Google Ads vs Amazon Ads : e-commerce 2026

According to FEVAD data and the 2025 e-commerce Federation, about 62% of French mid-market e-commerce retailers (annual revenue between €1M and €50M) are present simultaneously on their DTC site and on Amazon in 2025-2026, with a projection at 70% in 2027. On this hybrid panel, Amazon typically captures 25 to 55% of the brand's total e-com revenue, and 40 to 65% of the paid acquisition budget depending on vertical. The 2026 strategic question is almost never "Google or Amazon" but "how to allocate between the two to maximize consolidated net margin, not apparent ROAS"

This article exposes the complete mechanics: the structural DTC vs marketplace dichotomy (two strategies, two economies, two customer relationships); the compared Google Shopping/PMax vs Amazon Sponsored Products mechanics; real net margins DTC website vs Amazon FBA after all fees; comparable ROAS and cross-channel cannibalization; the full Amazon stack (Sponsored Products, Sponsored Brands, Sponsored Display, DSP); allocation by 3 typical business profiles; and the budget arbitrage methodology based on net margin rather than platform ROAS. For Google Shopping fundamentals and Merchant Center feed optimization, see our Google Shopping setup and optimization guide. For the full Google Ads e-com stack, see our 2026 e-com Google Ads playbook. Official Amazon Advertising documentation available on advertising.amazon.com. Our MER calculator (Marketing Efficiency Ratio) measures overall marketing efficiency, not channel-by-channel.

DTC vs Amazon marketplace: 2 strategies, 2 economies

DTC (Direct-to-Consumer via own merchant site) and Amazon marketplace are not two interchangeable channels, they are two distinct economic models requiring two distinct strategies. DTC is an owned-audience strategy: the brand owns the customer, first-party data, transactional history, direct email/CRM relationship, and the full net margin (minus logistics and payment fees). Amazon marketplace is an inventory and visibility strategy: Amazon owns the customer (the email stays with Amazon, not the brand), takes a category commission, provides FBA logistics if activated, but offers in return already mature and qualified shopper traffic.

This structural dichotomy implies radically different economic arbitrages. In DTC, every sale feeds the owned audience, LTV (lifetime value) builds over time, retargeting and email marketing widen the margin long-term. On Amazon, every sale is purely transactional — the customer is not directly re-engageable, LTV is limited to the rebuy rate on Amazon (and can be captured by other brands on Amazon), net margin is compressed by category commission + FBA fees + returns.

Direct operational consequences for paid acquisition steering:

• In DTC, marketing investment can accept a lower apparent ROAS short-term because CRM LTV compensates (a loyal DTC customer is typically worth 2.5 to 6x their first order over 24 months). Google Shopping/PMax + retargeting + post-purchase email amplify this mechanic.
• On Amazon, marketing investment must be amortized on the first transaction because LTV-via-marketplace is mechanically lower (Amazon rebuy rate from the same brand without direct email is 30-50% lower than DTC rebuy rate with active email + retargeting). Sponsored Products + Sponsored Brands must generate positive net ROAS transaction by transaction.

Typical profiles observed on aggregate Google Ads data 2025-2026:

• Pure DTC without Amazon: about 18% of FR mid-market e-com in 2026. Typical of very premium or very brand-driven brands (luxury, editorial niche, custom-made products) that refuse Amazon dilution or cannot exist there without destroying their positioning. Paid allocation: 100% Google Ads + Meta + complement.
• Hybrid DTC-dominant: about 32% of the panel. The brand has a mature merchant site and does Amazon as volume complement. Paid allocation: 60-75% Google + Meta DTC, 25-40% Amazon Ads.
• Balanced hybrid: about 25% of the panel. DTC site and Amazon contribute comparably to revenue. Paid allocation: 45-55% Google + Meta DTC, 45-55% Amazon Ads.
• Hybrid Amazon-dominant: about 18% of the panel. Amazon generates 60-80% of revenue, DTC site is residual. Paid allocation: 25-40% Google DTC (often mostly brand defense), 60-75% Amazon Ads.
• Pure marketplace: about 7% of the panel. No DTC site, exclusive Amazon sales (sometimes extended to Cdiscount + Fnac). Paid allocation: 100% Amazon Ads.

Initial allocation choice depends mainly on 4 factors: differential DTC vs Amazon net margin (detailed in section 3), DTC site maturity (UX, checkout funnel, reviews), commercial Amazon dependence (account suspension risk weakens the entire business), and long-term brand strategy (owned audience vs rented traffic). For e-com strategy and Shopping vs PMax arbitrage on the Google side, see our Shopify vs PrestaShop setup guide for Google Ads.

