By Diosh — Founder, AHAeCommerce | eCommerce decision intelligence for $50K–$5M GMV operators
A brand selling a $40 product on Amazon FBA keeps roughly $15 after the referral fee, fulfillment fee, storage, and the PPC spend required to maintain category ranking. The same product on a Shopify store at $40 keeps roughly $27 after payment processing, fulfillment, and customer acquisition cost — when the brand can drive its own traffic. The gap between $15 and $27 per unit is the price of Amazon's audience, and it is also the price of not owning your customer relationship. Which side of that trade-off is correct depends on where you are in the business, not on which channel is "better."
The Amazon-vs-own-store decision is the most consequential channel choice an eCommerce brand makes, and it is consistently framed as a binary when it is actually a stage-by-stage allocation problem. Run the wrong allocation for 18 months and you build either a margin-starved Amazon dependency or a customerless DTC store with zero discovery flywheel.
The Default Assumption (and Why It Fails)
The standard framing treats Amazon and DTC as alternatives — "should we sell on Amazon or build our own store?" — when the actual question for most brands is the mix. Amazon's marketplace serves 310+ million active customers in the US alone (Statista, 2024) and captures 37.6% of US eCommerce spend (Insider Intelligence, 2024). DTC stores have to manufacture every customer relationship from acquisition spend or owned audience.
This framing fails because it ignores that the two channels measure different things. Amazon optimizes for unit volume and discovery; DTC optimizes for margin and customer lifetime value. A brand that posts $1M of GMV on Amazon and $200K on DTC is not "80% Amazon" — it is a brand whose customer relationships are concentrated where they can be most easily disintermediated, and whose margin pool is concentrated where it can be reinvested.
The relevant comparison is not revenue per channel. It is contribution margin per customer per channel, and the strategic optionality each channel preserves.
What the Decision Actually Hinges On
The True Amazon Take Rate (30–50% of Revenue)
The Amazon take rate is consistently underestimated because the referral fee (8–15% depending on category, with 15% as the default for most categories) is the only number visible at the checkout level. The total cost stack: referral fee (15%), FBA fulfillment fee ($3.45–$5.55 for standard-size items per item shipped, per Amazon's FBA fee schedule, 2024), FBA storage fee ($0.87–$2.40/cubic foot/month for standard storage), inbound shipping to Amazon warehouses, PPC spend (typically 8–15% of revenue to maintain category ranking), and the 25% return rate in apparel/30%+ in some categories that compounds storage and processing cost.
Stacked end-to-end, the average Amazon take rate for a brand running steady-state FBA with normal PPC investment runs 35–45% of gross revenue. For apparel and high-return categories it pushes to 50%+. This is the gross-to-net delta most operators don't model before launching on Amazon, and it is the number that determines whether Amazon is profitable at your margin profile.
The Customer Data Asymmetry
Amazon does not give you the customer relationship. You receive a name, a shipping address, and order metadata — and you are prohibited from marketing to the customer outside the platform without explicit opt-in (which converts at 2–5%). You cannot run remarketing. You cannot run win-back. You cannot run loyalty. You cannot run LTV-based bidding on paid channels because you don't have the LTV data.
DTC gives you the email, the purchase history, the browsing behavior, and the right to use all of it for retention marketing. This is the asymmetry most underestimated by founders new to Amazon. A $50K monthly Amazon revenue stream produces zero compounding marketing asset; the same $50K monthly DTC revenue stream produces ~1,000 new customer email records monthly that compound into the most valuable owned audience the brand will ever build.
The Discovery Cost That Amazon Solves
The thing Amazon does that DTC cannot do without significant spend: it produces customer discovery without paid acquisition. A brand ranking organically for a category keyword on Amazon receives free traffic that, on DTC, would cost $15–$60 per click in a competitive category. This is the structural advantage Amazon provides — it is also why dependency on Amazon is so easy to slip into, because the unit economics of Amazon-driven volume look strong relative to the painful CAC math of building DTC traffic from scratch.
The decision hinge: discovery on Amazon is paid for in margin (35–45% take rate); discovery on DTC is paid for in cash (CAC of $15–$80 depending on category). One is a variable margin tax, the other is a fixed cash burn. They are not equivalent.
