Google Ads for E-Commerce: How to Structure Your Campaigns (Without Letting the Algorithm Eat Your Budget)

Most e-commerce brands running Google Ads aren’t losing because of bad products or weak margins. They’re losing because their campaign architecture is held together with duct tape — a Shopping campaign someone set up in 2021, a Search campaign with 400 keywords in one ad group, and a Performance Max campaign that Google convinced them to turn on and that now silently consumes 60% of their budget.

We’ve audited hundreds of e-com Google Ads accounts. The structural problems repeat themselves almost exactly every time. The good news: once you fix the architecture, performance often improves within weeks — before you touch a single bid or write a single new headline.

Why E-Commerce Campaign Structure Is Harder Than It Looks

Lead gen accounts have one job: generate a form fill or a phone call. E-commerce accounts have to do something harder — match buyer intent across an enormous range of product-specific queries, manage margin differences across SKUs, protect branded traffic, re-engage abandoners, and do all of that simultaneously across campaign types that increasingly overlap each other’s territory.

That complexity is also why the generic advice — “just run Shopping and add a remarketing campaign” — falls apart so fast for growing DTC brands. It’s not wrong exactly. It’s just two levels too shallow to be useful.

The brands that consistently win on Google are the ones treating their ecommerce Google Ads strategy as an architecture problem, not a settings problem. They’re deciding deliberately which campaign type owns which part of the funnel, how budget flows between them, and what guardrails prevent the algorithm from making decisions you’d never make yourself.

The Three-Layer Framework: Search, Shopping, and PMax Working Together

Think of your Google Ads account as a three-layer stack. Each layer has a job. Overlap is acceptable. Chaos is not.

Layer 1: Google Shopping — Your Highest-Intent, Lowest-Friction Driver

For most e-commerce stores, Shopping campaigns generate the most efficient revenue at scale. A shopper searching “women’s trail running shoes size 8” and clicking a Shopping ad with a photo, price, and your brand name is far closer to buying than someone clicking a text ad. The visual format does your pre-selling for you.

But Shopping campaigns live or die on two things most brands underinvest in: feed quality and campaign segmentation.

Your product feed is your “creative” in Shopping. Google can’t run a great Shopping campaign on a bad feed any more than you can run great Search ads with bad headlines. Product titles need to front-load the most important attributes — brand, product type, key specs — because that’s what Google matches against search queries. If your feed title says “Blue Widget 12oz” and the search is “12oz stainless steel water bottle blue,” you’re leaving impressions on the table.

On segmentation: don’t dump your entire catalog into one Shopping campaign and let Google decide what to push. Segment by product category, margin tier, or performance history — whichever dimension matters most for your business. This lets you assign meaningful ROAS targets per segment and pull budget toward your most profitable SKUs, rather than letting Google optimize toward your highest-volume, lowest-margin products. We have a full breakdown of Shopping ads optimization tactics that goes deep on feed structure and bidding if you want the full picture.

Layer 2: Search Campaigns — Intent Capture and Competitive Defense

Search campaigns in an e-com account serve three distinct roles, and mixing them together is one of the most common structural mistakes we see.

Role 1 — Branded Search: Run a dedicated branded campaign. Always. Even if your organic listing ranks #1, competitors are absolutely bidding on your brand name — and without a branded campaign, either they’re stealing your clicks or your own PMax campaign is serving instead (at higher cost and less control). Branded CPCs are cheap, conversion rates are high, and impression share should sit at 90%+. There is no argument for not running this.

Role 2 — Non-branded, high-intent product and category Search: These campaigns capture people who know what they want but haven’t chosen you yet. “Best ceramic cookware set,” “buy standing desk under $500,” “noise cancelling headphones for office.” These deserve tightly themed ad groups, strong product-specific ad copy, and landing pages that actually match the query — not your homepage. If you want to write ads that convert at this level of specificity, the principles in our guide to writing high-converting Google Ads copy apply directly.

Role 3 — Competitor conquesting (optional, and often overhyped): Bidding on competitor keywords in e-com can work, but the economics are brutal — low quality scores, high CPCs, skeptical searchers. It’s worth testing, but only after your core campaigns are dialed in. We’ve written an honest framework for deciding whether competitor campaigns are worth it that will save you from a common budget sink.

Layer 3: Performance Max — Amplifier, Not Foundation

Performance Max is the most misunderstood campaign type in e-commerce right now. Google pushes it hard. A lot of brands run it as their only campaign and wonder why results are volatile and attribution looks weird.

