eCommerce Fastlane

Free every Thursday · 59,694 Readers

Fastlane Insider Newsletter: What's working for Shopify in the AI era?

Ecommerce Advertising for Shopify Brands: Where Your Next Ad Dollar Should Go

Steve Hutt
Published:
November 16, 2026
Clay render of a coin on a cost dial linked by one blue line to one of four channel pedestals, a storefront
A 9x ROAS on your own brand name is usually a receipt for customers who were already on their way to buy.

Quick Decision Framework

  • Who This Is For: Shopify founders and marketers doing roughly $500K to $10M a year who already spend a few thousand dollars a month on Meta or Google and are being pitched Amazon ads, TikTok Shop, connected TV or another retail media network.
  • Skip If: You are still finding product market fit and have not yet run paid ads with any consistency. The product page and one working email flow come before any of this.
  • Key Benefit: One number (your maximum new customer cost) and one score (the readiness assessment) that tell you which channel gets the next dollar and which pitch to turn down.
  • What You’ll Need: Your average order value, product and shipping costs, payment fees, return rate, and a 90 day repeat purchase rate from Shopify’s customer reports.
  • Time to Complete: 13 minutes to read. About two hours to set the number and score your channels.

Matt Raminick, founder and CEO of the growth agency Sunnyside, said on episode 454 of the podcast that one of the first things his team checks when it audits a new ad account is who the ads are actually reaching: “And it’s really easy to inflate ROAS, you just target past customers.” That one sentence explains most of the paid media conversations that go sideways at the $500K to $10M stage.

The pitch you hear this year is diversification. Retail media budgets are growing quickly, Amazon and Walmart have closed loop purchase data, and connected TV is now sold through self serve platforms as well as big agency buys. Some of that is true at the market level. None of it tells you whether your next dollar should go there, because the question a Shopify brand needs answered is narrower: does this channel buy new customers for less than they are worth to you?

This piece is for the brand that already runs Meta or Google and is weighing what comes next. It gives you the number to set first, the channel order by stage, a scoring method from Thriving with Shopify for any new network, and a labelled worked example where the right call is to add less.

What Ecommerce Advertising Covers for a Shopify Brand in 2026

Ecommerce advertising for a Shopify brand now spans three groups of channels: paid social, search and shopping ads, and commerce media, which is advertising sold by a business that also sells or processes the purchase. Each group buys a different kind of attention, so they are not interchangeable line items in one budget.

Paid social (Meta, TikTok, Pinterest, Snapchat) creates demand from people who were not looking for you. It rewards brands that can produce new creative every week, and it is where many Shopify brands spend their first serious money. Search and shopping ads (Google, Microsoft) capture demand that already exists, which tends to make them efficient when people search for your category by name and weak when they do not.

Commerce media is the newer term, and it confuses people because it covers several businesses at once. Retail media networks are the biggest part: Amazon Ads, Walmart Connect, Target’s Roundel and dozens of grocery, pharmacy and delivery networks selling ads against their own shoppers and purchase data. The money is concentrated. EMARKETER’s December 2025 forecast puts US retail media spending at $71.09 billion in 2026 and has Amazon and Walmart capturing 89% of the year’s incremental spend. For a Shopify brand, that means the retail media question is mostly an Amazon question, with Walmart a distant second.

Shopify runs a commerce media program of its own. Shop Campaigns places your products in the Shop app, on shop.com, across other Shopify stores through the Shopify Product Network and on optional third party placements, and as of November 2026 Shopify’s help center says campaigns run at an average customer acquisition cost you set and you are charged only when a customer converts, with no upfront budget or subscription fee. Paying per conversion against a target you chose is a different deal from buying clicks or impressions, which is why it belongs in this conversation even though few people call it commerce media.

Social commerce (TikTok Shop, shoppable Instagram) sits across the line between paid social and commerce media. Its content and creator economics are their own topic, covered in the social commerce archive, so this piece treats it only as a channel to score.