The Cost Reality
The following table shows unit economics for a $40 SKU on Amazon FBA vs. Shopify DTC, assuming $12 COGS, 25% return rate in apparel scenario, 10% PPC on Amazon, $30 blended CAC on DTC.
| Cost Line | Amazon FBA ($40 sale) | Shopify DTC ($40 sale) | |---|---|---| | Gross revenue | $40.00 | $40.00 | | Referral fee (15%) | −$6.00 | — | | FBA fulfillment | −$4.75 | — | | FBA storage (allocated) | −$0.40 | — | | PPC spend (10%) | −$4.00 | — | | Return processing (allocated) | −$1.20 | — | | Payment processing (2.9% + $0.30) | — | −$1.46 | | Shipping (own) | — | −$4.50 | | CAC (allocated per order) | — | −$10.00 | | COGS | −$12.00 | −$12.00 | | Contribution per unit | $11.65 | $12.04 | | Take rate (total fees/CAC) | 41.1% | 39.9% | | Customer data captured | Limited (shipping address only) | Full (email + behavior + history) | | Retention marketing optionality | None (without opt-in) | Full |
The contribution margins are near-equivalent at this unit economic profile. The strategic value diverges sharply. The Amazon $11.65 produces zero compounding asset; the DTC $12.04 produces a customer relationship worth, on average, $25–$80 in lifetime value through email-driven repeat purchase. The DTC unit is 2–6x more valuable on a 24-month horizon — but it requires the brand to have the marketing infrastructure to capture that value.
This is also where marketplace fees compound into a problem that looks like a cost line but is actually a strategic constraint — the 41% Amazon take rate is the price of not owning the customer.
The Trade-Off Map
Pre-PMF and $0–$100K GMV: Lead with Amazon
For brands without product-market fit, Amazon's discovery is the fastest validation surface available. Listing a SKU on Amazon and watching the organic-vs-paid sell-through rate tells you within 60–90 days whether the product, the price, and the listing copy work. The same validation through DTC requires $5,000–$25,000 of paid acquisition spend before you have enough sample size to read signal. At this stage, Amazon is the cheapest market research available.
The trap at this stage: ranking on Amazon produces revenue that feels like product-market fit when it is actually marketplace-fit. A product that does $30K/month on Amazon's category page may sell zero on a DTC store because no one is searching for the category outside Amazon. The Amazon validation tells you the SKU works inside the Amazon traffic flow; it does not tell you the brand can be built outside it.
$100K–$500K GMV: 70/30 Amazon-DTC, Start Capturing
The transition at $100K GMV is to begin building DTC infrastructure while Amazon still drives most of the volume. Launch the Shopify store; insert a card in every Amazon shipment offering a benefit (discount, content, warranty extension) for opting into the brand's owned channels. The expected opt-in rate is 5–8% — modest but durable. Run a small DTC paid acquisition test ($1,500–$3,000/month) to begin learning your CAC curves and audience signal outside of Amazon.
The trap at this stage: trying to scale DTC paid acquisition to match Amazon revenue. Your CAC math on DTC is not yet good enough — you don't have the LTV data, the email flows, or the conversion-rate optimization to make paid acquisition profitable at scale. Use this stage to capture, not to outrun the marketplace.
$500K–$2M GMV: 50/50 or 60/40, Focused DTC Investment
By $500K GMV the question is whether to make a deliberate investment in DTC traffic flywheel — content, paid social, email-driven repeat, influencer seeding — to reduce Amazon dependency before it becomes a single-point-of-failure. Brands that successfully transition through this band invest 6–12% of revenue in DTC marketing infrastructure for 18–24 months, accepting flat or briefly declining margin to build the audience asset.
The trap at this stage: cutting Amazon investment to fund DTC growth. Amazon advertising and inventory positioning are pay-to-stay; reducing PPC or letting inventory go thin causes ranking decay that takes 3–6 months to recover. Fund the DTC investment from new capital or from margin elsewhere, not by reallocating from Amazon while you still depend on it.
$2M+ GMV: DTC-First with Amazon as Discovery Channel
Mature brands that have built DTC repeat customer cohorts can flip the model — DTC becomes the margin and LTV channel, Amazon becomes the prospecting and discovery channel that introduces new customers who are then migrated to owned channels. This is the configuration most defensible brands settle into. The trade-off is operational complexity: managing two channels with different fulfillment, different pricing constraints (Amazon's price parity requirements), and different inventory rules requires meaningful infrastructure that smaller brands cannot afford.
