Google Ads
Google Ads account structure for ecommerce
Build the account around the decisions you will actually make: brand separation, catalog segmentation, conversion actions, budgets and naming.
By CartKernel · Published
Account structure is the set of boundaries that let you make a decision. If you cannot pause a segment, fund it separately or give it a different return target, it does not need its own campaign. If you can and you want to, it does. Every other reason for splitting an account, including tidiness and habit, costs conversion volume that automated bidding needs.
Start from the decisions. Then draw the campaigns.
The jobs an ecommerce account has to do
Most stores are running five distinct jobs at once, and they have different economics:
- Defend the brand. Capture people who already know the store and typed its name.
- Sell the catalog. Put products in front of people searching for what you sell, priced by the shopping auction.
- Find new demand. Reach people describing a problem rather than a product.
- Return to people who visited. Remarketing across the surfaces the account can reach.
- Build awareness. Video and demand generation where the store has a reason and a budget for it.
The first two produce the majority of revenue for most catalogs, and they behave nothing alike. Brand queries convert at high rates because the decision was already made elsewhere. Catalog queries are the auction the store actually competes in. A structure that mixes them reports an average that describes neither.
Separate brand from anything that can absorb it
Brand traffic is the cheapest, highest-converting inventory in the account, which means any campaign allowed to serve on it will look excellent and will spend there first. Keep it in its own campaign, with its own budget and its own target, and keep every other campaign out of it with negative keywords or the brand controls your campaign types offer.
That includes broad-coverage campaign types that serve across surfaces. When a catalog-wide campaign is free to serve on brand queries, its reported return rises without any additional sale, and the store loses the ability to see what non-brand prospecting actually costs. The symptom and the remedies are covered in Performance Max spending on brand, the standing decision behind it is in should an ecommerce brand bid on its own name, and the campaign itself is described in brand search campaigns.
One campaign per decision, and no more
The tension in every ecommerce account is between control and learning. More campaigns give more control over budget and targets. Fewer campaigns give each bid strategy more conversions to learn from. Resolve it with one test: would you set a different budget or a different return target for this segment tomorrow? If not, it belongs inside an existing campaign as an asset group or ad group.
A workable starting shape for a mid-sized catalog:
| Campaign | Purpose | Typical target logic |
|---|---|---|
| Brand search | Capture existing demand for the store name | Efficiency target, funded to full coverage |
| Shopping or Performance Max, core catalog | Products with healthy margin and stock depth | The store’s return target |
| Shopping or Performance Max, secondary catalog | Low margin, clearance, thin stock | Lower spend, different target or excluded |
| Non-brand search | Category and problem queries with commercial intent | Acquisition target, lower than core |
| Remarketing | People who visited and did not buy | Efficiency target |
Small accounts should collapse this further. A store with a modest budget is usually better served by brand plus one catalog campaign, because splitting a small conversion count across five campaigns leaves every bid strategy short of data. Is Performance Max worth it for small stores works through that trade-off, and how much should a store spend on Google Ads covers the budget floor.
Segment the catalog by economics, not by department
The temptation is to build campaigns that mirror the site navigation. Navigation reflects how shoppers browse. Campaigns should reflect how products earn. Group products by margin, price band, stock depth, seasonality and role, and carry that grouping into the feed as custom labels so the same segmentation is available in every campaign type.
That gives you the ability to say, honestly, that clearance products can run at a thinner return while flagship products carry the store, rather than averaging both into a single target that fits neither. The method is set out in margin-based Shopping segmentation, and where query-level control is needed within Standard Shopping, query sculpting across Shopping campaigns covers the priority mechanism.
Get the conversion actions right before anything else
Structure built on bad conversion data optimizes toward the wrong thing at speed. Three rules:
- One primary conversion action for bidding. Purchase, with accurate transaction value passed at the order level. Everything else, including add to cart, sign-ups and phone clicks, is recorded as a secondary action for observation.
- Value that reflects revenue, and ideally margin. Passing a fixed value per order tells the bid strategy that a low-margin accessory is worth the same as a flagship product.
- One measurement path. Duplicate tags on a Shopify theme produce inflated counts that no structure can correct, which is the situation described in why does Google Ads report more conversions than my store.
Set the return target from real economics rather than from a habit. The break-even ROAS calculator gives the floor below which a campaign is buying revenue at a loss, and target ROAS bidding explains how the target becomes a bid.
Name campaigns so reports read themselves
A naming convention is worth an hour once. Use a fixed order of fields, separated consistently, so anyone can filter a report without knowing the account:
Market | Channel | Segment | Intent | Type
For example: CA | Shopping | Core | Non-brand | PMax. The value is not neatness. It is that a filtered view of every non-brand campaign across every channel is one text filter away, which is what makes the brand and non-brand reporting in brand versus non-brand possible without manual tagging.
Budgets, bid strategies and the settings that quietly split spend
- Budget per campaign, not shared, wherever the campaigns have different jobs. A shared budget lets the campaign with the easiest conversions take the money.
- Fund brand to full coverage first, then spend the remainder on acquisition. Brand impression share below full coverage is usually the cheapest gap in the account.
- Separate search and display, and check that search campaigns are not opted into partner networks you did not choose.
- One market per campaign where currency, shipping, tax display or margin differ. Averaging two markets with different economics under one target produces a target that is wrong in both.
- Language and location settings reviewed on creation, since the defaults are broader than most stores intend.
Restructure only when a decision is blocked
Rebuilds are expensive: bid strategies re-enter learning, historical data becomes harder to compare, and the account spends a period of weeks less efficiently. Restructure when you can name the decision the current structure prevents, such as being unable to fund a product line separately or unable to see non-brand cost. Do not restructure because the account looks untidy.
When you do rebuild, change one thing at a time, keep the naming convention, and record the date so the performance change has a cause you can point to. The wider program, including the feed work the catalog campaigns depend on, sits in Google Ads, Performance Max for ecommerce and ecommerce search campaigns.