Why Your Ad Platform ROAS Looks Better Than Your Bank Account

Platform ROAS looks better than your bank account because it credits every sale the ad touched, including sales from returning customers and people who were already searching for your brand. The number that decides profit is what you paid for each new customer, and most dashboards do not show it by default.

Three guests on the show described the same problem from three angles. Ben Dyer, head of growth at Webtopia, told the story of a hair care brand on episode 451 whose branded search campaign was spending real money: “it’s super easy to stick a brand campaign in there and just capture everybody that’s coming to it and be like, We’re hitting a 10x for us. But I suppose the real side of that is people were going to buy anyway.” His team moved that budget into finding people who had never heard of the brand. Fastlane’s look at branded search cannibalization covers how to size that kind of campaign with a test instead of a dashboard.

Scott Desgrosseilliers, founder and CEO of the attribution company Wicked Reports, explained on episode 421 why automated Meta campaigns drift the same way. When new customer cost comes in high on an Advantage campaign, he said, “it probably means the conversion event you set up for that campaign isn’t just sending new customer sales, it’s sending repeat as well. And then what happens is the pixel learns and bids on the repeat because they’re cheaper and faster and all that.” The algorithm is doing what you asked. You asked for purchases, and your own customers are the cheapest purchases on the platform.

Raminick’s point closes the loop: an agency or a team judged on in channel ROAS will drift toward past customers without anyone deciding to. The fix is not a better dashboard. It is a different scoreboard: new customer acquisition cost (ad spend divided by first time buyers), first order contribution margin, and a blended view across channels. If you have not already split channel performance by profit, the POAS versus ROAS breakdown shows how a cheap channel can still need the highest ROAS to break even.

When I was a Senior Merchant Success Manager at Shopify, the paid media reviews that went badly usually started with a blended ROAS nobody had split into new and returning buyers. The account looked healthy right up until someone asked how many of those orders came from people who had never bought before.

Set Your Maximum New Customer Cost Before You Choose a Channel

Your maximum new customer cost is the first order contribution margin plus the contribution you can reasonably expect from repeat orders inside a payback window you can afford. Any channel that buys customers below that number is worth testing, and any channel above it is losing money no matter what its ROAS says.

Here is the math on an illustrative store, not a real one, assuming a $70 average order, a 10% first order discount, product cost of $21, $9 to pick, pack and ship, Shopify Payments at the Basic plan’s US online rate for standard cards of 2.9% plus 30 cents (as of November 2026), and a 4% return allowance. The discounted order brings in $63. Subtract $21, $9, about $2.13 in payment fees and about $2.52 in returns, and the first order contributes roughly $28. If one in four new customers orders again within 90 days and that second order contributes about $30, each new customer is worth another $7.50 inside the window. The 90 day break even is about $35.85, so a $35 target leaves a little room, and both numbers sit before fixed overhead and before any profit you want from the customer.

Now compare that to what an ad platform would tell you. At a $35 cost per new customer, first order revenue of $63 is a new customer ROAS of 1.8. That is roughly break even for this store over 90 days, and it would look alarming in a report where returning customers are lifting blended ROAS to 3 or 4. A brand reading only blended ROAS can end up switching off a campaign that buys customers near its break even point while funding retargeting that mostly buys orders it already had.

The window is a stage decision. A Starting brand should hold to first order break even or close to it, because cash is the constraint. A Growing brand can use 90 days once its repeat rate is measured rather than hoped for, using the cohort reports covered in the retention playbook. Scaling and Leading brands can stretch to six or twelve months only if the business can carry the gap, which is a cash question the ecommerce accounting pillar walks through. Write the number down. It is the ceiling you check any average CAC target against (Shop Campaigns, a new customer Meta campaign) and the line any agency proposal has to beat.

Which Channel Gets the Next Dollar at Your Stage

The next dollar goes to the one new channel where your customers already show intent and your team can do the work well, which for most Shopify brands means Meta or Google Shopping first, Shopify’s own pay per conversion program alongside, and a marketplace only once you sell there. Channels that already work keep running while you test the next one.

