buying (SellerAmp shows it), keep IP-complaint-prone brands (Nike, Disney, LEGO tier) off your menu, keep every receipt/invoice for 2 years, and register for sales tax where required — marketplace facilitator laws cover the sale, but resale certificates keep you from paying tax on the BUY.

#4.4 Why it's a treadmill — and why you should run on it anyway

The model's flaw is structural: profit only exists while you're shopping. Inventory doesn't replenish itself; every dollar earned must be re-deployed by YOUR hours; Amazon pays out on a ~2-week cycle while your credit card is due in 30 days — the float IS the business. Ceiling: most solo operators stall at $20–60k/mo revenue because sourcing hours don't scale (VA teams and prep centers extend it, at margin cost). Run it anyway if you're under-capitalized, because: first profit in days–weeks (vs 4–9 months PL), total tuition ~$500–2k, mistakes cost $50 not $15,000, and after 6 months you can read a Keepa chart like a radiologist. That skill transfers directly into wholesale scouting and PL competitive analysis.

---

#5. THE MIGRATION LADDER — SEQUENCING MODELS

CAPITAL & SKILL LADDER (the proven sequence)

RA ($500+)          →  learn: fees, prep, Keepa, account health
  ↓ 3–9 months, bank $5k+
OA ($2–5k)          →  learn: sourcing systems, VAs, cash-flow mgmt
  ↓ 6–12 months, bank $10k+, spot repeat winners
WHOLESALE ($10k+)   →  learn: brand relationships, replen ops, negotiation
  ↓ 12–24 months, bank $25k+, know a niche's buyer cold
PRIVATE LABEL ($15–30k) → build: brand, audience, launch machine
  ↓ 24–48 months, $250k+/yr profit, clean books, 2+ yrs history
EXIT (2.5–3.5x SDE)  → sell to search funds/individuals (aggregators are gone)

Rules of the ladder: skipping rungs is buying tuition at PL prices (a failed PL launch costs 10–30x a failed OA buy). But don't squat on a rung either — arbitrage past ~18 months without systematizing means you built yourself a job. Meadors/Frost climbed arbitrage→wholesale; Moran preaches straight-to-PL only for the funded and obsessed. Both are right for their capital profiles. The wholesale rung is optional but powerful: replen wholesale income is the most stable cash to fund PL launches. And PL→exit only works if you run clean books from day one (separate LLC, accrual accounting, SDE tracking) — Module AMZ_11 covers the exit mechanics.

---

#6. THE 2026 BATTLEFIELD — WHAT CHANGED, WHERE MARGIN HIDES

#6.1 The Great Compression (Marketplace Pulse's term, and it fits)

• New-seller collapse: only ~165,000 new sellers launched a first listing on Amazon.com in 2025 — a decade low, down 44% YoY and ~73% off the 2021 peak (700k+). Of those, 59.9% were Chinese sellers; US sellers just 16.3% of new launches.
• Fee extraction at maximum: Amazon's revenue mix is now ~60% services / 40% retail — sellers ARE the product. All-in Amazon take (referral + FBA + storage + mandatory-in-practice ads) commonly lands at 45–55% of retail price.
• Temu/Shein/Haul pincer: Temu scaled to ~$22B US GMV; Amazon Haul (sub-$20, direct-from-China, ~1M+ items, est. ~$2B year-one US sales) means Amazon now competes with its own sellers at the bottom shelf. If your product's Chinese factory can list it on Haul, your PL "brand" is a countdown clock.
• Tariff floor: post-truce Section 301 rates settle most FBA categories at 20–30% effective duties; de minimis is dead globally (Aug 2025). Everyone's landed costs rose 15–40%; the winners re-quoted, consolidated freight, or moved origin.

#6.2 Where margin still lives (operator targets)

1. $35–80 price band, differentiated PL — under $25 you're fighting Haul/Temu economics you cannot win; over $35 the direct-from-China players thin out and PPC math works.
2. Bulky/heavy/fragile items — bad unit economics for cross-border direct shipping = structural protection.
3. Brand-gated wholesale accounts — the gate is the moat; boring replenishables (auto parts, toys, grocery, pet, hardware) with authorization letters.
4. Compliance-heavy categories (topicals, supplements, kids' products) — paperwork is a filter that scares off the horde. Do the paperwork.
5. Domestic/nearshore sourcing stories — "Made in USA" now carries both tariff immunity and conversion lift.
6. Arbitrage in the thinned field — with active sellers down ~30% from peak and traffic per remaining seller UP, clearance cycles at Target/Walmart/Home Depot are less picked-over than 2021–22 lore suggests.

---

#7. DECISION FRAMEWORK — PICK YOUR MODEL IN 20 MINUTES

#7.1 The three gates (answer honestly)

1. Capital gate: liquid risk capital (money whose loss changes nothing)? <$3k → RA/OA only. $3–10k → OA/wholesale-lite. $10–30k → wholesale or PL. $30k+ → any model.
2. Time gate: real weekly hours? <10 → Merch/KDP or don't start. 10–20 → OA/wholesale. 20+ → any. RA specifically needs daytime store hours.
3. Risk gate: could you emotionally survive a $10k inventory write-off? No → stay arbitrage/wholesale. Yes → PL is available to you.