Google Shopping/PMax vs Amazon Sponsored Products mechanics

The two native e-com mechanics differ deeply in auction logic, exposure surface and intent signal. Google Shopping (Standard + Performance Max e-com) captures the Google Search query with explicit commercial intent — the user types "blue navy cotton summer dress" in Google, sees Shopping ads at the top/side of the SERP, clicks to the DTC merchant site. Official documentation support.google.com Shopping campaigns.

Amazon Sponsored Products captures the Amazon Search internal query with already mature commercial intent — the user types "blue navy cotton summer dress" directly in the Amazon engine, sees Sponsored Products ads in the results, clicks to the Amazon product page. Official documentation advertising.amazon.com Help.

Google Shopping operational mechanics: product feed managed in Google Merchant Center, Shopping Standard campaign or integrated into Performance Max, CPC bids with Smart Bidding (Target ROAS, Maximize Conversion Value). The algorithm distributes impressions across Search SERP, Shopping tab, YouTube, Display Network, Discover (in PMax). The click leads to a product URL on the DTC site, with full brand control over the checkout funnel, post-purchase experience, retargeting, email marketing.

Amazon Sponsored Products operational mechanics: product feed managed in the Amazon catalog (Seller Central or Vendor Central), Sponsored Products campaign with auto-targeting (Amazon ML) or manual-targeting (keywords or competitor ASINs). CPC bids with Dynamic bids or Fixed bids strategies. The click leads to the Amazon product page, with brand control limited to listing content (title, bullet points, A+ content, videos), not the checkout funnel (Amazon controls), not post-purchase (Amazon controls), not direct customer retargeting.

Key differences in steering:

• Google Shopping is multi-surface via PMax: a single container can serve Search + Shopping + Display + YouTube + Discover. This "broad signals + ML decisioning" logic is powerful but requires clean Customer Match, optimized Merchant Center feed, and well-calibrated Smart Bidding to avoid drift. For full PMax e-com mechanics, see our Performance Max 2026 guide.
• Amazon Sponsored Products is mono-surface (Amazon): simpler auction, ROAS observable transaction by transaction, but no native multi-channel extension (except moving to Sponsored Display or DSP — section 5).
• Product page content: on Amazon, the listing is shared between your ad and all competing ads that would lead to it; your competitors can bid to appear at the bottom of your own listing in the Sponsored Products banner. On DTC, the product page is entirely yours.
• Reviews and organic ranking: on Amazon, the organic ranking of your ASINs depends on sales velocity, number of reviews, listing conversion rate. A high-performing Sponsored Products campaign feeds organic ranking by boosting initial sales — that's the Amazon "halo effect". This mechanic has no direct equivalent on Google Shopping.

Amazon organic ranking mechanics (the halo effect): Amazon A9 algorithm weights organic ranking on Amazon Search SERPs based on conversion rate, sales velocity (7-30 day sales), price competitiveness, fulfillment (FBA Prime favored), reviews count and rating. A high-performing Sponsored Products campaign that pushes a new or losing-momentum ASIN regenerates sales velocity and can improve organic ranking by 30-150 places on SERPs in 2-6 weeks. This mechanic is the main argument justifying aggressive Amazon Ads investment for product launch (new ASIN launch) or listing recovery (ASIN whose ranking dropped). On Google Shopping, the equivalent exists but with a more diffuse halo effect — Google Merchant Center values regularly performing feeds and product Quality Score, but not with the same algorithmic ranking brutality.

Net margins: DTC website vs Amazon FBA

This is the heart of cross-channel budget arbitrage and it's what makes apparent ROAS largely misleading. The Amazon Sponsored Products ROAS displayed in Amazon Ads interface never tells the truth about real net margin, because it deducts neither Amazon commission, nor FBA fees, nor returns, nor indirect costs (annual Pro account, supplier support, catalog management).

Here's the net margin calculation mechanics per channel, on an example product at €50 incl. tax, COGS €15, sold in fashion (Amazon category 17% commission):

Critical reading: on this typical fashion example €50 incl. tax, the DTC net margin is 2.57x higher than Amazon net margin (€11.17 vs €4.34). The Amazon platform ROAS may show 7x or 8x apparent, but the net margin per Amazon sale remains structurally lower than the DTC net margin. That's why apparent ROAS is never enough to arbitrate cross-channel.