Stage
Where the next dollar goes
Hold off on
Starting, under $500K a year
One channel, Meta or Google Shopping
Branded search, retargeting, every marketplace
Growing, $500K to $2M
Second channel, Shop Campaigns under your ceiling
Connected TV, a third attribution tool
Scaling, $2M to $10M
Amazon test if you already sell there
More than one new network per quarter
Leading, $10M and up
Walmart, connected TV, measured with holdouts
Channels nobody owns full time

Choose Meta first if your product looks good in a few seconds of video and someone on the team can ship new creative every week, because creative is what the platform now targets with. Choose Google Shopping first if people already search for your category by name (a replacement part, a supplement ingredient, a specific shoe width), because you are buying demand that exists instead of creating it.

Choose Shop Campaigns alongside either one if your store is eligible and you want to pay per conversion against an average acquisition cost you set in advance. Shopify says setup needs no new assets, which suits a Growing team stretched thin. An average target is not a guaranteed price for every customer, so treat its results like any other channel’s: check the realized cost per new customer against your ceiling. If Shopify Plus is in your near future for other reasons, Shopify’s Audiences tool is limited to eligible Plus stores in the US and Canada using Shopify Payments as of November 2026; the Shopify pricing breakdown covers when that upgrade pays for itself, and an audience tool alone rarely does.

Choose Amazon Sponsored Products only if your products are already listed on Amazon at a margin you are happy with. Choose connected TV only if you can measure it with a geographic holdout, because branded search lift is easy to claim and hard to prove. And wherever the click lands on your own store, the page has to match the ad, which the guide to paid traffic landing pages covers in detail.

Retail Media Is a Marketplace Decision Before It Is an Ad Decision

The retail media most Shopify brands get pitched (Amazon Sponsored Products, Walmart Sponsored Search) is a marketplace decision first, because those ads promote listings you sell on that marketplace, and selling there changes your margin, your customer relationship and your channel mix before a single ad runs. Decide whether the marketplace belongs in your business, then decide how much to advertise on it.

The entry requirements make the order clear. Amazon’s seller advertising page says Sponsored Products needs a Professional seller account, and Sponsored Brands, display ads and Brand Stores also need a brand enrolled in Amazon Brand Registry. Walmart Marketplace sellers buy Sponsored Search through Walmart Connect’s self serve ad tools, which sit behind Walmart’s own seller approval. Many of the other networks are sold mainly to brands that already supply that retailer.

There is one exception worth knowing. Amazon Ads says brands that do not sell on Amazon can still buy display, video and audio through its DSP, including ads that send shoppers to your own site. That is a reach and awareness budget, closer to connected TV than to Sponsored Products, and it should be judged the same way: with a holdout, against your new customer ceiling.

Danan Coleman, CEO of the Amazon agency eCom Triage, put the cultural gap plainly on episode 402: “Most Shopify sellers and Amazon sellers, they are not the same person, and they are definitely not the same business.” Amazon rewards product selection and constant ad and listing management. A Shopify brand that treats it as one more ad account usually underinvests in the work and overestimates the return.

Chapter 6 of Thriving with Shopify names the two ways retail media hurts a Shopify brand that does go in. The first is dependence: when Amazon grows past roughly half your revenue, one suspension or policy notice can freeze the business, so the chapter’s rule is to keep any single platform you do not own below 50 to 60% of revenue. The second is cannibalization. Amazon revenue climbs, the Amazon ads report looks great, and a quarter later blended margin is down because customers who used to buy on your store at full margin now buy the same product on Amazon. The defence is selective catalog presence: list your proven sellers or Amazon-only bundles, which may reduce direct substitution, keep the full range and the best margin on your own store, and treat a falling Shopify new customer count while Amazon grows as a warning sign to investigate, not proof on its own.

