Campaign Architecture

DTC Google Ads Campaign Architecture

A proven framework for structuring high-performance Google Ads accounts for DTC brands, derived from accounts spending $100K-$200K+/day profitably.

Core Philosophy

Build a PMax-first account with aggressive audience segmentation, systematic naming conventions, and device-level isolation. Treat each campaign as an independent test cell with its own audience, offer, landing page, and budget — then scale winners through duplication.

Campaign Type Mix

Allocate spend across campaign types using this ratio as a starting point:

Performance Max handles the heavy lifting. Demand Gen extends reach via YouTube at low CPMs. Search is only for brand terms if needed.

Audience Segmentation Framework

Segment campaigns by audience persona. Each segment gets its own campaign(s) with tailored creative, landing pages, and offers.

Primary segments to test for any DTC brand:

• Women — largest addressable market for most health/wellness/beauty DTC
• Seniors (55+) — high intent, high AOV, responsive to urgency
• Moms — responsive to "reset" and transformation messaging
• Men — often lower volume but can be efficient
• Geographic — state-level campaigns for localized offers or compliance

Naming convention pattern:

{platform}-{price/offer}-{audience}-{sequence}-{modifier}

Examples:

gpm-$149-women-4           → PMax, $149 offer, women audience, 4th iteration
gpm-$149-senior-1-phone-excl → PMax, $149, seniors, 1st, phone traffic excluded
gpm-$149-feb-13            → PMax, $149, February launch, 13th iteration
directmeds-seniors-3       → PMax, DirectMeds brand, seniors, 3rd iteration
gdg-youtube-seniors-8      → Demand Gen, YouTube, seniors, 8th iteration

Encode the offer price, audience, launch date/event, iteration number, and any device modifiers directly in the campaign name. This makes performance analysis instant at a glance.

Device Segmentation Strategy

Device performance varies dramatically. Isolate devices at the campaign level:

Three campaign tiers by device:

1. All-device campaigns — default starting point, let PMax optimize
2. Phone-exclusion campaigns (-phone-excl) — exclude mobile phones to isolate desktop + tablet traffic, which often converts at lower CPA
3. Tablet-only campaigns (-tab or -tablet) — isolate tablet traffic for specific testing

Typical device performance benchmarks:

• Mobile: ~83% of spend, highest volume, slightly higher CPA
• Desktop: ~15% of spend, lowest CPA, highest CVR
• Tablet: ~2% of spend, variable performance
• TV screens: negligible spend, no conversions

Create phone-exclusion variants of top-performing campaigns to capture desktop traffic at lower CPAs.

Campaign Duplication & Scaling

Scale winners through duplication, not just budget increases. The framework:

1. Identify winners — campaigns with CPA within 10% of target and stable for 7+ days
2. Duplicate with variation — create a new campaign with the same audience but a different:
   • Landing page variant (different uid parameter)
   • Slightly adjusted tCPA ($5-$15 higher or lower)
   • Modified budget (start at 30-50% of original)
   • Different creative rotation
3. Iterate naming — increment the sequence number (women-2 → women-3 → women-4)
4. Run in parallel — let both original and duplicate compete

Scaling signals (when to duplicate):

• Campaign hitting daily budget consistently
• CPA stable at or below target for 5+ days
• Conversion volume > 5/day
• "Limited by budget" status in Google Ads

Pause signals:

• CPA > 120% of target for 5+ consecutive days
• ROAS < 0.70 with no improvement trend
• Zero conversions in 3+ days with meaningful spend

Budget Allocation Framework

Distribute budget based on proven performance tiers:

Use non-round budget numbers (e.g., $8,888.88 instead of $9,000) as a convention to quickly identify campaigns managed by the team vs. auto-suggestions.

Additional Resources

Reference Files

For detailed implementation guides, consult:

• references/segmentation-deep-dive.md — Complete audience segmentation patterns with examples for 10+ DTC verticals
• references/naming-convention-guide.md — Full naming convention system with parser logic and examples