Paid Media · Pillar guide
The Complete Guide to Paid Media for Ecommerce
By Milton von Hesse · Updated July 2026 · 22 min read
How to structure, fund and measure paid media for an ecommerce brand: break-even ROAS, channel selection, campaign structure and budget allocation.
Most ecommerce brands do not have a paid media problem. They have a math problem that shows up in their paid media account.
They set a ROAS target because someone said 4:1 was good. They spread budget across five channels because a competitor is on all five. They judge campaigns on a 7-day window because that is the default. Then they wonder why revenue grows while profit does not.
This guide covers how to structure, fund and measure paid media for an ecommerce brand, in the order the decisions actually need to be made. It assumes you sell physical or digital products online and that you care whether the spend is profitable, not just whether it produces sales.
What is paid media for ecommerce?
Paid media for ecommerce is the practice of buying advertising placements that drive traffic to an online store, with the goal of generating profitable revenue rather than reach or awareness alone. The main channels are Google Ads (Search, Shopping and Performance Max), Meta Ads (Facebook and Instagram), Amazon Ads for brands selling on the marketplace, and LinkedIn Ads for high-consideration or B2B products.
Ecommerce paid media differs from brand advertising in one structural way: every transaction is tracked, so performance can be measured against margin instead of estimated. That makes it accountable in a way most marketing is not, and it also makes it unforgiving. If the unit economics do not work, no amount of creative or optimization will fix it.
The one number that determines everything: break-even ROAS
Before choosing a channel, before writing an ad, calculate this. Almost every failed paid media program I have seen skipped this step.
Return on ad spend (ROAS) is revenue divided by ad spend. Spend $1,000, generate $4,000, that is a 4:1 ROAS.
The problem is that 4:1 is meaningless on its own. Whether it makes you money depends entirely on your contribution margin.
The formula
Break-even ROAS = 1 ÷ Contribution Margin
Where contribution margin is the percentage of revenue left after the cost of goods, payment processing, shipping and fulfillment, everything that scales with each additional order.
What that looks like in practice
| Product type | Contribution margin | Break-even ROAS | 4:1 ROAS verdict |
|---|---|---|---|
| Digital product | 85% | 1.18:1 | Very profitable |
| Beauty / supplements | 70% | 1.43:1 | Very profitable |
| Apparel | 55% | 1.82:1 | Profitable |
| Consumer electronics | 30% | 3.33:1 | Barely profitable |
| Furniture (free shipping) | 22% | 4.55:1 | Losing money |
Two brands can both report a 4:1 ROAS. One is compounding, the other is quietly funding its own decline. The number on the dashboard is identical.
Working example
An apparel brand sells a jacket for $120.
- Cost of goods: $42
- Payment processing (2.9% + $0.30): $3.78
- Shipping and fulfillment: $9
- Contribution per unit: $65.22 (54.4% margin)
Break-even ROAS = 1 ÷ 0.544 = 1.84:1
At 4:1, this brand keeps roughly $54 per jacket sold after ad spend. At 2:1, it keeps about $5. At 1.5:1, it is paying for the privilege of shipping jackets.
Set your target ROAS above break-even by the margin you need to cover fixed costs and profit. For most brands that means targeting 1.5x to 2x your break-even number, not an arbitrary industry benchmark.
Which channels to use, and in what order
The instinct to be everywhere is the most expensive mistake in ecommerce paid media. Each channel needs enough budget and enough conversion volume to exit the learning phase and optimize. Splitting $3,000 across four platforms gives you four underfunded campaigns and zero usable data.
Start with one channel. Prove it. Then add the next.
Google Shopping and Performance Max
Best for: Products people actively search for. If someone types "waterproof hiking boots," they have already decided to buy something in that category.
This is the highest-intent traffic available and usually the first channel to test. You are capturing demand rather than creating it.
Start here if: Your product solves a problem people know they have and search for by name.
Meta Ads (Facebook and Instagram)
Best for: Visually distinctive products, impulse-friendly price points, and categories where people do not know the product exists yet.
Meta creates demand rather than capturing it, which makes it harder but also expands your total market beyond existing search volume.
Start here if: Your product is visual, novel, or in a category with thin search volume.
Amazon Ads
Best for: Brands already selling on the marketplace with reviews in place.
The intent is exceptional, Amazon users are in buying mode. The tradeoff is that you build Amazon's customer relationship, not your own, and you compete with your own store.
Add this when: Marketplace is already a meaningful revenue channel, not as a first test.
LinkedIn Ads
Best for: B2B products, wholesale programs, or high-ticket items with a considered purchase cycle.
CPCs run several times higher than Meta. It only works when order values justify the cost.
Skip this if: You are consumer DTC with an average order value under a few hundred dollars.
How to structure campaigns that can actually be optimized
A well-structured account makes optimization obvious. A poorly structured one makes it guesswork.
Segment by margin, not by product category
This is the highest-leverage structural decision most brands get wrong. Grouping products by category ("all outerwear together") feels tidy but hides the thing that matters. If your outerwear line ranges from 25% to 70% margin, one target ROAS cannot be correct for both.
Group products by contribution margin band instead. Each band gets its own campaign and its own ROAS target, calculated from its own break-even. Now the algorithm is optimizing toward a number that is actually true for those products.
Separate branded from non-branded search
People searching your brand name were already coming. Mixing those conversions into your prospecting campaigns inflates your reported ROAS and disguises whether your acquisition is working.