Net margin variations by category:

• Fashion/textile: 17% Amazon category commission, 10-18% returns, DTC net margin typically 1.8 to 2.8x higher than Amazon. Verticals where Amazon dilutes margin most.
• Beauty/cosmetics: 8% Amazon commission (Beauty category), 4-8% returns, DTC net margin typically 1.3 to 1.8x higher. More moderate gap.
• Mass-market electronics: 8% Amazon commission (Electronics category), 6-12% returns, DTC net margin typically 1.2 to 1.5x higher. Small gap — price competition pulls DTC margins down, the Amazon FBA logistics advantage partially compensates.
• Food/grocery: 8-15% Amazon commission depending on subcategory, 1-3% returns, DTC net margin often comparable or even lower than Amazon (own food logistics is expensive, FBA Pantry/Fresh can be more efficient).
• Tools/DIY: 12-15% Amazon commission, 4-8% returns, DTC net margin typically 1.5 to 2x higher. Verticals with bulky/heavy products where FBA fees are disproportionate.
• Premium/luxury (cart above €200): variable Amazon commission (often 8-15%), 8-15% returns, DTC net margin typically 2.5 to 4x higher. Verticals where DTC is nearly mandatory to preserve margin.

Direct operational implications: integrate into your business reporting a net margin calculation per source channel, going beyond the native platform ROAS. This requires precisely tracking: Amazon commissions per category, FBA fees per SKU, returns per channel, DTC logistics fees, payment fees, ad spend per source channel. Most mid-market e-com retailers do this calculation only once per quarter, or even once per year — creating an operational blind spot on cross-channel budget allocation.

Comparable ROAS and cannibalization

Google Shopping ROAS and Amazon Sponsored Products ROAS are not directly comparable, and their naive addition leads to systemic overestimation. Three structural biases make any cross-platform comparison delicate:

Bias 1 — Different ROAS definition between platforms: Google Ads calculates ROAS on conversion value pushed via GA4 or Enhanced Conversions (gross revenue incl. tax typically, or excl. tax depending on configuration). Amazon Ads calculates ROAS on revenue sold via the platform (typically gross revenue incl. tax, without commission/FBA fees deduction). On the same typical sale, Amazon ROAS is mechanically inflated by 15-25% compared to equivalent Google Shopping ROAS — commission and FBA fees are not deducted in the Amazon ROAS numerator.

Bias 2 — Different attribution window: Google Ads uses data-driven attribution (DDA) with 30-day post-click + 1-day post-view default window. Amazon Ads uses a 14-day post-click window for Sponsored Products (configurable up to 30 days), with last-touch attribution within the window. These different windows differently capture long decision cycles (premium, premium fashion, furniture) — Amazon mechanically underestimates cycles longer than 14 days, Google captures them better.

Bias 3 — Cross-channel cannibalization: a user searching "men's running shoes Nike Pegasus 41" on Google may click on the Shopping ad to the Nike DTC, or scroll to the bottom of the SERP and click on the Amazon Sponsored Products ad to the Amazon listing. The same user may also have seen a Meta Reels ad the previous week, triggering the search. Each platform will attribute the conversion to itself, creating massive cumulative over-attribution.

The Google Shopping ↔ Amazon cannibalization specifically: on referenced hybrid DTC + Amazon accounts, the query overlap between Google Shopping and Amazon Search internal reaches 35-55% depending on category. On this overlap portion, the user arbitrates between clicking on Google or going directly to search on Amazon. The direct consequence: scaling Amazon Sponsored Products beyond the incremental threshold mechanically cannibalizes Google Shopping (and vice versa), because both channels fish in the same pool of users with active commercial intent.

How to measure real cannibalization: quarterly cross-channel geo holdout test. Cut Amazon Sponsored Products on 30% of the territory for 28 days, measure the total revenue delta (DTC site + Amazon sales) in test regions vs control regions. If Amazon revenue test zone drops by 25%, but DTC revenue test zone bounces back by 8%, the Amazon → DTC cannibalization on this perimeter is around 32% of the Amazon delta (8/25). This allows calibrating the real incremental ROAS of Amazon vs its apparent platform ROAS.