I’d argue most brands under $2M a year should leave marketplace ads alone until the Shopify side has a known maximum customer cost and a measured repeat rate. Without those two numbers, you have no baseline for telling whether Amazon is adding customers or relabelling them.

Score Any New Channel Before You Fund It: A Worked Example

Score a new channel on five criteria from zero to ten each (customer presence, product and platform fit, resources you have today, margin after all costs, and room to scale). A total of 40 or more is a priority one channel to start with now, 30 to 39 is worth testing once your first channel is working, 20 to 29 waits until you are strong elsewhere, and under 20 is a no for the next 12 to 18 months.

This is the Platform Readiness Assessment from Chapter 6 of the book, and its value is that it forces an honest resource score. Here it is applied to a labelled composite, built from patterns that come up repeatedly on the podcast and in merchant conversations, not a real store: a skincare brand at about $1.4M a year, roughly 80% of paid spend on Meta, a branded search campaign reporting 9x, and three pitches on the desk for Amazon, TikTok Shop and connected TV.

Channel pitched
Presence, fit, team, margin, scale
Call
Amazon Sponsored Products
7 + 8 + 4 + 6 + 7 = 32
Priority two: test once Meta is fixed
TikTok Shop
6 + 7 + 3 + 4 + 6 = 26
Priority three: wait for creator capacity
Connected TV
4 + 5 + 2 + 3 + 4 = 18
No for 12 to 18 months

All scores here are illustrative. Nothing clears 40, and that is the useful answer: the brand’s real opportunity is inside the account it already runs. Its Meta campaigns optimize for all purchases, so on the pattern Desgrosseilliers described the pixel may well have drifted toward repeat buyers; switching the conversion event to first time customers and judging results against a ceiling like the $35 one from the earlier section costs nothing. Its branded search campaign is the kind Dyer described, so a geographic holdout (a few weeks is a planning window; the real test length depends on order volume and how comparable the regions are) can show how much of that 9x would have happened anyway. If the regions are too small to show a clear difference, the result is inconclusive, and the campaign stays capped rather than cut or scaled.

What I’d remove before adding anything: pause the connected TV conversation entirely, cap retargeting at a fixed share of spend instead of letting it grow on its own, and drop the third attribution tool the team is trialling, because a third dashboard that disagrees with the first two does not settle anything. Amazon is a priority two channel, so it waits until Meta is working. When Meta has been buying new customers under the ceiling for long enough to trust the number, and someone on the team has the hours Amazon’s thin resource score says are missing, rescore it and run the test with an Amazon-only bundle, which lowers the chance of pulling buyers away from the store’s best seller.

This is the premature complexity pattern I watched repeat with brands at the $500K to $2M stage during my Shopify years: more channels before the first one is measured properly. Every new network adds a dashboard, a reporting argument and a person’s time, and none of that shows up in its ROAS.

How to Measure When Every Platform Claims the Same Sale

Measure paid media with three numbers you control rather than the platforms’ reports: blended ad spend per new Shopify customer, blended marketing efficiency (total revenue divided by total ad spend), and periodic holdout tests that switch a channel off in some regions to see what it actually adds. The first two tell you whether the whole program is healthy; only the third tells you which channel caused the customers.

Every platform takes credit for every sale it touched, so Meta, Google, Amazon and Shopify will never add up to the same total, and they are not supposed to. Stop trying to reconcile them. Pull first time customers by month from Shopify’s customer reports and divide your store focused ad spend by that count. That is a blended number, not a channel CAC: it includes customers who found you organically, it leaves out buyers who purchased on a marketplace (keep marketplace ad spend out of it too), and it covers media only, not agency fees or creative costs. Its value is that no single platform can inflate it. Track it beside blended efficiency so you can see whether growth is coming from new buyers or from spending more on people you already had.