Here’s the honest take on PMax for e-com: it works best as an amplifier on top of an already-functioning Search and Shopping foundation. It can find incremental demand, surface your products across YouTube, Display, Discover, Gmail, and Maps, and scale what’s already converting. What it cannot do is substitute for intentional campaign structure.

The risks are real. PMax will preferentially eat your branded traffic if you don’t exclude it. It will default to your easiest conversions — often existing customers or people already deep in the funnel — unless you actively push it upstream. And its reporting is still opaque enough that diagnosing problems requires real skill. Before you scale PMax, you need to understand how to actually control it. Our complete Performance Max guide for 2026 covers asset group structure, audience signals, exclusions, and the specific settings that prevent it from quietly wrecking your efficiency.

Shopping vs. Search: When Each One Should Win the Budget Argument

The Shopping vs. Search question comes up constantly. The real answer: for most e-com stores, Shopping should get the larger share of non-branded budget — typically 60–70% — because it closes more efficiently at the bottom of the funnel. But Search has advantages Shopping can’t replicate.

Search wins when:

• Your product needs explanation before a shopper will convert (high consideration, complex spec, premium price point)
• You sell in a category where Google Shopping is thin or highly competitive and Search CPCs are actually cheaper
• You’re going after informational queries that suggest early-stage intent you want to own before they comparison shop
• You’re running a promotional campaign with time-sensitive messaging that Shopping’s static format can’t communicate

Shopping wins when:

• Your product is highly visual and the photo + price comparison does the selling for you
• You’re selling commoditized or well-understood products where price and availability close the deal
• You have strong feed data and can surface the right product for a long-tail query at scale
• Your conversion rates are higher on Shopping — which is typical and worth confirming in your own data

The deeper breakdown of this trade-off, including when to lean into one vs. the other based on product type and funnel stage, lives in our dedicated piece on Shopping vs. Search for Google Ads e-commerce.

The ROAS Target Mistake That Quietly Kills Ecom Campaigns

Setting your ROAS target too high is one of the most common ways a technically correct campaign structure falls apart in practice.

Here’s what happens: you set a 500% tROAS because your CFO wants it and it “sounds right.” Smart Bidding interprets that as a hard constraint and starts restricting auction participation aggressively. Impression share drops. You’re only competing for the very highest-intent, most obvious queries. Volume shrinks. You show your CFO the 520% ROAS number and feel great — but revenue is down 40% because you’ve basically turned your campaigns into a remarketing loop for people who were already going to buy.

A sustainable ROAS target for a Google Ads for online store account needs to account for your actual margin structure, your new-vs.-returning customer mix, and the role each campaign plays in the funnel. A top-of-funnel Shopping campaign should not be held to the same ROAS bar as a branded Search campaign. If you haven’t thought through how to set this number correctly, our guide on setting a realistic Google Ads ROAS target walks through the math and the mental model in detail.

Account Structure Best Practices for E-Commerce at Scale

Good campaign architecture isn’t just about campaign types — it’s about how your account is organized at every level, from campaigns down to ad groups to keywords and assets.

A few structural rules we don’t bend on for e-com accounts:

One clear job per campaign. A campaign that’s trying to do branded Search and non-branded Shopping simultaneously is a campaign you can’t optimize. Every campaign should have a single, clear objective, a single budget, and a single bid strategy aligned to that objective.

Ad groups organized by product theme, not keyword volume. If your ad groups are defined by whatever keywords you felt like adding, your Quality Scores will suffer, your ad copy won’t match your landing pages, and Smart Bidding won’t have clean signal to work with. Organize by product category or sub-category, match ad copy to that theme, and send traffic to the most relevant product or category page.

Negative keywords are non-negotiable. In e-commerce, the wasted spend from irrelevant queries is often catastrophic — people searching for information, for competitor products, for items you don’t sell, for DIY solutions. A rigorous negative keyword strategy is not a nice-to-have. It’s the difference between a 300% ROAS and a 150% ROAS on the same traffic volume.

Separate campaigns for new customer acquisition vs. retention. If you care about new customer ROAS (and you should — it’s a fundamentally different metric than blended ROAS), you need the structural ability to measure it separately. That means either separate campaigns or very disciplined use of audience exclusions for existing customers.

What to Measure — And What to Stop Letting Distract You

E-commerce Google Ads accounts generate an enormous amount of data. Most of it is noise. Here’s what actually tells you whether your structure is working:

Revenue and ROAS by campaign, not blended. A blended 400% ROAS can hide a branded campaign doing 1,200% and a Shopping campaign doing 180%. You need to see each layer independently.