The trap at this stage: undercutting Amazon pricing on DTC and being delisted, or running into Amazon's "buy box" pricing enforcement. Maintain price parity at the listed level; offer DTC benefits through bundles, loyalty discounts, and exclusive products rather than headline price.
When to Act (Specific Triggers)
Trigger 1: Audit Amazon Take Rate at 6-Month Anniversary
If you have been on Amazon for 6+ months, audit the actual take rate by pulling the last 90 days of payment reports, summing all Amazon fees (referral + FBA + storage + PPC + return processing), and dividing by gross revenue. If the rate exceeds 45%, you are running an unprofitable Amazon program at most margin profiles; restructure inventory mix, listing optimization, or PPC strategy before adding more SKUs.
Trigger 2: Insert Capture Cards When Crossing $50K Monthly Amazon
Every Amazon order is a one-time customer relationship by default; an insert card converts a percentage to an owned-channel customer. At $50K monthly Amazon revenue (~1,250 orders at $40 AOV), even a 5% opt-in rate produces 62 new owned customers/month — a meaningful audience-building rate. Insert cards must comply with Amazon's terms of service (no off-platform pricing comparisons, no anti-Amazon language); content-led capture (recipe books, care guides, warranty registration) outperforms discount-led capture at 2x rates.
Trigger 3: Diversify When Amazon Exceeds 70% of Revenue
If Amazon exceeds 70% of revenue for 6+ consecutive months, the brand is structurally dependent on a single channel that can suspend you for policy violations, change category fee structures (it has), or surface a competing private-label SKU that outranks your listing. Diversification is not a profitability move at this point; it is risk management. Allocate 15–25% of marketing budget to DTC growth even if it produces lower short-term ROI than additional Amazon PPC.
Trigger 4: Compare LTV Net Margin, Not Gross Margin, Across Channels
Run the customer-level LTV calculation separately for Amazon-acquired customers (limited to first purchase + opt-in repeat) vs. DTC-acquired customers (full email/SMS LTV). The gap is typically 2–4x in favor of DTC over a 24-month horizon. If your channel mix decisions are still based on first-purchase contribution margin (which looks nearly equal across channels), you are systematically undervaluing DTC.
What Operators Get Wrong Most Often
Mistake 1: Treating Amazon Revenue as Equivalent to DTC Revenue
The most common error is reporting blended revenue and blended margin without segmenting by channel. Amazon revenue is non-compounding; DTC revenue is compounding through the customer relationship. A brand at $2M GMV with 80% Amazon revenue and a brand at $2M GMV with 30% Amazon revenue are not the same business — they have different defensibility, different exit multiples, and different strategic optionality. The second brand is worth materially more, and the gap widens every year.
Mistake 2: Underestimating Amazon's Pay-to-Stay Mechanics
Amazon ranking is not earned and held — it is purchased continuously. The brand that ranks #2 for its category keyword today is ranking there because of a specific PPC bid level, a specific conversion rate, and a specific inventory depth. Cut PPC by 30%, drop into low inventory for two weeks, or get hit by a return-rate spike, and the ranking degrades within 2–4 weeks. Recovering the ranking requires 6–8 weeks of overspend at higher CPCs. Operators who treat their Amazon ranking as a stable asset rather than a continuously-rented position consistently underestimate the maintenance cost.
Mistake 3: Building DTC Too Late
The opposite mistake at the mature stage: brands that wait until Amazon represents 85%+ of revenue to start building DTC infrastructure usually find the acquisition spend ramp impossibly steep, because they are trying to manufacture audience the same year they are trying to fund growth. Brands that begin DTC investment at the $100K–$200K GMV stage — when Amazon is still doing the work of revenue — are dramatically better positioned for the transition than brands that wait until they "have the margin to invest" at $1M+ GMV.
The Verdict
Amazon-vs-own-store is not a channel choice. It is an allocation question that shifts by stage: lead with Amazon pre-PMF, begin capturing at $100K GMV, invest in DTC infrastructure at $500K, flip to DTC-first at $2M+. The brand that picks one channel and stays there for the full lifecycle is making a mistake regardless of which one they pick.
This week: Calculate your actual Amazon take rate from the last 90 days of payment reports. If you do not have a DTC store yet and are above $50K monthly on Amazon, set up Shopify and insert capture cards into the next 1,000 Amazon shipments. If you have both channels but have never compared 24-month LTV by channel, run that segmentation in your analytics — the gap will reframe how you think about channel mix.
