Google Ads Account Structure for eCommerce: How to Set It Up Properly

Campaign structure is the question eCommerce advertisers ask most and get the fewest straight answers to. Every guide says something different, Google’s own advice changes every year, and the right setup really does depend on your account: budget, catalogue size, conversion volume and how hands-on you want to be.

This guide gives you a clear starting structure for a Shopify store, the reasoning behind each choice, and when it makes sense to move away from the defaults.

The Core Principle

Structure exists to give Smart Bidding clean, accurate signals. That is its main job. A good structure keeps different types of traffic apart so Google’s algorithms can optimise each one without the signals blending together and producing poor decisions.

Brand searches (people typing your store name) convert at a very different rate from non-branded commercial searches such as “best waterproof backpack”. Put both in one campaign and the bidding model has to average them. It ends up under-bidding on brand traffic (strong conversion rate but limited volume) and over-bidding on non-branded traffic (because brand’s high conversion rate inflates the average).

Keep traffic with different intent in separate campaigns. Everything else follows from that rule.

The Starting Structure for Most eCommerce Stores

For a Shopify store with an established catalogue and enough monthly conversions for Smart Bidding (20+ a month), start here:

Campaign 1: Brand Search

• Type: Search
• Match types: exact and phrase match on your brand name and its variants
• Bid strategy: Maximise Clicks or Manual CPC (brand campaigns convert so reliably that Smart Bidding adds little; the aim is a low CPC, not conversion optimisation)
• Budget: small (brand traffic is low volume and cheap)
• Purpose: capture people searching for your store at low cost and stop competitors taking that traffic

Campaign 2: Performance Max (Non-Branded)

• Type: Performance Max
• Brand exclusions: applied (your brand terms excluded)
• Bid strategy: Maximise Conversion Value, moving to Target ROAS once you have 50+ conversions
• Budget: your main acquisition budget
• Purpose: broad non-branded acquisition across Shopping, Search, Display and YouTube
• Audience signals: Customer Match buyer list, website visitors, and in-market audiences relevant to your category

Campaign 3: Competitor Search (Optional)

• Type: Search
• Keywords: exact and phrase match on direct competitor brand names
• Bid strategy: Maximise Conversions or Manual CPC
• Budget: small test budget
• Purpose: reach people researching competitors who might switch

Competitor campaigns are worth testing if you are clearly different from specific competitors and can write ad copy that makes the comparison compelling. They are not a default for every store.

That is the baseline: three campaigns (two without competitors), clean separation of intent, brand exclusions on PMax, and Smart Bidding driving non-branded growth.

When to Add More Structure

The baseline suits most accounts. These are the situations that justify more complexity.

Separate PMax campaigns per product category:

If categories differ significantly in margin, conversion rate or target ROAS, separate PMax campaigns give you budget and target control for each.

For example, a store selling high margin skincare (target ROAS 400%) and lower margin supplements (target ROAS 250%) cannot hit both targets in one PMax campaign with a single ROAS target. Split them, set a suitable target for each, and budget them independently.

Do not split simply because you have several categories. If margins and targets are similar across the catalogue, one PMax campaign works better: more impression data, more conversion data, better optimisation.

Standard Shopping for your most valuable products:

If five to ten products are your highest margin and best performers, consider running them in a Standard Shopping campaign with specific bids, excluded from PMax using product filters. You get precise bidding on your most important products while PMax handles the rest.

The catch: PMax still wins the auction when a product is eligible in both campaigns. To split product groups cleanly, you must use PMax product filters to exclude the products you want in Standard Shopping.

Separate campaigns for international markets:

If you sell into several countries with different prices, conversion rates or currencies, separate campaigns per market are cleaner than one global campaign. Geo-targeting within a single campaign can work, but market-specific campaigns with their own budgets and ROAS targets give better control and clearer reporting.

A dedicated remarketing campaign:

A Display or Demand Gen campaign aimed only at past visitors (basket abandoners, product page viewers) with its own budget keeps retargeting spend visible and under control. PMax includes retargeting automatically, but if you want to set the retargeting budget yourself rather than let PMax split it between prospecting and retargeting, a separate campaign is the way to do it.

What Not to Do

Do not build a campaign per product. That is over-segmentation. A single product campaign never builds enough conversion data for Smart Bidding to optimise, and it adds management work without improving results.

Do not run Shopping and PMax on the same products without exclusions. Without product-level exclusions, both campaigns bid on the same searches and you cannot tell which performs better. Run one or the other, or split the catalogue between them explicitly.

Do not run several PMax campaigns on the same products without different signals or targets. PMax campaigns covering the same products with the same signals compete with each other, pushing up your own CPCs and fragmenting conversion data. If two PMax campaigns do not differ meaningfully in product scope, target ROAS or audience signals, merge them.

Do not keep campaigns that spend under £5 a day. An underfunded campaign does not get enough impressions for Smart Bidding to work. Consolidate the budget into a campaign that can spend properly, or pause it.

A Naming Convention That Works

Name campaigns so their purpose is obvious without opening them. One convention:

[Market] - [Type] - [Scope] - [Match/Bid]

Examples:

• UK - Search - Brand - Exact
• UK - PMax - All Products - tROAS
• UK - Shopping - High Margin Products - tROAS
• US - PMax - Full Catalogue - MaxConvValue

Consistent names make filtering, Looker Studio reporting and handovers to colleagues far easier.

Ad Groups and Asset Groups

In Search campaigns, organise ad groups by theme: a tight cluster of related keywords with ads written for that theme. One ad group per keyword is too granular, and one ad group for everything is too broad. A useful test is whether every keyword in the ad group could sensibly share the same headlines and descriptions.

In PMax, organise asset groups by the audience signal they target or the product category they feature. If one asset group uses a Customer Match signal and another an in-market signal, give each creative and landing pages that suit that audience.

How to Split the Budget

As a starting point, put 80 to 90% of your Google Ads budget into the non-branded PMax campaign. Brand Search is efficient but small, and 5 to 10% of budget is usually enough to capture all brand traffic. Competitor and other campaigns share what is left.

This split will change as you optimise. If PMax spends efficiently and ROAS is at or above target, give it more budget. If PMax is underspending (common early in Smart Bidding’s learning phase), hold off on extra budget until it has enough conversion data to use it well.

Review the Structure Every Quarter

Structure is not a one-off decision. As your catalogue grows, conversion volume rises and Google’s campaign types evolve, the best structure changes.

A quarterly review should ask:

Are any campaigns consistently underfunded compared with their potential? Consolidate them or raise their budget.

Are any campaigns producing too little conversion data for Smart Bidding? Merge or remove them.

Has a new campaign type become relevant? Demand Gen, for example, is now a real option for stores with visual products and budget for upper funnel activity.

Are brand exclusions and product filters still set correctly? They tend to drift when new products are added or campaigns are duplicated.

The best structure is the simplest one that separates genuinely different traffic and gives every campaign enough data to optimise. Start simple, add complexity only for a specific reason, and review it regularly.