Impression share by campaign type. If your branded campaign isn’t at 90%+ impression share, you’re losing traffic you already paid to earn through organic, social, and PR. If your top Shopping campaigns are below 50% impression share, budget or bid constraints are likely capping your upside.

Search term reports, weekly. What queries is Google matching to your campaigns? In e-com, this is where you find both new negative keywords and new positive keyword opportunities you hadn’t thought to target. PMax makes this harder — another reason to maintain dedicated Shopping and Search campaigns where you have full visibility.

New vs. returning customer conversion split. If 70% of your Google Ads conversions are from existing customers, your campaigns are functioning as a retention tool, not a growth engine. That’s not necessarily wrong, but you need to know it and make deliberate decisions about it.

Frequently Asked Questions

Should I run Shopping campaigns or Performance Max for my e-commerce store?

Both — but in the right order and with the right relationship between them. Start with well-structured Standard Shopping campaigns to establish conversion data and ROAS benchmarks. Once Shopping is profitable and stable, layer in Performance Max to find incremental volume across Google’s broader network. Running PMax first, without Shopping giving you baseline data and control, is putting the amplifier before the signal.

How many Shopping campaigns does an e-commerce store need?

More than one, almost certainly. The right number depends on your catalog size, margin structure, and performance variation across product categories. As a starting point: separate campaigns for your highest-margin products, your highest-volume products (which may not be the same), and any categories with meaningfully different ROAS targets. A 10-SKU store might run 2 campaigns. A 10,000-SKU store might run 15. There’s no universal number — it depends on where you need independent budget and bid control.

Is a branded Search campaign worth running for a small e-commerce store?

Yes. Even if you’re small. Branded CPCs are typically $0.20–$1.00 for most e-com brands. Your competitors are either bidding on your name already or will be once you grow. And your PMax campaigns will steal branded traffic if you don’t carve it out explicitly. The ROI on a branded Search campaign is almost always the best in your account. Run it.

What’s the right budget split between Search, Shopping, and Performance Max?

A reasonable starting framework for an established e-com account: 60–70% to Shopping (the highest-intent, most efficient format), 20–25% to Search (branded plus high-intent non-branded), and 10–15% to PMax to test incremental reach. Adjust based on where your actual ROAS data tells you to invest more. Don’t let PMax grow above 30–35% of total budget until you have strong confidence in its incrementality — and know how to measure that.

How do I stop PMax from cannibalizing my Shopping and Search campaigns?

A few specific tactics: add your branded keywords to a negative keyword list applied to your PMax campaign. Segment your PMax asset groups by product category so Google has focused signals rather than a free-for-all. Monitor your Standard Shopping impression share — if it drops after you launch PMax, PMax is eating your Shopping traffic. And use audience signals aggressively to push PMax toward new customer types rather than remarketing to your existing audience.

What ROAS target should I set for my Google Ads e-commerce campaigns?

Start with your break-even ROAS — the point where ad spend equals gross margin. If your average gross margin is 50% and your average order value is $100, your break-even ROAS is 200%. Set your initial target 20–30% above break-even to protect profitability while still giving Smart Bidding enough room to find volume. Resist the urge to set aspirational targets like 600% or 800% until you have enough conversion data (50+ conversions per month, per campaign) to support that kind of optimization without starving reach.

If Your Current Setup Doesn’t Look Like This, It’s Worth Getting a Second Opinion

Campaign architecture is the thing most in-house teams and generalist agencies get wrong — not because they’re not trying, but because e-commerce Google Ads structure is genuinely complex, and the defaults Google pushes you toward are designed to maximize Google’s revenue, not yours.

If your Shopping campaigns are one big catch-all with no segmentation, your PMax is running without branded exclusions, and you’re not sure what percentage of your conversions are new customers vs. returning — you don’t have a bidding problem or a creative problem. You have a structure problem. And structure problems don’t get fixed by tweaking a headline or adjusting a bid modifier.

The fix is rebuilding the architecture with a clear-eyed understanding of what each campaign type is supposed to do, what it’s not supposed to do, and how they’re supposed to hand off to each other. That’s exactly what we do for e-commerce brands.

If you want us to take a look at your account structure and tell you honestly what’s working and what’s broken, request a free account audit. We’ll review your campaign architecture, ROAS by layer, Shopping feed quality, and PMax configuration — and give you a straight answer on whether your setup is built to scale.