Contents 13
Build Google Ads structure around catalog economics, goals and data quality. Margin, stock, seasonality and product roles determine the control you need. Two stores using the same Search, Shopping and PMax campaign types may need different architectures. Check your products and conversions before copying another account.
Three campaign types, three forms of control
Search connects query intent to keywords, ads and a landing page. Standard Shopping builds product ads from Merchant Center data; bids and exclusions apply to product groups. PMax combines campaign goals, assets, a feed and signals across Google inventory.
Give every campaign a job. Separate budgets are useful when goals, markets, economics or priorities need independent control. A site category does not automatically require its own campaign. PMax asset groups are not separate budgets or individual product bids.
| Campaign | Control unit | Advertiser controls | Useful when |
|---|---|---|---|
| Search | Keywords and ads | Intent, matching, copy and landing page | Queries and messaging need direct control |
| Standard Shopping | Product or product group | Group inventory, bids and exclusions | Product-level control is needed |
| Performance Max | Goals and asset/product groups | Goals, feed, assets, signals and exclusions | Automation across Google inventory is useful |
The table does not assign universal roles. Tasks can overlap in a real account. Product-group bids are available with the appropriate strategy, such as manual CPC; they do not replace automated strategy goals.
How Performance Max interacts with Search
Priority depends on query matching and eligibility. An eligible exact Search keyword identical to the query takes priority over PMax. Identical phrase/broad keywords and search themes can share priority; relevance and Ad Rank then matter.
Budget limits, keyword status, approval and targeting can change eligibility. Naming a campaign Brand Search does not automatically reserve every brand query. Separate brand/non-brand control helps budgeting and analysis; exclusions still need actual query evidence.
Google: Search keyword priority ↗
Segment products using feed data
Use category, product_type, brand, condition, item id and custom_label_0–4 where they help control. Standard Shopping uses product groups; PMax uses listing groups within themed asset groups.
Custom labels are five optional internal fields. Define their meaning, source and refresh rule instead of treating them as mandatory for every product. Submit identifiers such as GTIN according to category and manufacturer requirements; do not invent them to fill a field.
| Dimension | Attribute | Control | Review question |
|---|---|---|---|
| Category or product type | category / product_type | Product/listing group inventory | Are different product roles mixed together? |
| Brand and condition | brand / condition | Groups, bids or exclusions | Does this affect sale value? |
| A specific SKU | item id | Product inclusion or exclusion | Does it need separate control? |
| Margin, season, priority | custom_label_0–4 | Filters, reporting and bidding | Who updates the label, and is it current? |
Margin and product role are an operating method
Margin, seasonality and stock can inform working labels. This is a team decision, not a Google requirement to divide every catalog into four groups. Use reliable data and identify estimates explicitly.
One PMax campaign is not inherently wrong. Consolidation may suit a small or consistent catalog. Do not split campaigns only because of SKU count; consider distinct goals, budget needs, messages and signal quality.
Standard Shopping for testing and PMax for scaling is one scenario, not a rule. Another campaign can provide control, but may unnecessarily fragment a small catalog.
Check the foundations before automation
Reconcile feed prices and availability with product pages. Make a test purchase and check transaction id, value, currency, duplicates and the actual order. Conversion value is not automatically profit.
For store sales, purchase is usually the primary optimization action while add_to_cart and views remain diagnostic. Check each campaign goal: custom goals can use selected actions differently. Customer data and enhanced conversions require appropriate permission, lawful use and consent configuration.
Check three layers before restructuring.
- Product data: title, price, availability, GTIN and other attributes must match product pages and the actual catalog.
- Conversion tracking: purchase, amount, currency and transaction id must arrive correctly without duplication.
- Conversion goals: the primary goal must represent the business outcome, not arbitrary intermediate activity.
Google Merchant Center setup guide →
Build the structure step by step
Map your catalog and markets, verify measurement and choose useful product attributes. Record each segment’s job, goal, landing page, included products, budget, exclusions, owner and review rule.
Prepare the architecture without enabling it. Check queries against ads, feed and pages, plus payment and delivery readiness. Make structural changes deliberately, with a baseline and rollback plan, rather than repeated small resets.
- Describe catalog economics: margin, stock, seasonality and product-group roles. Record unknown data as a limitation.
- Reconcile purchase and value with store orders. Check amount, currency, duplicates and campaign goals.
- Put useful dimensions into the feed. Assign a label owner and update rule.
- Give each segment a role and choose the appropriate campaign type. Do not create a separate campaign without a separate task.
- Record boundaries for brands, queries, products and URLs. Every exclusion needs a reason.
- Allocate budget by priority and actual demand. Another store’s metric is not your target CPA or ROAS.