Typical cannibalization case observed: on a mid-market fashion brand (revenue ~€12M annual, 60% DTC + 40% Amazon in revenue), Amazon Sponsored Products ROAS displayed was 6.8x. After 4-week holdout cutting Amazon on 4 regions (Brittany, Hauts-de-France, Bourgogne-Franche-Comté, Centre-Val de Loire), Amazon test zone drop was −38% from usual zone Amazon revenue, but DTC test zone bounce was +14% from usual zone DTC revenue. Effective Amazon → DTC cannibalization: 36% of Amazon delta (14/38). Consequence: real Amazon incremental ROAS was 4.3x, not 6.8x. On net margin, the gap widened further in favor of DTC, justifying a budget allocation rebalance toward Google Shopping for this brand.

For Google Shopping incrementality mechanics and Search-Shopping-PMax cross-channel deduplication, see our ROAS CPA CPC guide which lays the arbitrage fundamentals applicable to any paid mix.

Sponsored Brands, Sponsored Display, DSP: the full Amazon stack

Amazon Ads doesn't reduce to Sponsored Products. The 2026 Amazon Advertising platform exposes 4 distinct paid products covering awareness, consideration, conversion and retargeting. Mastering the full stack allows scaling beyond Sponsored Products when it caps in volume.

Sponsored Products: core format covered in section 2-4. Targets the Amazon Search internal query and Amazon product pages (bottom-of-listing placement). CPC format, manual targeting (keywords/ASIN) or auto-targeting. It's the format that covers 65-80% of the Amazon budget for most mid-market brands.

Sponsored Brands: display banner format at the top of Amazon Search SERPs (under the engine, above the results). Allows pushing an Amazon store front (multi-product branded showcase), a hero product, or a video. Targets Amazon Search keywords with mature commercial intent. Sponsored Brands ROAS is typically higher than Sponsored Products (because the top-of-SERP placement captures the first clicks), but available volume is more limited (single Sponsored Brands placement per SERP). Ideal for brand defense (bidding on your own brand) and category hunting (appearing at the top on category queries where you're not organic leader).

Sponsored Display: display retargeting + prospecting format on the Amazon inventory (product pages, Amazon home, Amazon app, and Amazon-owned external inventory like IMDb). Targets users having viewed your Amazon product pages without buying, or having bought a competitor ASIN in the category, or having interacted with similar listings. CPC or CPM format. Practicable from €2-3k/month Amazon monthly budget, it's the natural extension of Sponsored Products when you want to capture on-Amazon retargeting without moving to DSP.

Amazon DSP (Demand-Side Platform): programmatic platform that allows buying display + video + audio inventory on the Amazon ecosystem (Amazon, IMDb, Twitch, Fire TV, Audible) and on external inventory via partnerships. Premium format, typical minimum spend €35-50k/month to have a viable setup, custom audiences based on Amazon shopping signals (audience that bought in category X, audience that searched competitor ASIN Y). DSP opens audiences inaccessible to Sponsored Products/Display, but the opportunity cost (setup, team, reporting) often exceeds the gain for accounts under €30k/month Amazon budget.

Practical Amazon stack rules:

• Under €3k/month Amazon budget: Sponsored Products almost exclusively (90-100%), focus on optimizing manual targeting + auto-targeting and identifying best-performing keywords/ASINs. Sponsored Brands useful in brand defense (5-10%) if active brand competition.
• €3-12k/month: Sponsored Products 70-80%, Sponsored Brands 10-20% (brand defense + category hunting), Sponsored Display 5-15% (on-Amazon viewer retargeting).
• €12-35k/month: Sponsored Products 60-70%, Sponsored Brands 15-25%, Sponsored Display 10-20%. At this level, measure if DSP becomes practicable for custom lookalike prospecting.
• Above €35k/month: open DSP gradually (10-25% of Amazon budget), test for 90 days minimum before scaling. DSP requires heavier analytical setup and Amazon Marketing Cloud reporting tracking.

Branded Amazon Stores (free multi-product showcase) are an underrated complement of the paid stack. A well-built Amazon Store (4-6 pages, navigation by collections, brand videos, cross-referenced A+ content) increases the conversion rate of Sponsored Brands clicks by 25-45%, and allows retargeting Store visitors via Sponsored Display. Setup 4-8 hours for a 50-200 SKU catalog, near-immediate ROI on accounts having €5k+ monthly Amazon budget.

For brands with mature Amazon presence wanting to industrialize steering, see official documentation advertising.amazon.com Library which details best practices per format. Third-party tools (Helium 10, Jungle Scout, Pacvue, Skai) automate cross-format Amazon bid management but require a license investment €200-2,000/month depending on volume.