#7.2 Scoring worksheet

Score each 1–5, multiply by weight, highest total wins:

MODEL SELECTION SCORECARD (1 = poor fit for you, 5 = strong fit)
                                weight   PL   WS   OA   RA
Capital available vs required     x3     __   __   __   __
Hours available vs required       x2     __   __   __   __
Risk tolerance vs model risk      x3     __   __   __   __
Patience (months to profit OK?)   x2     __   __   __   __
Sales/negotiation appetite        x1     __   __   __   __  (wholesale lives here)
Marketing/brand appetite          x1     __   __   __   __  (PL lives here)
Physical hustle appetite          x1     __   __   __   __  (RA lives here)
Need income within 90 days? (5=no)x2     __   __   __   __
                                TOTAL:   __   __   __   __
Tie-breaker: pick the LOWER-capital model. You can always climb;
you can't un-spend a failed PL launch.

#7.3 The 90-day commitment rule

Whatever wins: 90 days, one model, zero channel-hopping. Model-switching every 3 weeks is the #1 beginner death pattern — every model looks broken in week 2 and printable in month 6 of someone else's YouTube video.

---

#8. STEADY-STATE P&Ls — THE THREE MAIN MODELS, REAL NUMBERS

#8.1 Private Label at steady state (2 SKUs, ~$33 ASP, ~50 units/day)

Note: cash-locked in inventory ≈ $30–45k at all times. Profit ≠ payout.

#8.2 Wholesale at steady state (12 brand accounts, ~$25 ASP)

No ad spend, no launches. Scale lever = more accounts, not more margin.

#8.3 Online Arbitrage at steady state (solo + 1 VA)

~28% monthly ROI on deployed capital — best cash-on-cash of the three, zero enterprise value, and it stops the week you stop sourcing.

---

#9. FAILURE MODES — HOW OPERATORS DIE HERE

#9.1 Private Label deaths

• The me-too launch: 47th garlic press, zero differentiation → PPC bleeds $5k → liquidate at a loss. (Antidote: Marziani's rule — visibly different or don't launch.)
• Reorder starvation: PO#1 sells, no cash for PO#2, stockout kills rank, relaunch from zero.
• Tariff blindness: quoted 2024 landed costs in a 2026 world; margin evaporates at customs.
• Haul-vulnerable product: factory lists your exact item at 40% of your price.
• Review-gaming shortcut: account gone. No appeal wins against manipulated reviews.

#9.2 Wholesale deaths

• Distributor-only sourcing: 8% "margins" that are really 0% after returns and storage.
• Brand pulls authorization (or goes exclusive/gated against you) on your #1 account — 40% of profit gone overnight. (Antidote: no account >20% of profit.)
• Race-to-the-bottom listings: 14 sellers join, price drops 22%, your 6-month stock is underwater.
• Fake "wholesale" suppliers: counterfeit goods → IP complaint → Section 3. Verify authorization chains, always.

#9.3 OA/RA deaths

• Treadmill burnout: month 14, sourcing 25 hrs/week, income stops the week you stop. No system, no VA, no exit plan.
• Buying today's Keepa spike: price normalizes 35% lower before your stock lands.
• IP complaint roulette: selling gated/litigious brands without invoices Amazon will accept.
• Tax mess: no resale certificates, no bookkeeping, then a five-figure surprise in April.
• Ungating after purchase: garage full of inventory you're not allowed to list.

#9.4 Dropshipping (FBM) death — singular

• You run retail dropshipping (Walmart→customer), a buyer photographs the Walmart invoice, Amazon deactivates under Seller Code of Conduct, your funds sit frozen 90 days, and the "automation agency" that set it up doesn't answer. This isn't a risk; it's a schedule.

---

#MODULE SUMMARY — THE TEN COMMANDMENTS OF MODEL SELECTION

1. Match the model to your capital, not your ambition. <$10k liquid = arbitrage or wholesale-lite. PL with lunch money is a donation to Amazon's ad business.
2. Only private label builds an asset you can sell (2.5–3.5x SDE in 2026). Every other model is income, not equity — plan accordingly.
3. Arbitrage is the tuition, not the career. Ride it 6–18 months for cash and Keepa literacy, then climb the ladder: RA → OA → wholesale → PL → exit.
4. In wholesale, go brand-direct or go home. Distributor markups + Amazon's ~30% fee load = zero. The account-winning pitch (fix listings, police the channel) IS the business.
5. Retail dropshipping on Amazon is a suspension with a delivery date. Seller of record, your packing slips, your returns — or don't.
6. Rank drops are sales. Never buy, launch, or negotiate anything you haven't validated against a 90/180-day Keepa view — and model exits at the 90-day low.
7. Price above the Haul line. Sub-$25 undifferentiated products now compete with Amazon's own direct-from-China channel. $35–80, differentiated, defensible — or skip.
8. Budget tariffs like a fee, not a surprise: 20–30% on China-sourced landed cost is the 2026 baseline; de minimis is dead. Requote everything.
9. Diversify what kills you: no wholesale account >20% of profit, no PL SKU >40% of revenue, no arbitrage brand you can't invoice.
10. Pick once, commit 90 days. The 165k new sellers of 2025 (vs 700k in 2021) prove the tourists left. The model you actually execute beats the model you keep researching. Fewer competitors, higher bar — 2026 punishes hobbyists and pays operators.

---

Next module: AMZ 03 — Product Research & Validation (where PL and wholesale scouting get their full toolkits).