Skip to main content
Browse topics
Back to Docs

Ad Platform Playbook

Campaign Structure Principles

Branded vs. non-brand separation, Shopping vs. PMax tradeoffs, and intent-based organization

For · Operators planning or restructuring campaigns
Time · 7 min read
Next · Match type strategy

How you organize your campaigns shapes the data your bidding strategies have to work with. Structure decisions made early tend to compound: good segmentation surfaces clear performance signals, while poor segmentation can mask problems for months.

The Short Version

  • Keep branded and non-brand keywords in separate campaigns. Sharing one campaign lets strong branded performance mask unprofitable non-brand ad groups behind a healthy account-level CPA.
  • Standard Shopping and Performance Max are a control-versus-reach trade, not a better-versus-worse one. The table below names which side each account should land on.
  • Group by search intent, not only by product. Purchase-intent and research-intent queries carry different conversion rates and warrant different bid strategies.
  • Consolidation helps automated bidding until it merges segments with different intent, margin, or geography. The deciding question is whether a segment produces enough conversions to sustain its own campaign.
  • Campaign names follow [Channel] | [Brand-bucket] | [Intent], and *, ?, & are forbidden because they break downstream formulas and queries.

Branded vs. Non-Brand Separation

Branded search terms (your company name, product names) and non-brand terms (category searches, problem-based queries) behave differently. Branded traffic typically converts at higher rates and lower CPAs because those searchers already know you.

If branded and non-brand keywords share a campaign, the strong branded performance can mask underperforming non-brand terms. You may see a healthy account-level CPA while individual non-brand ad groups are unprofitable. Separating them gives you independent performance signals and independent budget control.

Shopping: Standard Shopping vs. Performance Max

For e-commerce advertisers, the choice between Standard Shopping and Performance Max involves a tradeoff between control and reach.

Standard ShoppingPerformance Max
Inventory reachedShopping placementsAll Google inventory from one campaign: Search, Display, YouTube, Gmail, Maps
Who controls structureYou. Custom product groups, bids per product typeGoogle’s bidding handles placements and bids
Query visibilityYou can see exactly which queries trigger which productsLimited visibility into which channels and queries drive results
Management loadHigher. The structure is yours to maintainLower. Fewer levers to manage
Choose this whenYou have well-defined product groups and enough volume to make the transparency actionableYour catalog is thinner, or management time is the binding constraint

Neither is categorically better. The trade is control for reach, and the right side of it depends on whether you have the volume and the time to use the control that Standard Shopping gives you.

Intent-Based Organization

Grouping campaigns by search intent, rather than just by product, can improve bidding signal quality. A campaign targeting “buy stainless steel water bottle” (purchase intent) behaves differently from one targeting “best water bottle for hiking” (research intent). Conversion rates, CPAs, and appropriate bid strategies may all differ.

When Consolidation Goes Too Far

Google generally recommends fewer, larger campaigns to give automated bidding more data to work with. That guidance has merit. Splitting volume across too many campaigns can leave each one below the conversion thresholds needed for effective automation.

But consolidation has limits. Combining campaigns with fundamentally different intent, margin profiles, or geographic targets can dilute the signals that bidding strategies rely on. The right balance depends on your conversion volume: if a segment generates enough conversions to sustain its own campaign, it’s often worth keeping separate.

Canonical Naming

Campaign names follow the schema [Channel] | [Brand-bucket] | [Intent]: pipe-delimited with spaces, title-case attributes. The six canonical campaign names Trellis recognizes:

  • Search | Brand | TM: trademark / exact-match brand defense (tightest match discipline)
  • Search | Brand | Evergreen: always-on broader brand-adjacent expansion
  • Search | NonBrand | Products: generic non-brand product search
  • Search | Competitor: rival-brand conquesting (singular; DKI must be disabled in ad copy)
  • Search | DSA: Dynamic Search Ads (channel prefix required so the name parses by the same regex as other Search campaigns)
  • Shopping | All Products: full-catalog Shopping

Forbidden characters in any campaign name: *, ?, &. These break Excel formulas and SQL queries. Avoid retired vocabulary like Core+, Brand+, Smart Shopping, Discovery, Similar Audiences, BMM, ETA, and Advantage+ Shopping; the platform-renamed forms are in active use today (Evergreen, PMax, Demand Gen, Optimized Targeting, Phrase Match, RSA, Advantage+ Sales / ASC).

Trellis Perspective

Trellis audits evaluate campaign structure in the context of your actual performance data, not general best practices alone. A structure recommendation is grounded in whether your current segmentation is producing clear, actionable signals or obscuring them.

Sources