Allocation by profile: pure DTC, hybrid, pure marketplace

Based on differential net margins, attribution biases and cross-channel cannibalization, here is the practicable Google Shopping/PMax vs Amazon Ads budget allocation matrix for 2026. The matrix targets 3 typical business profiles covering most cases observed on the French mid-market e-com market.

Profile 1 — Pure DTC without Amazon presence, budget €5-50k/month. Allocation: 100% Google Ads + Meta + complement (TikTok, retargeting). No Amazon arbitrage. The Google stack typically breaks down into: Google Shopping/PMax 35-50%, Search 20-30% (brand + generic), Demand Gen + YouTube 10-20%, Display retargeting 5-10%. Typical verticals: luxury, premium-niche, very brand-driven brands, custom-made or complex products.

Profile 2 — DTC-dominant hybrid (Amazon 25-45% of revenue), budget €8-80k/month. Allocation: 60-75% Google + Meta DTC, 25-40% Amazon Ads. Google focus Shopping/PMax + Search brand + DTC site retargeting. Amazon focus Sponsored Products + Sponsored Brands brand defense. Typical verticals: mid-premium fashion, indie beauty, premium food, home/DIY equipment. DTC net margin typically 1.5 to 2.5x higher than Amazon, justifying DTC dominance.

Profile 3 — Balanced hybrid (Amazon 45-60% of revenue), budget €15-150k/month. Allocation: 45-55% Google + Meta DTC, 45-55% Amazon Ads. Typical verticals: mainstream fashion, mass-market beauty, mass-market electronics, food/grocery. DTC net margin typically 1.2 to 1.5x higher than Amazon — the margin gap no longer justifies clear DTC dominance. Almost-systematic 50/50 allocation with quarterly arbitrage by holdout.

Profile 4 — Amazon-dominant hybrid (Amazon 60-80% of revenue), budget €20-200k/month. Allocation: 25-40% Google DTC (often mostly brand defense + site visitor retargeting), 60-75% Amazon Ads (Sponsored Products + Sponsored Brands + Sponsored Display + possible DSP). Typical verticals: commodity electronics, consumables, commodity food/grocery, standard tools. DTC net margin often comparable or lower than Amazon (expensive own logistics vs FBA), justifying major investment on Amazon.

Profile 5 — Pure marketplace (no DTC site), all budget levels. Allocation: 100% Amazon Ads (sometimes extended Cdiscount Ads + Fnac Ads). No Google Shopping (no DTC merchant site to land on). Amazon stack: Sponsored Products 65-75%, Sponsored Brands 15-20%, Sponsored Display 10-15%, possible DSP. Typical verticals: white-label commodity products, brands without mature brand presence, pure Amazon FBA sellers.

Cross-channel hybrid e-com scaling rules:

• Measure differential net margin quarterly: recalculate net margin per source channel by integrating all variable fees. If the gap evolves (Amazon category commission changes, FBA fees increase, DTC logistics fees decrease), readjust allocation.
• Annual minimum cross-channel holdout test: cut Amazon on 25-30% of the territory for 28 days, measure real Amazon → DTC cannibalization, calibrate incremental Amazon ROAS vs platform ROAS.
• Strong seasonality: on Black Friday, sales, holidays, optimal ratios may drift by 10-20 points for 3-6 weeks (Amazon typically captures more in seasonal peak). Redo holdout post-season to validate the 2026 base allocation.
• New ASIN launch: temporarily allocate more on Amazon Sponsored Products to boost organic ranking (A9 halo effect), then rebase after 4-8 weeks.

For brands wanting to industrialize cross-channel Google + Amazon steering with consolidated net margin reporting, see our Google Ads vs Meta Ads 2026 allocation comparison which lays the cross-channel mechanics applicable to global paid mix. For Merchant Center feed and Google Shopping catalog fundamentals, see our Google Shopping setup and optimization guide.

Margin arbitrage methodology

Arbitrating paid e-com budget between Google Shopping/PMax and Amazon Ads in 2026 requires a methodology that goes beyond native platform ROAS. Here is the practicable protocol for a mid-market e-com account, broken down into 5 operational steps.

Step 1 — Calculate net margin per source channel per product category. For each product family or top-SKU, calculate: gross sale price incl. tax minus recoverable VAT minus Amazon commission (per category) minus FBA fees (per weight/dimension) minus returns (per category) minus DTC payment fees minus DTC own logistics minus COGS minus ad spend. Build a DTC vs Amazon net margin table per SKU or SKU family. Update quarterly.