Run branded search as its own campaign, measure it separately, and never let it justify a non-branded budget decision.
Give campaigns enough volume to learn
Every platform's algorithm needs conversion data to optimize. Below roughly 30 to 50 conversions per month per campaign, the system cannot distinguish signal from noise and your optimizations are effectively random.
If you cannot fund a campaign to that threshold, consolidate. Fewer, better-fed campaigns beat many starved ones.
Budget allocation and the testing reserve
A useful default for a brand past initial validation:
- 70% to proven campaigns. The ones already hitting target ROAS. This is your engine.
- 20% to scaling. Pushing winners harder to find where efficiency breaks down. It always breaks down eventually.
- 10% to testing. New channels, new audiences, new creative angles. You should expect most of this to fail.
The testing reserve is what brands cut first when results dip, and it is exactly the wrong instinct. Your current winners will decay, creative fatigues, competitors enter, platforms change. The 10% is what replaces them. Cutting it optimizes this quarter at the cost of next year.
Measurement: what to actually watch
Attribution windows change the story
The same campaign will report wildly different ROAS depending on the window. A 1-day click window shows only immediate purchases. A 28-day window credits sales weeks later.
Neither is wrong. But the window must match your buying cycle. A $30 impulse product and a $2,000 considered purchase should not use the same window. Pick one based on your actual time-to-purchase, document it, and stop changing it. Most reporting confusion comes from silently comparing numbers measured different ways.
Blended ROAS versus platform ROAS
Every platform over-reports. Meta claims a sale, Google claims the same sale, and if you add up platform-reported revenue it exceeds your actual revenue. That is not fraud, it is each platform counting conversions it touched.
Blended ROAS, total revenue divided by total ad spend across all channels, is the number your bank account agrees with. Platform ROAS is useful for optimizing within a channel. Blended ROAS is the truth about whether your program works.
Watch both. Trust blended.
The metric that matters more than ROAS
ROAS measures efficiency on the first purchase. It says nothing about whether the customer comes back.
Customer acquisition cost against lifetime value is the actual business question. A brand with strong retention can profitably acquire at a ROAS that would bankrupt a brand with none, because the second and third purchase carry no acquisition cost.
This is why paid media and email marketing are not separate disciplines. Your email program determines how much you can afford to pay for a customer. Brands that treat them as separate line items systematically underinvest in acquisition or overspend on it.
Common mistakes worth avoiding
Optimizing to a target that was never calculated. If your ROAS goal did not come from your margin, it came from nowhere.
Judging campaigns before they exit the learning phase. Two weeks of data on a new campaign is noise. Let it gather conversions before drawing conclusions.
Letting branded search inflate the picture. It flatters your numbers and hides acquisition problems.
Ignoring creative as a performance lever. At scale, creative variation drives more incremental gain than bid tuning. Most accounts are over-optimized and under-tested creatively.
Bundling management fees with ad spend. If a vendor quotes one number covering both, you cannot verify how much reaches the platform. Always demand the split in writing.
Frequently asked questions
What is a good ROAS for ecommerce?
There is no universal good ROAS. The right target is derived from your contribution margin. Calculate break-even ROAS as 1 divided by your contribution margin, then set a target above it with enough headroom to cover fixed costs and profit. A brand with 70% margins can thrive at 2:1, while a brand with 25% margins loses money at 4:1.
How much should I spend on paid media?
Enough for each campaign to reach roughly 30 to 50 conversions per month, or the algorithm cannot optimize. In practice this means most brands should fund one channel properly rather than spreading budget thin across several. Start with the channel where your customers already show intent, prove profitability, then expand.
How long before paid media shows results?
Paid media generates data immediately, so measurable signal typically appears within 2 to 4 weeks. Reliable optimization requires longer, since campaigns must exit the learning phase and accumulate enough conversions to distinguish real performance from variance. Anyone guaranteeing profitable scale in days is misrepresenting how the platforms work.
Should I run Google or Meta first?
Start with Google Shopping if people actively search for your product category, since you are capturing existing demand at high intent. Start with Meta if your product is visual, novel, or sits in a category with low search volume, since Meta creates demand rather than capturing it. Running both underfunded is worse than running one properly.
Why does my platform-reported ROAS not match my revenue?
Each platform claims credit for conversions it touched, so adding up platform-reported revenue across channels overstates actual revenue. Use blended ROAS, total revenue divided by total ad spend, to evaluate whether the overall program is profitable, and platform-reported ROAS only to optimize within a single channel.
Where to go from here
If you take one thing from this guide, calculate your break-even ROAS today. Most brands discover their target was wrong by a factor of two, in one direction or the other. Everything downstream, channel choice, budget, campaign structure, what counts as a win, depends on that number being right.
If you want a second set of eyes on your account, Tirna's Diagnostic covers exactly this: a full audit of paid media, email, conversion and tracking, with a prioritized roadmap of what to fix first.
About the author
Milton von Hesse is the founder of Tirna, an ecommerce marketing consultancy. He holds a Post-Degree Diploma in Marketing from Douglas College in Canada and has spent 8+ years running paid media, email and CRO for brands across North America and Latin America. At Imprint Plus he restructured a $1M+ annual ad program, lifting ROAS from 250% to 400%, and has taught digital analytics and search advertising at the university level.
Related guides: The Complete Guide to Ecommerce Email Marketing · The Complete Guide to Ecommerce CRO