Use holdouts for the channels where the doubt is largest, and decide before the test how big a difference would count, because a test without enough orders in each region ends inconclusive. Branded search, retargeting and connected TV are the usual suspects, because all three show up just before purchases that might have happened anyway. A post purchase survey asking how a first time buyer heard about you is cheap and directional, and it often surfaces podcast, creator or word of mouth sources no pixel can see. Raminick’s episode walks through checking several sources against each other rather than trusting one; at the Growing stage, a spreadsheet with those three inputs is usually enough, and a dedicated attribution product earns its place when spend is large enough that a 10% reallocation pays for it.

Before any of these numbers mean anything at higher spend, the store has to hold up under the traffic. If you are about to double a budget into a launch or peak season, the guide to what breaks when traffic doubles covers the checks to run first. And keep paid placements separate in your head from being recommended inside AI answers, which is earned rather than bought and lives in the generative engine optimization pillar.

Names on this page may link through eCommerce Fastlane /recommends/ links, which can earn a commission.

Frequently Asked Questions

What is ecommerce advertising?

Ecommerce advertising is paid media that drives online purchases, and for a Shopify brand it covers paid social, search and shopping ads, and commerce media such as Amazon Ads, Walmart Connect and Shopify’s Shop Campaigns. Paid social creates demand from people who were not looking for you, search captures demand that already exists, and commerce media reaches shoppers inside a retailer or shopping app using that business’s purchase data. The useful test for any of them is the same: does the channel buy new customers for less than the contribution margin they return within a payback window you can afford? Judging channels by platform ROAS alone tends to reward ads that reach your existing customers.

What is the difference between commerce media and retail media?

Retail media is the largest part of commerce media: ads sold by retailers such as Amazon, Walmart and Target against their own shoppers and sales data. Commerce media is the broader term and also includes ads sold by other businesses that see purchases, such as delivery apps, payment companies, travel brands and Shopify’s own Shop Campaigns, which places products in the Shop app and across other Shopify stores. For a Shopify brand the distinction matters less than the concentration. EMARKETER forecasts US retail media spending of $71.09 billion in 2026, with Amazon and Walmart taking 89% of the growth, so most retail media decisions come down to whether you sell on those two marketplaces.

How much should a Shopify store spend on ads?

A Shopify store should spend on ads only up to the point where each new customer costs less than their first order contribution plus expected repeat contribution inside your payback window. Start from your numbers, not a percentage of revenue: average order value, product cost, shipping, payment fees, returns and discounts give you first order contribution, and Shopify’s customer reports give you a repeat rate. In an illustrative store with a $70 average order, about $28 of first order contribution and a 25% repeat rate, the 90 day break even is roughly $36 per new customer before fixed overhead. Spend can grow while new customers keep arriving under that ceiling, and should stop growing when they do not.

Should a Shopify brand advertise on Amazon?

A Shopify brand should run Amazon’s product ads only after deciding that selling on Amazon belongs in the business, because Sponsored Products requires a Professional seller account and Sponsored Brands also needs Brand Registry. Amazon can add customers who start their product search there, but it can also move your own customers from your store to a lower margin channel. Put proven sellers or Amazon-only bundles on the marketplace, keep the full range on Shopify, and keep any single platform you do not own below roughly half your revenue. Watch whether your Shopify new customer count holds steady while Amazon grows. If it falls, treat that as a warning that Amazon may be relabelling customers rather than adding them, and test before you conclude it.

Why is my ROAS high but my profit low?

ROAS can be high while profit is low because the platform counts every sale its ads touched, including returning customers and shoppers who searched your brand name and would have bought anyway. That is how a branded search campaign can report 10x, the figure one agency guest described on the eCommerce Fastlane podcast, and automated campaigns optimizing for all purchases can drift toward repeat buyers because they are the cheapest conversions available. ROAS also ignores product, shipping, payment and return costs. Split results into new and returning customers, divide ad spend by first time buyers from Shopify’s reports, and compare that cost to first order contribution margin. That comparison shows whether the ads are growing the business.

FIND US ONLINE

WEEKLY DTC INSIGHTS

TRUSTED BY THOUSANDS

TRUSTED PARTNER

Choose a language