Step 2 — Calibrate target ROAS per channel on net margin, not gross revenue. If DTC net margin is 22% of gross revenue and Amazon net margin is 9% of gross revenue, the minimum acceptable ROAS must be calibrated to reach the business target net margin. Example: to target 4% global net margin after ad spend, target Google Shopping ROAS will need to be around 5.5x (margin 22% → max ad spend 18%) and target Amazon ROAS will need to be around 9x (margin 9% → max ad spend 5%). This differential calibration is the basis of rational arbitrage.

Step 3 — Measure cross-channel cannibalization quarterly by holdout. Cut Amazon Sponsored Products on 25-30% of the territory for 28 days, measure DTC test zone bounce vs control zone. Calculate effective cannibalization ratio: (DTC test zone bounce / Amazon test zone drop) × 100. This ratio gives the % of Amazon revenue cannibalized from DTC. If above 30%, rebase allocation toward DTC; if below 15%, Amazon is largely incremental.

Step 4 — Implement DTC offline conversions tracking for margin measurement. Google side, configure Enhanced Conversions for Web and Merchant Center conversions to push revenue excl. tax per order. Amazon side, export Amazon Marketing Cloud (AMC) reporting or monthly Amazon Ads API to reconstruct net margin per campaign. Consolidate in a single business dashboard (Looker Studio, Power BI, or dedicated tool) crossing gross revenue + net margin + ad spend per source channel.

Step 5 — Readjust allocation by 90-day cycles on consolidated net margin. Each quarter, review cross-channel budget allocation based on consolidated net margin of the past 90 days. Never arbitrate on an isolated month (seasonality, launch, event create too much noise). The 90-day cycle absorbs cyclical variations while allowing adaptation to structural margin change.

Practicable 2026 operational rhythm:

• Weekly: operational steering Smart Bidding Google Shopping + manual bids Amazon Sponsored Products. Micro adjustments on target ROAS, negatives, exclusions.
• Monthly: performance review per channel, identification of best-performing keywords/SKUs, optimization of Merchant Center feed + Amazon catalog (title, A+ content).
• Quarterly: differential net margin recalculation, cannibalization holdout, cross-channel allocation readjustment, seasonality review.
• Annual: full Amazon stack audit (Sponsored Brands, Display, DSP), DTC vs marketplace long-term strategy review, annual budget planning.

Common mistakes in cross-channel e-com margin arbitrage:

1. Comparing Google vs Amazon platform ROAS without net margin normalization → systemic Amazon overestimation.
2. Ignoring Amazon → DTC cannibalization → Amazon over-allocation, DTC under-investment.
3. Not tracking FBA fees + Amazon commissions in Amazon ROAS → apparent ROAS inflated by 20-40%.
4. Steering Amazon without Sponsored Brands brand defense → competitor bids on your brand on Amazon.
5. Duplicating Google Shopping negatives to Amazon Sponsored Products → loss of category opportunity.
6. Reusing the same Google Merchant + Amazon catalog product feed without distinct optimization → sub-optimal title/description on both sides.

For hybrid e-com advertisers wanting to industrialize cross-channel arbitrage without restarting a manual calculation each quarter, our free Google Shopping + Amazon Ads audit consolidates net margin per source channel, identifies cannibalization patterns, and proposes an allocation plan per business profile. The report is delivered within 72h with actionable recommendations — target ROAS parameters to recalibrate, holdout test to launch, Amazon stack structure to optimize, target allocation per product category.

Building a coherent Google Shopping/PMax vs Amazon Ads allocation for 2026 is less a question of channel arbitrage than a question of net margin discipline. Apparent platform ROAS is misleading on both sides (inflated on Amazon by non-deduction of commission/FBA fees, biased on Google by non-deduction of DTC payment and logistics fees), net margin per sale after all fees is the only relevant business metric, and the only rational allocation derives from a quarterly consolidated net margin calculation + annual cannibalization holdout. E-com brands that steer on consolidated net margin and not apparent ROAS typically capture 18 to 32% of additional margin at identical paid budget. This differential is precisely what separates an e-com account driven for long-term profitability from an e-com account driven for vanity volume.

Sources

Official sources consulted for this guide:

• gs.statcounter.com
• gs.statcounter.com
• similarweb.com
• blog.google

FAQ