Amazon Ads vs Google Shopping: Where Product Budgets Win

Amazon Ads and Google Shopping sell product clicks at nearly the same moment of purchase intent, and the trade is cleaner than most channel matchups: Amazon Sponsored Products capture buyers already inside the store at roughly $0.90–1.00 per click with ACOS norms of 25–35% (Adbadger and published agency medians), while Google Shopping buys open-web reach, lands the click on a site you own, and leaves the customer relationship — email, remarketing, repeat purchase — in your hands. Which channel deserves the next dollar is margin math: Amazon's fee stack against your DTC fulfillment costs, single-order economics against lifetime value.

What does each channel actually do?

Amazon Ads is shelf-space advertising. Sponsored Products put your listing inside marketplace search results and on competitor product pages; Sponsored Brands and Sponsored Display extend the same intent upward and across the funnel. Every click resolves to a purchase that happens on Amazon — with stored payment details, Prime shipping, and a trust default that unknown DTC sites spend real money to earn. For a large slice of physical-product categories, the shopping journey now starts inside Amazon's search bar, which makes these placements feel closer to slotting fees than to classic advertising: you are buying position at the exact point of purchase.

Google Shopping is open-web product advertising. Feed-driven listings surface across Search, Images, YouTube, and the display network — today mostly orchestrated through Performance Max — and the click lands on your own product page. Conversion becomes your problem and your prize: pure ecommerce Shopping/PMax activity converts around 2–4% (WordStream/LocalIQ), because your site has to do the convincing a marketplace checkout gets for free. In exchange, every completed order creates a customer record you can email, remarket, and grow for years.

Intent is nearly identical on both sides; economics and ownership are opposites. That single asymmetry drives everything below, the same way the capture-versus-create line drives Google Ads vs Facebook Ads.

How do the numbers compare?

If ACOS is new vocabulary: it is ad spend divided by attributed sales — the inverse of ROAS — so a 30% ACOS is a 3.3x ROAS, and the 25–35% norm translates to roughly 2.9–4x. Both channels price clicks through the same kind of auction, mechanics covered in our CPC glossary entry, and our free ROAS & Break-Even Calculator converts between ACOS and ROAS while marking the break-even line your margins actually require.

Resist the urge to crown the cheaper click. A $1.00 marketplace click and a $1.50 open-web click buy different things: one buys a conversion-optimized checkout with fees attached, the other buys a prospective customer with acquisition friction attached. Cost per profitable outcome, after each channel's fee stack, is the only comparison that survives contact with a P&L.

One macro trend belongs in the picture: retail media is the fastest-growing major ad channel, compounding at roughly 20%+ a year per eMarketer/GroupM forecasts, and growth like that shows up as auction pressure. The numbers behind the trajectory — and what they mean for Amazon CPCs over the next few years — live in our retail media statistics roundup.

What does Amazon's fee stack do to the margin math?

The channel comparison is really a fee comparison, and it deserves pencil-level honesty. Take a $40 product with a $12 landed cost, sold both ways — round numbers, purely an illustration.

Sold on Amazon: the referral fee (typically 15% for most categories, per Amazon's published schedule) takes $6.00, FBA fulfillment takes roughly $6 for a small item, and the $12 landed cost leaves about $16 of contribution before advertising. Break-even ACOS is contribution over price: 40%. A 30% ACOS — mid-range against published norms — spends $12 of ad cost on the sale and nets roughly $4 a unit.

Sold DTC via Google Shopping: payment processing takes about $1.20, pick-pack-and-ship takes $7, leaving roughly $19.80 of contribution before ads. Break-even ROAS is price over contribution: about 2.0x. A 3.0x ROAS spends $13.33 on the sale and nets about $6.50 — and the order carries an email address, which is where the comparison stops being close. Email drives 25–30% of ecommerce revenue for mature brands (Klaviyo), all of it attached to customers acquired on a store you control.

Same product, same appetite for ad spend, different winners depending entirely on fee structure and repeat behavior. Two refinements keep the model honest. First, judge Amazon on TACOS (ad spend over total Amazon revenue) as well as ACOS, because paid velocity lifts organic rank — a falling TACOS is the flywheel paying you back. Second, credit the DTC side with a defensible slice of lifetime value rather than single-order margin, or subscription and replenishment brands will systematically underfund the channel that builds their file.

Where does Amazon cap brand building?

Four ceilings, worth naming plainly before budget season:

• The customer stays Amazon's. No email addresses, no off-platform remarketing audiences, thin post-purchase data. Winbacks, cross-sells, and reorder flows all run inside Amazon's ecosystem, on Amazon's terms.
• Price parity pressure. The buy box rewards the lowest competitive price, so your Amazon price disciplines your DTC price whether you planned that or not.
• A crowded shelf you pay to defend. Competitors advertise on your product pages and against your brand terms inside the marketplace, so defensive placements become a recurring tax on traffic you generated.
• Rented rank. Organic position follows sales velocity, so pausing ads slows the flywheel. The asset you are building compounds inside Amazon rather than on your own domain.

None of that argues against the channel. It argues for clarity about which asset each dollar builds: Amazon spend builds rank and velocity inside the marketplace; Google Shopping spend builds a customer file and first-party data on property you own.

When does each channel deserve the budget?

The conditions matter more than the averages, because category economics swing this verdict hard in both directions:

How do you run both as a portfolio?

Score each channel on contribution after its own fees rather than on platform-reported efficiency, then push budget to whichever clears its break-even at the margin. In practice: scale Amazon while marginal ACOS stays under break-even and TACOS holds or falls; scale Google Shopping while marginal ROAS clears the DTC line with a defensible LTV credit attached. Our Media Mix Planner pressure-tests any split against editable channel benchmarks, and the Paid Media Benchmarks report supplies the surrounding medians so neither channel gets graded on rumor.

Two operational notes from running this pair in live accounts. Measurement will disagree: Amazon's closed loop reports cleanly on itself while Google's results land in your own analytics, and reconciling both into one weekly contribution view is tedious, structured work — exactly the shape of job we map to automation in chatbots vs AI agents. And proven structure travels: the import-and-adapt economics that make Microsoft Ads vs Google Ads a cheap second search channel apply to your product feed as well, so a working Shopping setup extends further than it first appears.

This is one of twenty-four matchups in our marketing comparisons hub, and holding a marketplace-plus-DTC portfolio at the profit-maximizing edge — fee math, attribution reconciliation, saturation watch — is the daily work of a performance media practice.