- Set a review cycle for queries, products, goals, feeds and labels. Keep baseline data and a rollback plan before changes.
Search terms, negative keywords and brands
Distinguish account-wide negatives from campaign boundaries. PMax negatives apply to relevant Search/Shopping inventory; brand exclusions address brands. These controls are not interchangeable.
Do not automatically exclude reviews, affordable or similar words: they may indicate purchase research in your niche. Evaluate query, landing page and goal together. Audience signals guide PMax learning but do not restrict delivery only to those audiences.
The Search terms report shows actual queries. PMax search terms/insights access and completeness may vary by interface and report. Use available evidence, and check whether exclusions also reduce useful reach.
Google: PMax negative keywords ↗
Budgets and bidding without a magic threshold
Choose for your goal: conversion count or value, with appropriate CPA/ROAS targets where suitable. Data requirements and evaluation guidance differ. A sample of 30 or 50 conversions does not guarantee success.
Since March 2025, eCPC is unavailable for Search and Display. Target-strategy labels began changing in June 2026; interfaces may temporarily use different names without changing the underlying mechanics. For broad match, we recommend Smart Bidding and verified conversion signals.
Interface ROAS or CPA does not establish profitability without margin, returns, logistics and other costs. Set advertising targets against your own financial model.
- Is the budget sufficient for relevant auctions?
- Do purchase and value arrive consistently?
- Have important product prices or stock changed?
- Does the bidding goal fit segment economics?
- Are restrictive targets or exclusions limiting the campaign?
Google: Target bidding names ↗
PMax budget scenario calculator →
Diagnose the reason before restructuring
Review products, queries, goals and spend. Mismatched orders, weak coverage of an important segment, stale labels or campaigns without distinct roles warrant investigation. One symptom does not prove one cause.
Standard Shopping campaign priority applies to overlapping Standard Shopping campaigns promoting the same products in the same country. It is not a PMax control. PMax/Shopping overlap follows auction and Ad Rank considerations; normal impression distribution is not a controlled experiment.
Review the structure when these symptoms appear.
- PMax serves brand queries the team intended to control through Search.
- Advertising conversions rise without more confirmed store orders.
- One group mixes products with different economics and stock.
- A priority segment has weak coverage because of budget or grouping.
- Custom labels no longer describe the catalog.
- Campaigns share products and goals without distinct roles.
Google: Standard Shopping priority ↗
Google Ads audit → · Public audit sample →
UPLIFY’s approach to account structure
Start with catalog data, measurement and a task map. Then explain why a segment exists, what affects its budget and which evidence will be reviewed after changes. This is a transparent working method, not a profit promise tied to a campaign mix.
Use the Google Ads audit offer and public sample for an independent account review. A budget calculator helps build a scenario from your assumptions; its output does not guarantee sales.
Google Ads structure questions
What if a product is in Standard Shopping and PMax?
Check whether overlapping inventory serves a defined purpose. Separate products through item id, custom labels or listing/product groups if needed. Standard Shopping priority does not control PMax; ordinary campaign impression distribution is not a controlled experiment.
How should margin labels work with manually updated costs?
Define the data source, owner and update frequency. A controlled broad classification may be more useful than a stale precise number. Do not make margin the sole decision factor when data lags behind changes.
Does a small store need a complex account structure?
Not automatically. Start with a minimal structure where each campaign has a role. Separate budgets should reflect real differences in goals, markets or priorities, not every category or SKU count.
What if add_to_cart and purchase are both primary conversions?
Check measurement and each campaign’s goals first. Purchase is usually the main store-sales goal; use intermediate events for diagnostics. Also check custom goals, where selected actions can be used differently. Then assess bidding and structure.
Do out-of-stock products need a separate campaign?
Usually not. Check availability, feed updates, the product page and Merchant Center. A separate campaign will not fix mismatched data or update delays.
Do audience signals restrict PMax targeting?
No. They are learning suggestions, not hard targeting limits. PMax can serve outside those signals when it considers an audience relevant to the campaign goals.
Sources and scope of recommendations
Mechanics were checked against Google documentation. Margin segmentation and the review sequence are UPLIFY working methods, not a mandatory Google account template.
- Google: Performance Max ↗
- Google: PMax / Search ↗
- Google: Search keyword priority ↗
- Google: Standard Shopping priority ↗
- Google: Custom labels ↗
- Google: PMax negative keywords ↗
- Google: Audience signals ↗
- Google: Target bidding names ↗
- Google: eCPC ↗
Next step
Discuss your account structure
Describe your catalog, markets and goal. We will agree data checks and next steps before making advertising changes.